Thank you very much for being here. Yet another OpenWorld, yet another amazing week. Couple things to take care of here. First and foremost, you'll note the date is September 22. I want to take care of some personal business and extend a very happy, warm birthday wish to my mother. 78. I know she's listening. She's worried about her little boy. Is he okay? Is he hanging in there? There's a lot going on. Mom, I love you. Happy birthday. Other housekeeping items. If you have not seen Larry Ellison's Tuesday keynote, watch Larry Ellison's Tuesday keynote. Phenomenal. Just a lot of clarity on some topics that I know that are on your minds, and you can't get any clearer by what you're going to see. Watch it. Basic housekeeping. I think you've all done the drill. Thankfully, we've all been here many times.
The restrooms are over there. The Wi-Fi's in the air. The cables are on the floor. Don't kill yourselves. Now let me do my favorite slide. Dorian, are you here? Dorian? Hi, Dorian. Dorian, this one's for you, and the whole of the entire organization. I need to do this, and so just as a heads-up, you know how this works. Basically, I'm going to go through some of our slides around safe harbors and whatnot. The idea being here is that each of the speakers as they come up will make reference to what I'm talking about here and now, so that we don't have to run you through it each time. The statements in this presentation relating to Oracle's future plans, expectations, beliefs, intentions, and prospects are forward-looking statements and are subject to risks and material risks and uncertainties.
Many factors could affect our current expectation and our actual results and could cause actual results to differ materially. A detailed discussion of these factors and other risks that affect our business is contained in our filings with the SEC, including our most recent reports on Forms 10-K and 10-Q. Copies of these filings are available online from the SEC or by contacting our investor relations department. All information set forth in this presentation today is current as of today, September 22, 2016. Lastly, Oracle undertakes no duty to update any statement in light of new information or future events. In addition, we will be making forward-looking statements, and we will use non-GAAP information in the presentations. Lastly, as I just want to make sure each of you understands that the presentations are being made today for informational purposes only.
Thank you for indulging that. Now let's step through the agenda for the day. Welcome. Following myself, Thomas Kurian will be coming up, and he will be speaking with you all, talking about Platform as a Service as well as Infrastructure as a Service. He'll also be having a customer panel where we'll have a number of customers, you can see the chairs up here, coming up and sharing with you their experience using not only our infrastructure, but our Platform as a Service offerings. We'll take a break. Now that break will probably run right around 11:00 A.M. About 2 hours from now, we'll have a break.
As we come back from the break, Steve Miranda will come up, and he will speak with you about Software as a Service, and Steve likewise will have a customer panel where he will bring up some customers to share their experiences. That will take us to right around 12:30 when we will have a break. There will not be any presentations or discussion during the break. It's an opportunity for you to informally meet and talk with each other, and we will resume the conversation and the meeting at one o'clock. At one o'clock, Doug Kehring, our EVP of Corporate Development, he'll be up here speaking with you about a project that he's been working on, leading an initiative which has sponsorship from Larry, Safra, Mark at the highest level to basically how we are transforming the way that we work and engage with our customers.
He'll be up here speaking about that. Mark will come up, and he'll follow up the conversation we started last year. We had a long conversation with you about what was going on as we were migrating or transforming from an apps company to being a Software as a Service company. Started that conversation and said that, "Look, when I come back next year, I hope to talk more about what's going on from technology software going to Platform as a Service." He'll be starting that conversation with you today as well. Afterward, he will join Safra, and the two of them will do a Q&A with you to talk about the things that you would like to talk about, the executive Q&A at that point. We'll have a break just a little bit before three o'clock, and then we'll bring Larry back, and we'll do the Q&A.
This is the agenda for the day. We look forward to having a great day, I think what we'll do is we've got a video, and then we'll bring Thomas up.
The modern CFO is in an enviable position. Their point of view is distinct. Finance is uniquely qualified to guide the way through the upheavals of digital transformation. They manage the flow of money, giving them an unrivaled outlook on operations. To spot opportunities, eliminate barriers, create value. To effectively deal with disruption, modern CFOs must disrupt their own organizations in order to become the guidance system needed to move the business forward. Finance is no longer just about transactions, but transformation. The modern finance organization is becoming the go-to source for decision support and delivery of data and insight to drive strategy. Pair finance's unique position with the social, mobile, analytic, and cloud technologies of Oracle and successfully guide your organization through disruption.
Ladies and gentlemen, please welcome to the stage Oracle President, Thomas Kurian.
Good morning. Thank you for coming to this investor conference. I am going to focus the session on Infrastructure as a Service and Platform as a Service. To give you a sense, when we started our cloud effort, we said that we were going to build an integrated suite, Infrastructure as a Service, Platform as a Service, Software as a Service, and Data as a Service. We now have a fully integrated cloud offering that covers all these pieces. A customer can choose to use one or more of these pieces, and you will hear from several customers, and I will also give you examples as I go of customers who use one or more of these pieces together. Okay. Over the last several quarters, we have had very strong growth in customers.
Today, over 60,000 customers use our cloud, whether it is Infrastructure as a Service or Platform as a Service or Software as a Service. There is 270% growth over the last eight quarters or 10 quarters. 27 million people log in every day in the cloud. That is 170% growth in that period. They process 60 billion transactions every day in the cloud, 50 billion transactions every day in the cloud. Okay. All these growth metrics come because of the technology that we have built. I am going to start by talking about Infrastructure as a Service and what our strategy there is. What are we doing with Infrastructure as a Service? What we are giving customers is the ability through a browser or through an API, an Application Programming Interface, to create a very high performance, highly secure, software-defined virtual data center that runs in the cloud.
As a customer, you can go in and say, "I want a software-defined virtual data center. I need it near my operations, so I want to stand up in the western part of the U.S." You get an account. You can then go against an API and create a local area network. You can then carve it into virtual subnets. You can attach high performance compute servers to it. You can attach high performance storage to it. You can add network devices, firewalls, load balancers, et cetera. Then you can deploy a workload on top of it. It is all done elastically for you through an API. What we have really done is taken hardware and made it programmable so that you do not have to manually stand up these things, wire them together, and you get them elastically. Here are our regions where data centers exist.
I think you see a broad number. We continue to expand the coverage in many more countries every year. Okay. In every region, we have three highly available data centers. We call them availability domains. They are not on a single power grid. They are not on a single ISP network. Typically, they are peered nine ways with ISPs. They have a terabit per second bandwidth and less than five-millisecond latency between the regions. What that allows a customer to do is build a very highly available system with three replicas within a region. If you're worried about the entire western U.S. having a problem, you can also build an application that spans one region across for disaster recovery for remote DR. Our regions are connected on a fiber ring so that you have a lot of bandwidth and very low latency as well.
The heart of our innovation in Infrastructure as a Service is what we've done with the actual design inside the data center. If you look at a software-defined Infrastructure as a Service, historically, and in all other platforms, virtualization has been done in compute with something called a hypervisor. What we did was move virtualization from being done in compute to where it should be done into the network. Our data center network is a very broad network. Every customer's traffic is isolated in a private layer 3 virtualized software network. Okay, I'll talk about why that's better for security first. Most importantly, three other things. It allows us to present customers a compute server with zero Oracle software running on it. It's a raw physical host, which means there are three benefits.
First of all, you can install as a customer your own stack on top of it. It gives customers the ultimate in portability between their premise to the cloud. We have a lot of customers talking about how easy it is to move workload because they can bring their own stack on top of it. They do not have to use any Oracle software or any other operating system that they're not comfortable with. One is portability. Number two is security. Because there is no Oracle stack running on it, and there's no hypervisor running on it, not only are you protected from anybody's software coming in and providing a backdoor into your compute environment, you're also isolated from any other customer software running on the same physical server.
Okay, for example, there's a company that runs logistics for the U.S. into areas of threat like Syria, Iraq, Afghanistan. By law, they are required to run on compute hosts where there's no one else on it. Because our compute hosts are physically locked down, there is no ability to share an address space with anybody. They use our cloud for this purpose. Okay? The third benefit. One benefit is portability, the second benefit is security, the third benefit is performance. Because if you want the ultimate in performance, and you will hear from a customer today, you don't have to run a hypervisor. You get a raw physical host. You get really good performance. The first big innovation we've done relative to any other cloud vendor is to move virtualization into the core network.
As a result of it, we can present customers a raw physical host, we can provide them unbelievable security, and we can give them ease of migration. Now, if you're a bank or a financial services institution or a Department of Defense customer who on premise has standardized on a certain set of physical appliances. Let's say you've got a Palo Alto Networks physical security device, or you're using a vendor's security intrusion detection and monitoring device. In other clouds, it is very clumsy to plug those devices into the network, because the network is virtualized with a hypervisor. Because we have this new design for where we virtualize, we present the network as just an IP address and host name endpoint. You can attach your own physical devices into the network, just like plugging it in to an Ethernet port.
It is actually literally integrating an Ethernet port. Why is that relevant? When big corporations move their workload to the cloud, they want to make sure that the perimeter defense, the load balancing stack, et cetera, that they have standardized on their premise, they can continue to use in the cloud. With our cloud, it is very easy to do that. If you look at the traffic in the network, you've got three customers, one, two, three coming in. We have a shared physical network, but all customer traffic is isolated in their own private overlay. Think of it as this physical network looks to individual customers as though it's a private pipe. What's the benefit of that? It's security built in the lowest level of the infrastructure, right into the network.
You don't have to, therefore, do necessarily cryptographic encrypt, decrypt while your traffic is running because you're isolated at the physical network. Therefore, it gives you much better performance. It also gives you much better isolation. We also have quality of service on the network. Why is that important? In the cloud, performance is not just raw performance, but how predictable the speed is. In our network, because it's isolated in this way, if two customers, let's say customer one is running ERP and customer two is running a big Hadoop job, they don't step on each other, not just on the compute side, but even on the network. Therefore, we can not only guarantee better performance, but we also guarantee better predictability of that performance, which is very important for customers moving mission-critical workloads.
As a customer, we provide you a variety of choices on how you can connect your data center to the cloud. For example, if you want to see our cloud as just an extension of your corporate data center's IP address space, you can do that. If you want a 10 gigabit per second link into the data center so that your traffic doesn't go over the internet, you can do that. If you want to run a software-defined VPN, you can do that as well. They're all available. Attached to this network substrate, we also differentiate on compute and storage in two important ways. Our compute servers are ultra-dense. They are specially engineered so that you get ultra-dense footprint. On a compute node, you can run as much as a terabyte of DRAM and 29 terabytes of flash.
That's 30 terabytes, essentially, of local memory, which means you can fit almost every OLTP database directly in memory, which means you get amazing performance because you never have to go over the network to disk. Secondly, because that memory is local resident on the compute host, you get very, very predictable performance because you are not going over the network except if you want to do so. We've also attached high-performance storage to it. We have a distributed storage environment with local flash high-performance block and object storage. How many IOPS can be delivered to one physical host? 4 million IO operations per second. IOPS are essentially a direct correlation of read-write speed that you're going to get. You're welcome to look at our website and also the competition's website.
For 80% of the price, meaning for 20% less, 80% of the price, you get 2 times the number of cores, 2 times the amount of memory, 4.5 times storage, and 11 times the speed. It's a very different offering than is available in the market today for Infrastructure as a Service. We present this Infrastructure as a Service in four or five different ways to customers. A customer can say, "I just want a bare metal host on which I want to install my own operating system and my own hypervisor." We call that bare metal. A customer can say, "I really just want hypervisor, elastic compute." You can run elastic compute. We run Linux, Windows, Unix, Debian, CentOS. You can pick from an array of different operating systems to give you choice of what you want to deploy.
If you're a customer who is building a next-generation application, typically designed around an architecture called microservices, you can stand up something called a container. It's essentially a Docker container running either within the hypervisor or on raw physical host. The benefit of running on a raw physical host compared to everybody else is you don't pay the tax of the hypervisor's networking layer in order to then run Docker. You're running Docker on a raw host, there's only one network stack running, therefore you get way better performance. If you have VMware on your premise and you want to move it to the cloud, we provide a mechanism based on technology called Ravello to lift not just an individual VMware virtual machine, but a collection of a number of them. We preserve all the virtualization, network virtualization, everything, and you can move it directly into the cloud.
You will hear from a customer today who uses that technology. For customers who say, "Listen, I really, really want the maximum in isolation," you can buy a rack of compute. Same software, but fully isolated rack. It's called dedicated compute. If you want Exadata or high-performance database environment, for example, you can tie in Exadata into this environment. Exadata, SuperCluster, a variety of these. We present this high-performance Infrastructure as a Service in a variety of different ways. For storage, we offer similarly a broad profile of capability. The model essentially starts if you're running on a compute host, you get local flash. Unlike other solutions which require local spinning disk, we give you local flash, which is way faster than spinning disk. You can run all your calculations off flash.
If you want to do persistent writes, you go to high-performance block storage. If you want to replicate data across a bunch of virtual machines, you can take a snapshot, replicate it into the object store, the object store will replicate it across a bunch of virtual machines or hosts. If you want, let's say you're doing backups into object storage, and you say, "The latest last week's I want it in object storage, but anything older than a week, I want it archived," you can simply build an information lifecycle management policy, and we rotate it into archive. Archive is extraordinarily cheap. It is $12,000 a petabyte a year. Okay. $12,000 a petabyte a year. You will hear from some customers who are using it for archive. You can put file and database backups.
What we do with the storage cloud software appliance, see, big corporate customers have many applications they've written over the years that expect NFS or the Windows protocol SMB, which is the file system protocol. When you move objects to the cloud, they struggle because those applications don't understand object-oriented storage. What the storage cloud software appliance does is a Docker container you put on your premise. It frontends our object storage and makes it look like NFS. You can use the benefits of the object storage without touching any of your on-premise applications. That's in a nutshell what we're doing with Infrastructure as a Service. How can you migrate workloads onto this platform? We can migrate a range of different kinds of workloads. Let's take databases.
You simply back up a database using the same tools that you use to back up on-premise, whether it's Oracle or MySQL, then you can instantiate a database off the backup. We had a customer last week back up a 12 terabyte database into the cloud and instantiate a brand-new copy in less than two hours. Okay, it's literally trivial to move a database. We have a technology also available called GoldenGate, where if you're running an OLTP application on your premise You're worried about taking it down while you're migrating to the cloud? You can keep it up and running on your premise. We can replicate the copy into the cloud, you can switch over your DNS. It takes about eight seconds, you're up and running in the cloud. No one else offers such a capability.
You can migrate programming languages and your apps in a variety of different programming languages, not just Java, but Ruby, Python, Node, et cetera. Operating systems, I told you we have all the popular operating systems covered. We talked about what we can do for VMware and KVM, where there's a huge base of customers out there who want that capability. Then you can also migrate engineered systems to engineered systems because we have Exadata offered in the cloud. You have a variety of different choices. Here are some customers who use us for Infrastructure as a Service. You'll hear from several of them, I'm also going to give you some examples. Why do they use it, and what's the real business value of it? Think of a customer running a workload on their premise. They spend 30 This is a real customer's example.
It was for a typical workload, an application workload. It was a calculation of what they were spending on a monthly basis, in thousands of dollars. They were spending about $30,000 on facilities, $120,000 on hardware. It was a very big app. $120,000 on hardware. There was additional spend on software. When I say software, not Oracle software, because the database is above that layer. This is on operating system virtualization, Linux, hypervisor, and tools. They were not using any Oracle software in this. Then they had a set of people costs that were responsible for installing servers, wiring devices, managing the data center, connecting, handling SSL certificates, IP addresses, et cetera. They were spending about that much on infrastructure. When they move to the cloud, the total cost for them, including the cost of their DevOps people, was $440,000.
They were saving about 30% off the top on their infrastructure. For a customer, the value is they can save 30%, not on just the hardware, but on the total cost of running and managing infrastructure, which is big. For Oracle, the opportunity is none of this, almost none of this revenue came to Oracle in the past. It was somebody else's hardware, somebody else's virtualization technology, et cetera. Our differentiators. Much better performance. You will hear from customers, not just at this event, but there's lots of customers benchmarking and testing, and they find that we have not only better performance, but much more predictable performance. Because of the design of our network and because we don't do something called oversubscription. Oversubscription is where you say, customer A asked for a compute host, customer B asked for a compute host.
I'm going to try and put them on the same compute host because I'm assuming they're not going to be busy at the same time. If they do get busy at the same time, you're going to contend for resources. We don't do that in our cloud. A lot better scalability and availability. Scaling, the heart of it is we have so much denser servers and storage that there is a lot less traffic that goes over the network, and because of the design of the network, we have much better scale. Security. From a security point of view, I talked about the fact that every customer has three important things that they really like about the security profile. One is that their network traffic is isolated. Number 2 is that the physical host has no Oracle software running on it.
Number 3, they have a variety of things that we do for segregation of duties, where our people don't even look at their software. We don't have access to it. We're locked out from that environment. It is a completely isolated network and server environment. Therefore, they're very comfortable with it. A small example. Our control plane, which our administrators use to access all the servers, is not even reachable on the internet. You can't even access it from the internet. No one else has that capability. Defense departments, security-conscious companies really want that because they don't want a threat to happen from the internet into our control plane and out to the compute guests. Governance and compliance.
One common thing that customers want is the ability to say, "I want to procure centrally and then distribute it to my departments, and I want to monitor the usage." We have a capability called Compartments that allows a CIO to buy centrally, distribute budget and quotas to different departments, set up security policies on it, and then monitor usage. That avoids the sprawl of accounts that customers complain about when they go to the cloud. "Hey, everybody's buying and bypassing my central procurement process." That's also important for customers. Portability of workload. We've had customers migrate very complex applications, ERP, a 60 terabyte ERP instance was moved from premise to our cloud in 12 hours two weeks ago. Okay?
Portability, at some point, it becomes how well are you designed to mimic the protocol, the network design of the floor in the data center, thereby making the move of workloads much easier. Lastly, we recognize that a lot of customers say, "Hey, I really like the public cloud, but I'm not able to go there right now for a set of applications." It might be because of jurisdictional reasons, it may be because of regulatory reasons, or it may be just a comfort level that the board of directors has. We are the only cloud vendor that offers the entire stack that we run in the cloud, is available on a customer's data center floor as a subscription or metered service. In fact, the price of it is the exact same as in the Oracle Cloud. What's the benefit for a customer?
You can standardize your DevOps tools, meaning the tools your programmers use to create cloud environments across the public and private, because it's same software, same API, same tools. Your choice of deployment, you can decide later rather than upfront, because we give you choice commercially as well on where you can deploy. Okay? Here are some customers who use our cloud. Pfizer is a project we did with the Accenture Business Group, Oracle Accenture Business Group. They're building pharmaceutical science lab for their drug research teams. When they're doing research, they don't want to have physical infrastructure for their scientists. They want to just spin up a lab, put the statistical analysis software, their clinical data securely in the cloud, allow them to run their experiments, and quiesce it. They run it on our cloud.
General Electric is using our compute platform to run their Predix capability, which is their smart sensor digital industrial internet platform, as well as they're migrating a number of bespoke and packaged applications to our cloud. Grant Thornton is a public reference. They moved from a competitor's cloud to us, not just their test and development, but also their production applications. They're a large ISV. The U.K. Government is migrating a large number of legacy bespoke application workloads to Oracle's infrastructure and platform as a service. Comcast, the cable company, is using our platform to build the next generation suite of applications for its Xfinity service. Riot Games, they are the company behind the League of Legends. They're moving their production workloads to our cloud infrastructure as a service. T-Mobile, their leasing application, which they use for financial leasing, required 99.99% availability.
They run it on our Infrastructure as a Service, production and disaster recovery. Brink's, which is the security service company, the armored car service, is migrating their entire archive of documents, which is over 250 petabytes, to our cloud archival storage solution. The State Bank of India, which is the largest financial institution in India, had a regulatory requirement to keep the cloud platform within the company's boundaries, within its data center floor, and they're using our Cloud at Customer offer. There are many, many, many more such examples. These are a short list just to give you a flavor of what people are doing with our Infrastructure as a Service. Okay. Let's talk about Platform as a Service. What's the core strategy behind Platform as a Service?
The core strategy behind Platform as a Service is that companies spend a lot of money doing very mundane things with software. Installing software, configuring software, patching software, tuning software, backing up software, setting up disaster recovery, et cetera. If you took any piece of software, take Oracle Database, take MySQL, whatever it is, and ran it on Infrastructure service, what Infrastructure service is optimizing is giving you compute network storage elastically cheaper. As a customer, you are still responsible for installing the database, installing the application, patching it, backing it up, maintaining it. What we're doing with Platform as a Service is fundamentally eliminating the need for any human to ever have to do that ever again. We're doing it by automating it through software. Okay. You come in the same account. You can come in and log in.
You can choose from a catalog of these services. There's about 60 services in 8 categories. Let's say you want to create a database. You can come in and say, "I need a 10 terabyte, 16-core database. I need disaster recovery. I need clustering. I want it fully encrypted. I want it patched on the last Friday of the month. I want it backed up every night at midnight." You submit that both through an API or through the console. It's about five screens. You get a database 22 minutes later, and you never have to touch it ever again. You never have to patch it. You never have to back it up. You never have to set up disaster recovery. You never have to do encryption. All of that is done for you through software. Why is that attractive to customers?
It's attractive in three important ways to customers. The first way is it allows departments within an organization that never had IT resources like DBAs, or companies that never had IT resources to be able to use the software. Number 2, it allows an organization that is bottlenecked in legacy system maintenance to free up people to do projects because the work is automated in the software. For example, you can take all your sys DBAs, who are responsible for patching and backup, and help them focus on application performance tuning. The third thing it does, because the software is doing this, 100% of every piece, every database, for example, of which there are over 100,000 in our cloud, every instance is identically configured. For a customer, you get 100% repeatability of that config.
You don't worry about why is this issue happening in this test environment and not happening in production. They're all identically configured to software, you don't have idiosyncrasies because one human did it one way and another did it another way. The heart of what we're solving with Platform as a Service is to eliminate the bottleneck in IT by automating how people can get a platform and use it to build cloud services. There are eight categories. I'll go quickly through the major ones. Data management. We provide the ability to run Oracle Database, MySQL, and a couple of different flavors of NoSQL. You can also run SQL Server on our Infrastructure as a Service and other databases. Okay? With Oracle, we've automated everything. Same thing with MySQL, same thing with something like Cassandra, which is an open source key-value database.
You literally come in and say, "I need a key-value store. I need x amount of compute. I need so much scaling. I want you to monitor CPU. If CPU goes over 50%, scale it out. If it gets below 40%, scale it back down," and the software takes care of everything. Okay? We provide a lot of capability in our data management platform for people to run not just small databases. We introduced a product to allow you to run a very fast database for $175 a month. That's $2,000 a year. It is cheaper than Amazon's cheapest database, and it is way faster. The reason we can do that is not only is that available, but as a user, you never have to touch the database. You simply load data and write your program against it. To build apps, we support four styles of app development.
You can build apps in a variety of different programming languages, not just Java, but also the popular scripting languages. You can build mobile apps, mobile web, mobile apps, native apps for iOS, Android, and Surface. At OpenWorld, we introduced something called Mobile Bots, which allows you as a user, if you're a millennial, for example, and you really want to use Facebook Messenger or WeChat, but you want to interact with enterprise apps like ERP or HR, you can literally be inside your Slack channel on Facebook Messenger and interact with enterprise apps using this platform. We give you the ability to open source development. Okay? We certify close to 600 different open source packages, not just for development, but what engineers call the DevOps pipeline. I can check in source control, I can compile, deploy, and run.
There are many customers now using straight open source, no Oracle software, on our platform for development. Okay? For microservices, if you're building a new style of app, we also provide you a Docker platform with the microservices framework on top of it to allow you to build microservices. Okay? Here are some people who use our database cloud. Here are some people who build apps with our app dev platform. Here are people who build with mobile. Okay? Analytics. What are we doing with analytics? Our platform offers the following capability. If you're a customer who wants to move data into the cloud, you can sweep data into our object store. We have people running huge volumes, 50, 70 terabytes per day into the object store. You can also stream real-time data in through a technology called Kafka, which is very popular for real-time streaming.
The data comes into our object store, then you can instantiate a Hadoop cluster. It takes about two minutes to instantiate a Hadoop cluster. On top of which you can run MapReduce or Hive, most importantly, we have a lot of work going on with something called Spark, which is very popular in the Hadoop space now. It allows you to do streaming calculations, so you can look at a data stream coming in in real time and do calculations. You can do machine learning on the platform, and you can do SQL as well. Our differentiation here is the performance of the platform, the ease of use, because you don't have to manage the cluster. We're literally spinning up the cluster for you. You just describe the workload you want, and we spin it up for you.
Because we have very dense compute nodes with a lot of memory, you don't have to go from your Hadoop node back into the storage system for read/write performance, you just get it all locally. Many people use a technology underneath this called Tachyon for really great distributed I/O performance. Okay? If you're not a hardcore data scientist who loves writing Spark Scala, there are lots of people who like doing that, but there are many more who are like, "I don't even understand what that is," we give you a toolset on top of it. Okay? You can go into a graphical user interface and model ETL with something called Data Flow Machine Learning. You can cleanse your data using data preparation, so you don't have to go in and manually get your data lake dirty, as they call it.
Then if you're a scientist who wants to collaborate on the data, we offer something called Big Data Discovery, which allows you to build statistical models on sample sets of the data. Okay? We give you a full toolset to do analysis on top of this platform. End user analytics. Starting about five years ago, we wrote a brand-new toolset for Data Visualization and analytics. It was meant to solve a very simple problem. If you look at the average company, 20% of the people in the company have access to a data warehouse, the reason is pretty simple. It's a data warehouse, it has lots of data, people don't want to give everybody access because somebody may make off with the data.
For 25 years, almost 30 years, all software vendors in BI were focused on that 20% of users, who are the power users who access a warehouse or a mart. We asked a question, what happens to the 80% of people who don't have access but would like to do analysis? We built a brand-new product called Data Visualization. It had only two premises. You need a browser, you need to know how to load data using Excel, which we think most of the world knows how to do. You can bring your data in Excel, we load it, blend it. You can essentially do ETL within Excel. Okay? You can cleanse, standardize your data, then we generate underneath the entire OLAP model that you need.
You don't have to know SQL, you don't have to know what a cube is, you don't have to know what an index is. It's all done in memory. You can then visualize your data, and then we synchronize it to a variety of different devices. If you guys were using it and you're flying from New York here, you can detach, do analysis on your desk, on your laptop, get to your hotel, plug it back in. We synchronize it to our cloud and automatically distribute the analysis to all your devices and to the colleagues you want to share it with. We made analytics, the content of analytics, exactly the way that you would use iCloud to synchronize your contacts, for example. Here are some customers who run big data in our cloud.
Ambev, for example, the big distiller, they moved from another cloud vendor's MapReduce platform to our analytics big data platform because of the performance and scale of the system. We have lots and lots of users running us for analytics. Kentucky Fried Chicken, they want to do analysis for every restaurant, and they didn't want to have a desktop in each restaurant that's required to run analysis, so they do it all off the cloud. You'll hear more today from a couple of customers. Most large customers are going to end up with multiple clouds, we provide an integration layer in our cloud, connect our cloud with anybody else's cloud, including on-premise. It's based on an architecture which is very popular in the industry called SOA. You can build an API gateway, which abstracts your APIs.
You get a super high-speed messaging system to connect the cloud with other systems, then you can also build a business process or workflow on top of it. For security, we can centralize the way you manage users. We can do strong authentication if you're really concerned about security. All the data in the cloud is protected by a customer's encryption keys. We do not own the encryption key, the customer does. The reason for that strong segregation of duties posture is, even if we get a regulatory request from a government agency, we're not allowed, we cannot technically hand over the data because the customer owns the keys. All the systems in the cloud. If you wanted to monitor, administer your infrastructural platform, we give you a whole toolset to do that. You can monitor them yourself.
You can set up compliance rules to make sure they're in compliance with your best practice configurations. We give you a real-time dashboard where you can see security events. If you want to remediate those, we've got a scheduling and orchestration framework which auto-remediates everything. In everyone else's cloud, you have to do that with a different set of tools, you've got to sort of stand it up yourself. Here, we give it to you built in. Here are customers who use our management cloud. I mentioned we have a solution to run this entire IaaS and PaaS stack, not just in our cloud, but on your floor. The largest bank in Chile, in South America, moved from another cloud to our cloud for two reasons.
The first one is their latency from Chile to the data centers for these cloud vendors, which are not in South America, take a lot of time. Secondly, the government requires the production data to sit on premise. They run their test and dev in our cloud. Production is on the premise. We are the only vendor that offers this capability. It's been available since March. Okay. This is a real customer example. A real customer built a web app, 4 middle tiers, 4 web servers, a 2-node Oracle RAC cluster. They did it on premise, they did it on another platform, and they did it on the Oracle Cloud using our Platform as a Service. This is their own measurement of how long it took, how many clicks and commands. Okay. The speed is a measurement of how much automation there is.
The fewest clicks and commands means the fewest mistakes and the highest degree of automation. I think the numbers speak for themselves, and I think that's what I mean by the PaaS is not about just running on the infrastructure, which it certainly does, but by automating all the manual steps that you have to do. What's the business value to a customer? This is a metric from a financial institution. It is $ per core per month to run the Oracle Database at the customer. Okay. They run about 5,000 Oracle instances. Facilities cost is $30. Hardware cost is $120. The software cost, not just the operating system plus virtualization plus tools, but also the Oracle license and support, is $500 per core per month. Then there's $850 in maintenance costs. You're probably wondering, why is that maintenance cost higher?
Because it not only included the storage network and sys admin, but it also included the sys DBA and apps DBA. On our Platform as a Service, they can do it for $1,000, which is a 30% reduction off the top run and manage. If you look at the, let's call it the scarlet boxes, the software cost, that was all that Oracle was getting before. I think in the cloud, you can see we make a lot more. The addressable market for us is larger because we're not just moving your license to the cloud, but we're also eliminating that manual work and the value associated with eliminating that manual work.
Secondly, if you look at Infrastructure as a Service, if you took that software stack and ran it on Infrastructure as a Service, what you're displacing is that $30 plus $120, then there's some portion of that software cost, which is about $100. That's the service operating system and virtualization. Even if you cut that in half, what you're doing is taking 30 plus 120 plus 100, which is about $250. Even if you have that, you're saving $75. Here with the PaaS, we're saving the customer $500, which is a lot more. There's a lot more value in what we're doing with PaaS than just taking a stack and running it on an infrastructure. Okay. There's been a lot of speculation about how is Oracle Database as a Service different than Amazon Aurora, for example.
I thought it's better we just have a frank discussion about it technically. Amazon's a fine company. They have good products. I think there are differences in reality between what we do and what they do. Aurora is Amazon's OLTP database. It's very new. It was released less than two years ago, and it started as a fork of MySQL 5.6. Okay? MySQL is an open source database. However, Amazon Aurora is not open source. It is a private fork of MySQL, and there are no changes from Aurora that have been contributed back to MySQL. Even if you just look at raw MySQL, our MySQL service is based on the open source version of MySQL. It is not on a proprietary fork. Okay? First thing is, MySQL's existed a long time in the industry.
People have taken a private fork of it, they're using it to run an OLTP system. What is Redshift? Redshift is their analytic database. It's a fork of PostgreSQL by a small company called ParAccel, which went out of business, and they bought them in 2011. Redshift is also not open source. Redshift and Aurora are not compatible, they're not runnable on any other cloud. Okay? Our database service is open, meaning you can run our database on any cloud, including on your premise. We have a single source base for OLTP and data warehousing. Obviously, there are technical differentiators. Let's go through just a small set of them. Amazon can only run a database up to 16 cores because they don't have the ability to cluster to scale out. They have no scale-out capability.
Okay, the largest they can go is one physical node or one, what they call compute shape. We can go up to 352 cores. They cannot run a database with very high performance. The maximum IOPS they deliver is 48,000, which is the top end of elastic block storage. Today, we can run 3.6 million IOPS, which is slightly more than 48,000. Aurora cannot run a large database in size. Imagine you have a, let's say, your Dell computer, and you have a customer table with all orders, purchase orders of all the customers. That's over a billion rows. In order to scan that to say, "Find me all customers who have bought a computer in the last month," you need to be able to scan the table in parallel. We call that parallel query.
You don't have a single thread trying to scan that table sequentially. If you don't have that, you can't scan large databases with parallelism. Their maximum database size is 60 terabytes. We go significantly higher than that. If you look at high availability, just two examples. Imagine you're running order entry. If one physical host goes down with Oracle RAC, Real Application Clusters, you can continue a transaction on another physical host. Okay? The competition has something called a cluster. What a cluster is they have one instance, and they do asynchronous replication to another instance. It is not a single physical instance with real-time connection. So what happens if one node goes down? You cannot continue doing your updates because technically, because it's using asynchronous replication, you can only have one writer at a time. Okay?
You not only cannot scale out with a cluster, it also has much less availability. Take as an example, imagine you have a storage corruption, which happens reasonably frequently, like a storage volume got corrupted. Because they depend on block replication for HA, the DR location would also get corrupted just at the moment you really don't want that to happen. We don't use storage replication. We use a technology called Data Guard, which is based on the database replicating based on asynchronous logs. We protect you to a much higher degree than the competitors do. Okay? It also cannot protect you for high availability if you have a secure database. If you encrypt your database, you cannot replicate an encrypted database. You either get security or you get availability. You don't get both. In a nutshell, there are lots of technical differences.
I understand that they have a market that they're selling into, they're a fine company. When people say what they offer can easily replace you, it is absolutely not factually true technically. There are very elementary differences between what they do and what we do. Okay? Here are some customers using us for PaaS. Mazda, the car manufacturer, runs their replanning calculations. Supply chain replanning, it's a custom statistical program, 24 million parts and orders. Every time a specification changes, they run it on our Java Cloud Database Cloud. Samsung runs all of its in-store merchandising, sales, and transaction processing on the Oracle Cloud. There was a press release earlier this week about it. Manchester Airport, 48 million passengers travel through their airport every year.
They wanted to provide live updates to people on a mobile app saying, "Your bags are delayed," or, "Go to this checkout counter rather than this." They use our Mobile Cloud Service to build that. Toyota is moving many test and dev and production instances to our cloud. They also use our cloud in North America for all their analytics for motor sales. Skanska, a Swedish construction company, runs all its financial analytics and corporate management reporting, including defining what they're going to report in their 10-Ks and 10-Qs using our cloud. Gilead, which is a pharmaceutical company. They use us for analytics, they've opened it up to all their line of business employees because we've got this solution, which I mentioned, which is simpler to use, you don't need a data warehouse expert.
Avaya uses our Oracle Cloud to integrate its front office and back-office applications using our integration service. KPN, the national telecom company of the Netherlands, they wanted to run their test dev middleware database environments. Because they have many applications that have telephone numbers in them and telephone records, they are not allowed to put that in a public cloud. They, just like State Bank of India, they are going to run it on our PaaS on the Oracle Cloud at Customer. Trek Bicycles, for those of you who mountain bike. They have a fan club where people can challenge one another for races, something like 60,000 or 80,000 people. They collect the real-time information coming off the sensors on the bike, they share it with other people through our Mobile Cloud.
A variety of different examples, all kinds of different scenarios, and using a variety of different services from our cloud. That's a very quick tour of Infrastructure and Platform as a Service. I'd like now for us to get some customers up so we can talk to them about their experience using some of these services and the value that they get from these services. Thank you.
While we have our guests coming up on stage, for those of you out there listening on the webcast, please rejoin us in approximately one hour.
Good morning and thank you, everyone. Ken, I don't have to read this, right? Okay, good. Good morning. My name is Steve Miranda. I'm responsible for the applications development, our SaaS apps at Oracle. I'm going to go through today essentially our definition of our SaaS apps cloud. In particular, what I'm going to focus on and why I want to define it is to help you understand what we think are fundamental differentiators, both against other SaaS-based vendors, especially as we get into ERP and supply chain manufacturing against our traditional on-premise competitors, because we see that as a high importance to our customers. I'll also give you a sense of what we see differentiated going forward and where we're making our investments. What I'm going to really spend the bulk of the time is we feel great, particularly on SaaS.
This week's been a celebration of our customers. If you've been with us for a while, you've been at this show for the last several years, you've really seen our themes go from we're going to build a set of cloud-based applications. Do we have our first set of cloud-based applications? Do we have a set of cloud-based customers now? Last year, really completing our SaaS-based suite with the introduction of our manufacturing supply chain onto this year. You'll see this expressed in our customer panel, really hundreds if not thousands of customers live on every pillar in our suite, from CRM, HCM, financials, including EPM, and now manufacturing supply chain as well. I'll give you some of the examples as we go through and some of the importance of them.
Obviously the key highlights, we are now over 12,000 SaaS customers running our system in the cloud. We have more than 10 times as many ERP customers in the cloud as our closest competitor, with Workday. Over the last fiscal year, we're growing our HCM Cloud at about two and a half times the rate. In marketing and other CRM areas, being the top and extending our lead, and particularly with Data as a Service, which I'll cover in some detail near the end of the presentation.
What we are seeing now, what we've been saying for some time, is that it's not a question of if, it's a question of when 100% of our existing, and I'll show you not only our existing customers, but other customers will move from their traditional on-premise applications, either ours or our competitors' or third party, but into our cloud. Now to the differentiators. First off, I think it's important to understand in a SaaS definition of cloud, what cloud is not limited to. It is not simply a hosting of the applications. We offer that for a traditional service. Some of our existing competitors will try to talk about a cloud-based application, but is really their existing older platform repurposed and rehosted. It is not simply a subscription versus a license, a subscription fee or versus a license fee.
The reason why we think 100% of our customers are moving to SaaS, and the reason, frankly, why we did all these years of investment to rebuild our products from the ground up, because if you think about it, if that's all SaaS was defined to be, and those are the only things that would've been benefits for customers, we could have taken the E-Business Suite or Siebel or others, hosted that and called it a day. We didn't do that. We did significant investment, and why did we do that? The fundamental reason why is we could now deliver better software faster to our customers than we ever could before. We can deliver applications that we never had a hope of delivering and no one has a hope of delivering in the on-premise model. Save the new applications for a little bit later.
First off, what do I mean by we can deliver better applications faster than we ever could before? The example on the slide, it says three to four times faster. That is probably understated. In the old model of the world, we had a software release. We release it probably once every two years, maybe two and a half years, maybe sometimes up to three years. Why was that the case? We could have taken the same level of development, broken it up into smaller pieces, and released every six months like we do in the cloud. The fundamental impediment to that is that our customers couldn't consume it. Very common, almost universally in an on-premise implementation, the customers take our applications, whether it's the E-Business Suite or Siebel or SAP, by the way, and frankly, more common SAP, and they customize those applications.
The upgrade that we deliver, it's impossible for a customer to take that upgrade vanilla, run it, because they would have to check their customizations, make sure the upgrade work with their customizations, oftentimes redo their customizations and/or replace their customizations. As all of you know from our existing customer base in that profile, in that model, it's a very slow, very expensive model. We, and I'll use we the industry, ourselves, SAP, everybody in the on-premise model released roughly once every 2 to 3 years. The innovation cycle had that for real innovation cycle, not just developing software, but developing software and customers actually implementing it and using it conservatively was once every 2 to 3 years. That's if we just count how long we took to build it. More often than not, customers wouldn't start adoption.
You guys cover this when we have a new release. Adoption's going to start really a year after it's out, 2 years after it's out, et cetera, and then implementation takes another year or 2 years after that. You're talking very commonly 4, 5-year release cycles, real innovation release cycles. For our engineering teams, and again, I'll use the industry to actually get feedback on how well we did. Did what we build actually work? First of all, that release was usually long forgotten 5 years later. 5 years later, how do we do? The business problems have changed. The fundamental assets changed, and it was very, very imprecise. That's the old world, at least 5 years conservatively.
Today, every 6 months, we have a brand new release of our SaaS-based applications, and we deliver that to each and every one of our SaaS-based customers that I had on the prior slides immediately because they're all hosted in our cloud, and importantly, none of them have any customizations that preclude the upgrade and that need a redo. We deliver that to them. Because it's hosted in our cloud, we can measure precisely how well we did. What do I mean by that? As Thomas described, we never see, in fact, it's impossible for us to see our customer data. We do extensive analysis on what our customers use in our software. Every click the customer makes, we measure the time. Any slow clicks, we fix that immediately. We measure what features are being used. We measure what features aren't being used.
We measure how the features are being used. We measure certain ratios of what customers are doing in terms of feature usage, and it gives us immediate feedback to then incorporate into the next release. You've gone from at least a conservative 2 to 5-year release cycle to a 6 months release cycle. You've gone from little to no, and certainly at least imprecise feedback to extremely precise feedback. We can deliver better software faster and keep this product more current. As an example, you guys all live this in your own personal lives now, I'm sure every one of you gets frequent updates to the laptops that you're on right now. You get the updates. You get new features. You get new user interface going. Your phones, you have automatic updates on your phone. Automatic updates. Now, all of us have configuration in the products.
We all have our own calendar settings. We have our own contacts in our phone. Maybe you have certain apps that you run that you've extended it. Those are extending the applications. Those are not customizing the applications. You extend it, you can stay modern and current with the latest new features, but you don't need to redo everything you've done before, and all of us can be personalized even though we're on the same device. That exact same analogy in what we've done in rebuilding our SaaS applications. Frequency of updates, frequency of innovations, precise measurement without compromising on personalization, and you can't do that simply by taking your older technology and hosting that on the web. That technology was built with a customization, invasive changes in mind, and hosting is very different from SaaS on a go-forward basis.
By the way, if you want to do a test, just at superficial level to see what the benefit this is, if you have an old phone somewhere, or if you have an old tablet somewhere, and you haven't run those upgrades for, let's say, three years even, not that long, if you look at an enterprise software world, SAP on-premise, Oracle on-premise software, three years is not very old. Look at how old the user interface looks if you haven't upgraded your mobile device or laptop over three years. Look at some of the applications that will not work on that application. Same thing is happening for enterprise app, same need for immediacy of speed, and same reason we wrote the apps. In addition to that, we had to transform a lot of the ways that we actually engage and do business with the customer.
Doug Kehring, who follows, will talk a lot about our improvements in terms of zero hassle buying, simplified our pricing experience, having accelerated buying. We went to great lengths to ease the implementation process because the mindset of a customer can change now. When you get software delivery every six months and new and improved changes, as opposed to once every four to five years maybe, you can think in terms of phased implementations. Because you know that whatever you do in phase 1, you're going to get software delivered that's quickly for phase 2. If you were only going to get a delivery once every four or five years, the customer mindset used to be, "We have to get this all in now." We eased that implementation towards a much more adoptive implementation and made our entire service delivery much more cloud-based.
Again, that only happens from re-engineering your applications top to bottom, not from simply hosting your old application suite. That's what we mean by SaaS. Now, what are some of our key differentiators within that context versus other SaaS players? Our SaaS applications are complete, data-driven, personalized, connected, and secure. Let me go through them one by one. Is there a way to have the current slide updated here? It'd be helpful. First of all, complete. Complete, we mean robust functionality for not only multinationals, large enterprises, but also smaller enterprises across multiple product pillars.
You'll see when I bring up the customer panel, most of our customer base have started in one product pillar, whether it's CX or even a subset of a product pillar like marketing, expanded to sales, service, moved from that into the order management, order capture area, and/or supply chain manufacturing, moving to that into core financials, including consolidations, and moving to that into HR, including of talent management and payroll. A complete suite of applications that customers can adopt in phases wherever it makes sense for them and expand as you go, both moving from our applications and moving from any other customer application. Second one is key and critical is data-driven. We've talked about this in the past as embedded business intelligence, moving from transactional apps, which is really where ERP started, into much more business decision apps.
A key distinction here, where we've made the most investment in applications over the last several years, is data-driven not only means internal data, but external data. It used to be data-driven was you can run BI on things that happened in your enterprise, things you knew about your customer, things you knew about your supply chain, things you knew about your financial cash position. Now, we, like other applications, which I'll detail in a moment, could supplement that data with an extensive array of what we call third-party and third-party anonymous data. We supplement your internal data in the cloud with cookie data, mobile device data, consumer purchase data, and much, much more, which again, I'll detail in a moment.
The next fundamental piece that's necessary then in the cloud, because I said that we all had personalizations on our phone or our mobile device and our computers, but we were updating that. Personalizations does not necessarily complete everything for an enterprise app. You may need the ability to actually extend that application beyond what you can just tweak. By extension, we allow you to use our Platform as a Service, the exact same tools that we use to build our apps, Java, the Oracle Database, BI tools, mobile tools, all the tools that Thomas talked about available to our customers in SaaS, all the tools that many of our customers and ISVs use today to write their applications in the most popular programming language in the world and the most popular database in the world.
You can now run and extend our applications, but extend them in the same way as I described your phone, where you can still do that upgrade without having to worry about losing any of that information, or that it would block the upgrade going forward. That's what we call our SaaS and our SaaS with PaaS combinations. Then we're in a somewhat of a unique position in that we have not only our cloud-based applications, but our on-premise applications. Remember I described to you that though we have a complete suite, most customers will migrate to our cloud-based applications in a spot that makes best sense for them at the beginning and move over time. For example, we have many PeopleSoft customers who run HCM and Financials.
Their first step to the cloud has been to move HCM to the cloud, and we have a pre-built connector back to the existing PeopleSoft financial system so they can run Oracle in the cloud for HR and PeopleSoft on-premise for finance, and we pre-integrate that for them. Then over time, they upgrade the financials. Similarly, for our CRM customers in Siebel, you can move digital marketing to the cloud while still keeping your Siebel order management on-premise. We have pre-built integration for that, and now those customers are moving. What we're starting to see is not only movement around the edges, but now customers moving the core of their ERP systems to the cloud and integrating to third parties, which they've always integrated in the past, going forward. The final fundamental tenet is security.
This has really changed, in my mind, over the last year for customers really starting to understand instead of security being a concern or an impediment, I think was the earlier question a gentleman asked as far as move to the cloud, but frankly, one of the key drivers to move to the cloud. At the base level, what customers are starting to understand is security, in terms of the data, in terms of your software standpoint, it's not a one-and-done exercise. Oracle, from its very beginning and continues today, invest in security to make sure every part of the stack is secure, from our data centers to our hardware, to our database, to our application tools and middleware, on up to the applications.
At every single layer, there's innovation and security, which has to be kept up to date with the latest security threats that are out and available. We combine all of this in that investment level on an ongoing basis. It's frankly something that the vast majority, if not all of our customers, can't do in a cost-effective way in their own cloud. Most customers, if not all, are more secure in the Oracle Cloud. Therefore, it's starting to become a push to the move to our cloud for our SaaS-based applications. That's what we mean by complete data-driven, personalized, connected, and secure. All of those driving tenets apply to all the product pillars. Now let me get into the individual product pillars and detail some of the new investments and new product features that we've had.
Some of this we'll explain when Mark Hurd in the past, and I think on the call, talked about how we're expanding into new markets. I'll get into a little bit of what that means specifically with the applications investment, and then I'll get into how we're seeing the most common customer use cases and adoptions within each of the product pillars, and product families. The first one I'm going to start with is our CX Cloud. Just as a reminder, CX or customer experience is inclusive of marketing, sales, CPQ or configure price quote, basically a quote and order process for your sales, e-commerce, a cloud-based e-commerce platform, the service platform and service module, and social, and really social and data, which underpins all of the CX applications, which again, I'll talk about more in just a moment.
Huge investment that we made in the verticals because most of these applications were pretty well built out versus either what we've had in the cloud in the past, frankly, well beyond capability that Siebel or any other on-premise system used to have, and beyond what our largest competitors in CX Cloud had. What you've seen over the last year is increased investments in the more and more industry-specific functionality in automotive, telecommunications, financial services, high technology, CPG and retail, consumer type of applications, really geared towards having customers in those industries having now best-in-class capability within service, sales, marketing, et cetera. This is a sampling, which at this stage, as I mentioned before, this conference is really a celebration of our customers, really a who's who of any of the best brands in the world using our CX solutions.
Some of the most common adoption patterns, I sort of alluded to this a little bit at the beginning. In CX, more often than not, customers start CX with one particular product pillar. Many times, they already have a Siebel or some other legacy back-end system doing their customer management or order management. The number one area we're seeing uptake is in marketing, and the single biggest driver is marketing today is much different than marketing was even two or three years ago. There's multi-channel marketing, the need to market to customers not only in the traditional fashion but in email, in social, in mobile. There's digital marketing along that, the ability to personalize that marketing, not just to broad segments of the audience, but to very specific segments of the audience. I'll detail that more in a moment.
The other very popular case, we're seeing returns, for example, like Micro Focus, with 200% increases in their sales pipeline by using digital multi-channel targeted marketing from the Oracle Marketing Cloud. The other very popular one is the Service Cloud in the lower right. Once again, service, the dynamic is changing, the push is really from a consumer-driven push for service. Service used to be call centers, which again, many of our existing customers have had, but it's migrating very quickly to now call centers plus what's called knowledge management. If any of you have called up on the phone for service and they've directed you, "Please go to our website and you can find the following answers," that's usually driven by our knowledge management application, helping you answer it.
More and more service is not only phone, it's not only knowledge management on-premise web, but is now digital and social. Twitter, if you pick just about any airline, any cable company, any consumer, more and more of their traffic, their call centers are becoming a combination of call centers, traditional, and listening centers, meaning monitoring social activity and responding in that social medium, Twitter, et cetera. More and more these days, the next channel is through chat or SMS. A modern service application has not only the traditional channels, but all of the new channels. Because of the dynamic nature of those channels and that media form, to do this outside of a cloud-based environment where you're able to consume updates every six months is really impossible in an on-premise model.
Again, if you think back to an on-premise type of service, and you need to consume changes every 6 months because there's a new channel, really impossible for you to do because you can't even get your upgrades done more than once every 2 years at best. Again, as a traditional model. We've seen tremendous traction in our Sales Cloud for from Salesforce automation, territory planning, incentive compensation, and tremendous progress in our brand new introduced Commerce Cloud, which we just introduced last year for online e-commerce. I'm going to get into more in the commerce and sales in just a moment with some of our new apps.
Now, for the first time, you'll see on the panel, as customers have adopted one part of CX, we're starting to see them adopt a second, the third, the fourth, until they get a complete replacement of their on-premise CRM systems. The second product module we'll talk about is HR, which again, as a reminder, includes global HR, talent management, workforce management, including payroll benefits. But also some brand-new innovative applications we call Work Life Applications, which significantly differentiate us from our competition. If you just deal in HR and you just deal in U.S. HR, there's a sort of a limited amount of functionality.
However, if you deal with multinational and large corporations, if you deal with companies who have at scale, like L Brands, the parent company of Limited, who has to run payroll for over 80,000 retail employees with sophisticated retail for payroll requirements, that's where we significantly differentiate from competition. Also in what we call Work Life Applications. We've introduced two that are unique and differentiated from any competitor. The first is what we call employee wellness. It's an add-on to your benefits module to track and encourage employee activity, and/or have contests or gamifications around employees, integrates with Fitbit, so you can drive employees for better employee health, and internal competition and promotion of that. The second is for corporate giving application. At Oracle, for example, we have a corporate giving program where if we donate to certain charities, we get a matching program.
Corporate giving tied to your HR system allows you to register those, allows you to submit and/or request a new charity to be part of that, and allows you to organize broad events. Again, at Oracle this month or next month, we're going to host a Heart Walk for the American Heart Association, where the entire campus gets together and does this walk. It's an internal promotion. We allow you now to register that in your HR system, collect teams of people, get sponsorships for it, corporate giving. All of it's driving HR adoption. All of it's extremely popular, especially as the global workforce and the demographics of the global workforce change to give you a more comprehensive look at HR. None of the solutions come from even the new HR competitors that are in the cloud.
What you see in HR is large and small companies, big brand companies, U.S.-Pacific or North America-Pacific, and then many global companies. Just top of my head, running our core HR of over 70,000, 80,000 employees. Of course, Oracle ourselves with over 100,000 employees, UBS, British Telecom. You'll hear from AXA on the panel in just a moment. Siemens North America, L Brands, Macy's, Kaiser Permanente, to name a few in our cloud HCM. The HCM adoptions are starting to become. They used to be much more a start with talent management, either recruiting or performance management, or compensation, and that is still some of what we're having. CSX has increased their cycle time and hires by using our social recruiting by 38%.
Much more common today, as both the breadth of the cloud has solidified and customer confidence in cloud has solidified, we're seeing more and more customers go completely to HR in the cloud in one full swoop. Taking PeopleSoft, Core HR, Payroll, Talent Management, Recruiting, and moving that entirely to our HR cloud. You have companies like Pearson who's seen three months payback by moving their entire HR system to the cloud. The next area to cover and our fastest growing area now is our ERP cloud, financials, risk management, project portfolio, and procurement. We have over 2,800 customers in the cloud with over 400 go-lives just in the last six months alone. We're seeing speed and we're seeing depth and breadth across the financials. The 77 countries, that's 77 countries for which we have prepackaged legislative localizations to operate on behalf of our global companies.
That does not mean we only operate in those 77, because you can configure the system and write reports to comply for every country in the world. Those are 77 countries where we've pre-delivered it, where our customers don't need to do anything to run in terms of those global localizations. The reason we were able to get there so quickly is not only the investment we have in the software, but the experts who live around the world who understand the nuances of accounting and tax, and warehouse management rules and documents you need in all of these different countries have built these applications before. The same experts who built them for the E-Business Suite, for JD Edwards, for PeopleSoft, who understand not only today's, but the ongoing investment.
We have many people from our accounting team who are actually on government boards like IFRS, who help define and shape what the new rules and regulations are. That's what helps us stay compliant, all up-to-date, and in a broad basis, by far the largest, most complete financial set in the world. Here's where you'll probably see the biggest change. If you talked to us last year, we were seeing much more smaller medium business and/or companies who were starting to move pieces. Over the last year, we have more and more large enterprises like HSBC, and you'll hear from Orange in just a moment, who are moving their financials core to our cloud-based applications, not just the smaller and not just the acquisitions or different lines of business. In financials, really a few types of adoption methods that we see.
One is sort of what used to be the E-Business Suite sweet spot, no pun intended. The E-Business Suite sweet spot, these just now pre-IPO are now post-IPO company moving to their first financial system. Those are now moving directly to the cloud. People like Pandora Music, people like Boingo Wireless, people like LendingClub, starting in our cloud-based financials. Second, you're seeing companies who are large, like GE, moving either lines of business or lines of business and geography over. GE's moved GE HealthCare, starting first in Canada and Latin America, and they've now migrated since last year to over a dozen countries across Europe live with ERP. They also live with GE Digital, and live with GE Digital not only for the newly formed GE Digital but for any acquired company, going forward.
The third flavor is having companies do more wholesale migration to the cloud, large or small, across their financial fleet. Last year, we introduced our supply chain manufacturing products from planning, core manufacturing, order management, logistics, PLM, and procurement. Just in the last year, we've had over 640 customers to our manufacturing supply chain cloud. Again, you'll hear from one of them today. We're just starting to have the go-lives. The go-live, given that we introduced the application last year at this conference, and actually this year, it's earlier in the year, so it's within less than a year, is quite remarkable compared to the pace that you're used to seeing in a manufacturing type of product. Over $10 billion managed on their freight in our cloud, over $18 million, $10 billion, pardon me, and 18 million product management and large manufacturing customers.
The manufacturing adoptions, you're seeing really a sort of a surround strategy, companies moving logistics first, or customers usually moving PLM, their product lifecycle management or innovation management, or as an example you'll hear about today, a complete supply chain manufacturing move to the cloud. The final category I want to talk about is our Data Cloud, which many of you probably aren't as familiar with, a newer offering to the cloud, which is really a collection of investments and acquisitions we've made and organic investment. So formerly BlueKai, Datalogix, AddThis, and Crosswise. In that collection, we have through collections of data partners, cookies, mobile device information, and then connections we are able to make with the aggregation of all that data, we have over 5 billion cross-channel profiles.
A cross-channel profile being defined as an activity that you may have done on your mobile device, on your laptop, on your phone, and on that profile, we have information about purchases you may have made, things you may have researched, demographic information, and this is what today we use to drive our digital marketing. So I said in the marketing, there's digital marketing that's much, much different, and all of you have probably been personal examples of this, and frankly, you've already been personal examples of this over the last two days.
If you booked your trip to come to San Francisco for this conference on your laptop, almost immediately thereafter, you probably started to see on your Facebook page, on any other page you went to, or on your Google page or Google Plus, or on your mobile device or on your Twitter stream, advertisements for activities in San Francisco, hotels in San Francisco, events in San Francisco, rental cars in San Francisco. If you also happen to be a loyalty member at, say, Hilton or Hyatt or Marriott or pick any, I don't want to insult any of our customers or partner hotel chains. A loyalty card member at one of those, you probably also received an email from that hotel chain with offers you may have had for San Francisco.
Or if you're a loyalty card member at Avis or a rental car capability, an email with that going forward. That is very, very likely you being one of the 5 billion profiles that we have, collecting that data, doing a data mapping, which allows us to, with high level of certainty, understand that that profile is this email address, is this Twitter handle. Combine that with our customers, who's really a who's who of corporate brands. Their first-party data, so all we know is the anonymous activities you're doing. We don't know your name. We don't know your email. We know what activity you're doing in an anonymous source. However, we can combine that and connect that with our customer's first-party data to allow them to do targeted marketing.
We have by far the largest Data Cloud in the world, and combine this both with ours and third-party marketing to drive real-time digital marketing today. Announcing at this conference is we want to extend that into a whole host of applications we call Adaptive Intelligent Applications, and make Adaptive Intelligent Applications available across CX, HCM, supply chain, and ERP. What is it? I guess my simplest example, I'm going to give you on-premise to cloud, now to data-driven applications. I'm going to give it in an example that all of you are probably familiar with. If you came to this conference 10 years ago, and you needed to work your way around San Francisco, and you weren't familiar with it, you probably had a map.
I'm not sure how many people remember those, but I'm talking about the old-fashioned paper maps, and that's what you did. You bought it, you owned it. If something changed, you didn't get updates. If you needed to have an update, you needed to go buy another one. That's more or less your on-premise applications going forward. And maybe you customize it, meaning you scribbled out street names and things on your own. Didn't work on your new one because you had to redo that customization. Now, a few years after that, you got Google Maps on your mobile device or some other way. It was great. It updated itself. It was always current or reasonably always current. It had new UIs. It worked on mobile devices. That's our SaaS-based applications, much the way I just described.
You can personalize it, put your favorite locations, your home, your work address, et cetera. All of ours were different. Upgrades came, it was great. That's more or less SaaS-based application. Taking data-driven applications and adaptive intelligence is now like Waze, which Waze has all those same characteristics, except it has data, not just your data from, I'll say everybody in the world, obviously not everybody, data from all data is available out there on the web, and aggregates that and uses it to make better decisions on where you should go and how you should get there. In addition to that, based on feedback, it improves those algorithms. That exact same concept, we have the applications, in this case, a commerce application.
We have the data, what we've applied with adaptive intelligence is we've written the machine learning algorithms, in this case, for a next best offer to surface the offer in our commerce application, see how effective that offer is, then improve that algorithm going forward. What with the example here is meet Taylor. We have demographic information, 32, lives in San Francisco, coffee drinker. She likes hiking in the outdoors, maybe we know where she lives. What you would typically see today is, okay, I've got that information, I'm going to put some offers together. Buy six, get one free, some prediction on what a person of that sort of demographic information might do and have those offers. What we're going to do with adaptive intelligence is take that.
Let's take everything else that we know is going on with the data that we collect. We know it's been hot lately the last few days. We probably know she's been traveling because she clicked on that website to go forward, we know where she is. Maybe she lives in a different place, went to Southern California. It's even hotter. We're going to adjust and tune those algorithms to offer an iced coffee with a different upsell. It's not static information, it's updated information updated real time across all the data. We're going to take that because we have our data assets, we have hundreds of data scientists who've been doing this analysis on behalf of those name brands in marketing for years already.
By adaptive, we mean application to surface the recommendation, data both internal from our customers and external from our Oracle Data Cloud, data scientists who have not only expertise in the data, but domain expertise in CRM, supply chain for transportation planning, ERP for next best discount offer taking, HR for finding the best fit and the most available hire person, drive Adaptive Intelligent Apps that learn and get better over time. Significantly different from our competitors who don't have the external data that we have to supplement the data don't have the years and hundreds of data scientists to analyze the data, in most cases, don't have the breadth of the functional experts. The same functional experts who helped us drive the 77 enhancements and localizations, apply them to applying digital intelligence across the applications.
Significant ongoing differentiator for us in SaaS applications, which we just launched at this conference. We have a list of most of the examples I just talked through, adaptive intelligence to improve your offers, for supply chain to build stronger portfolio, for HR for better candidate fit, et cetera. With that, I don't think we're running a video, right? Please welcome our customers on stage to have you hear directly firsthand from SaaS customers about our applications. Just a note for those of you on the webcast, we'll be breaking shortly, and please join us here at 12:30. We look forward to speaking to you then.
Ladies and gentlemen, please welcome to the stage Oracle Executive Vice President, Doug Kehring.
Honestly, I'm still surprised that they're letting me up here. I've been at Oracle for 16 years, and not once have they allowed me to come up here. Now, it might be because I run the mergers and acquisitions program, but sometimes I'm not sure. Actually, the reason I'm up here is a little over a year ago, I expanded my role to include stewarding our cloud transformation from an internal operations standpoint. What I thought I would do today is give you a little bit of insight into what we're doing across the organization, and a little bit about what the impact has been for customers. The safe harbor language as usual. Everyone knows this.
We've spent all morning talking about technology, and it comes as no shock that I think we believe now we have the broadest and best and most complete set of cloud technologies in the industry. It's been more than that. It's way beyond engineering. It's really required a transformation across our entire company. If you go back and look at the series of technology transformations as we move from mainframe to client-server, client-server to the internet architecture, while the technology's changed, the business model underlying it didn't change. The reality is what our sales force sold, how they got commissioned, what the contracts look like, how the support model worked, it was still the same thing. It was just wrapped in a different type of technology.
As the cloud came along, that's all obviously changed dramatically, where as a company, we're moving from delivering products to delivering services. When we think about the support model, we're going from just a break-fix to one in which we have a continuous engagement, as you heard from a lot of the customers that were up here on the panel. The way I look at it is we really moved from being an antagonistic type of relationship with the customer to being a strategic partnership type of arrangement. When I say antagonistic, the notion is, as we went through, and once you'd selected the technology, all of a sudden we were at odds with each other. The whole contracting practice was about trying to get the best terms possible for each of our companies as we thought about the sale of this multimillion-dollar type of license.
When it got, again, to the support notion, we oftentimes had no idea what the customer had done with the technology. When you think about it, we sell a perpetual license. The customer takes the software, goes back to their internal operations of the company, and they connect it to who knows what. Right? It's the different hardware, different storage, different servers, different implementer, and oftentimes they actually change the software itself. When they had a problem, they'd call us and be like, "Hey, something's wrong with Oracle's technology. Fix it." Of course, our reaction is, "Well, what the hell did you do to it?" That got to this notion of you got to recreate the problem. As a result, it's really hard to be in a great partnership. Versus now, with the cloud, it's all on us, right?
To the degree that something goes wrong and they call us Hopefully, we already proactively called them and said, "Hey, something came up, we fixed it for you." The reality is we're running it all on their behalf. What that's really meant for us as a company is that we've had to change our whole way of thinking, really from a cultural standpoint, starting with Larry, Safra, Mark, all the way down through the organization, to really move from a risk management type culture to one in which it's all about a service-oriented culture.
As I like to say, the old Oracle is the just say no policy, which is as I went around internally to all the organizations, a lot of the response to change management was, "No." We're like, "Well, why don't we change?" "Well, that's not how we do it here." I said, "I know that's not how we do it, but why don't we think about doing it differently?" It was the old idea that nobody got fired at Oracle for just saying no, if you're in finance or legal, et cetera. Versus now our notion and orientation is this idea of saying, even if the answer is no, like a genius bar, if you will, within Oracle, from an internal organizational standpoint, the answer has to be, how can I help make it right for you and for your customer?
We've got a real different philosophy going on from that standpoint. This thing carries out across the organization. When I thought about it as we went and had these discussions about what does it mean to be a cloud provider, the first thing I'd tell you is that as a result of doing all of these acquisitions, every cloud company that we've purchased along the way had a different business model. Whether it's the customer success manager, you got paid as a hunter versus a farmer, all of these intricate notions of how should we run were all different in the cloud. That's very different from when I was doing integrations for on-premise type companies, because in the on-premise world, the business models were very similar to each other.
The reality, it was as easy for us to just move this company in and run it at Oracle the way we've always run Oracle. In the cloud, that wasn't the case. We had all kinds of different business models to think about. We took the opportunity to step back and say, how do we think the business model should operate? Beyond, again, the technology, now we get to the business model. How do we want that to be? We first said that we want to change how we think about engaging with and caring for our customers. It fell along three main pillars. How do we market to them? How do we sell to them? How do we service them? Internally, it was around this notion of what are the organizations that need to be in place?
What are the processes that need to be in place? What are the systems that need to be in place? We used some nice keywords, if you will, that have helped everybody internally really understand what are we talking about. We use the connected customer experience on the marketing side. That's a notion, especially around using Eloqua and a lot of our marketing cloud services, where the idea of tracking business customers throughout their journey and life cycle. When they touch oracle.com, what are they looking for? What is their profile using the Oracle Data Cloud? Is this an HR person versus is this an IT person? That then we can tailor the content and tailor the information and start to understand their body language to get to the point that we know when we should engage them in the selling experience.
For me, the way I've always looked at cloud versus on-premise is in the old world, the first call you made was to the sales rep. In the new world, it's often the last call you make. We've got to do, and a part of that is doing a very good job all on that front end from the contact of the opportunity. We've identified them, and this is a buying opportunity. We looked at this, and this is going to be what I'm going to spend more of my time on today, is around then what's the buying experience look like?
The history, again, at Oracle is one of those things where a lot of customers come to us and say, "You've got the best technology, but man, you're hard to do business with." That sort of philosophy, when I go back to being in a service-oriented culture, that's not what we want. Along those lines is how do we now sell to you to make it very quick and easy to purchase from Oracle. Finally, the superior ownership experience. When we think about, okay, now that you've purchased the cloud service, we've got to be on top of you understanding the deployment, the upgrades, the downtime, planned or unplanned, the statistics about your usage and adoption.
The entire notion of trying to drive their success in the deployment of that cloud service, really, again, get to the point where, as we saw from these panelists, every one of our customers is an advocate. Gets up there and holds the mantra for why Oracle's great. In fact, on this service side of it, around the superior ownership experience, again, the notion where they're engaging with our development teams, our support teams, that sense of bonding that's going on with our customers is really where we're focused. Now the beauty is this is also a way for us to then leverage our own cloud services. Everything that we've built around this entire journey and this life cycle utilizes our cloud services in some way.
It's also been amazing, and again, you heard it from the customers, that our ability to uptake more and more features as we think about how to craft this journey can be done faster because each new upgrade of the software enables us to take advantage of additional features that we need, versus you're stuck in this world where it's going to be years before you get to a new release of the software. As we get through here, I'll talk a little bit more about how we've leveraged some of these cloud services. I mentioned that I will go through the accelerated buying experience. In a nutshell, the notion is to make the ordering process at Oracle flexible, friendly, easy, fast, and with lots of help.
I'm going to go through each of these things in a second, our target, after we spent a lot of time analyzing all of the historical cloud transactions that we had done at Oracle, to look at if we change this, if we changed the contract, if we changed how people buy, if we changed how our internal organizations work, by using what we ended up with customers based upon actual transactions completed, we realized that this entire process could easily satisfy 90% of all of the cloud transactions at Oracle, that's been our target. What does it mean to buy Oracle Cloud services through the accelerated buying experience? It really is as easy as one, two, three.
The notion is that the sales rep, through our CPQ Cloud, is working with our customer, and oftentimes, in many cases, the customer just doing it on the Oracle Store themselves without intervention with a rep. It's just determining their cloud needs. What would you like to order? Oftentimes just crafting what the ordering is going to be in terms of which cloud services, how many users, that sort of thing. That gets all done in CPQ. The rep has all of it right in front of them. They click all the different options they need from that perspective in conjunction with discussing with the customer. Then it's done, and the order gets sent to the customer through a landing page. They open it up, look at it.
Hey, is this what I want to order?" Verify that's the case, and all they have to do is click to accept, and they're on their way to deploying the cloud service. Okay. I'm going to go through a little bit more of what that means in a second. Again, I want you to think about simple, fast. This is the type of transaction processing that we want to do because it's what customers want to do with us, and obviously, it's an advantage for us because it dramatically simplifies the ordering process internally. When we say about revamping the ordering process, it started with the terms, right. When we get back to that, how has that antagonistic relationship worked in the past?
The notion is the buyer decides what they want, the seller, our sales rep verifies what they want to sell to them, then historically, we turn it over to lawyers and to contract people. Months would go by while we beat the crap out of each other because some lawyer somewhere wants to prove that they can get the best term possible against our customer who happens to be the person we're trying to sell to, right. That's their job. That's their entire role. You get done, and we finally get that in place, and the taste left in the customer's mouth is, "God, that sucked." Right. Which is the opposite of what we want to have happening.
What we did is we looked at all these contracts, again, in the cloud space, and after analyzing, we said, listen, a lot of this stuff, it's all belts and suspenders. It's to take care of the most obscure possible outcome, right. Which just meant, again, another opportunity for everyone to sit around and try to figure out how to handle that obscure outcome. Reality is, as Larry would tell us when we were architecting this, he's like, "How many times are we really suing our customers and them suing us over the terms of the contract." It just doesn't happen. It doesn't happen. Mainly it doesn't happen because it's a business relationship, and we work it out anyway.
We spent, again, months with this entire process to take care of the most obscure possibilities that are in there for an outcome in which we'll never use, right? That only leaves customers with a horrible taste in their mouth. We sort of said, okay, take off the belts and suspenders. Let's move to taking a risk acceptance posture in these contracts because we already know we've completed contracts with these terms, so we know we accept this, right? We're okay with these type of terms. Why go through the process of trying to just get a little bit extra from the customer? Again, through that process, it dramatically shortened our document, and it made it dramatically easier for the customer to understand and read and look at it from more from a plain English standpoint.
We had this idea of how do we create options for flexibility where a customer may need a little something different in each contract. Again, looking back through all the contracts we had, we've tried to figure out what are all the options that everyone has been looking for. Then we had our IT team put all that stuff into CPQ Cloud. That, again, when the salesperson's in there, they're empowered to take a lot of the stuff themselves without any further approvals. They can select on all the options. There's a lot of sales rep-empowered options. Then there's a notion of a bunch of other options where it might need an approval. Okay? In each case, they can select it, and after the requisite approvals are received, the contract and the terms are automatically generated.
We no longer have to have this big group of deal management people who are simply rewriting a contract each and every time. Lots of manual labor, lots of extra time in the process, versus now it automatically can be spit out, which means now we have rapid ordering. Once that's all done, gets sent over to the customer, as we said. They can open it up. They can verify it. The ordering page is one to two pages long. It's just validating what's there. Of course, the lawyers are welcome to read it. The associated terms and conditions are in a link there. It's not like we're hiding anything, but it gives them the opportunity to validate what they care about. Then move on with, again, that rapid process, just clicking to accept. I mentioned the extensive support.
Extensive support is one of those notions where we took all these people that were just doing manual tasks with no actual added value, we reoriented to the process of being service first, right? We started with, in the actual cloud service applications, let's put more information right at the fingertips because a salesperson, at the end of the day, should be really good at understanding our products and services, understanding what the customer needs are, and connecting the two together. That's really it. They shouldn't be experts around limits of liability and obscure contracting terms. That's really what they had become, because the better you were at that, the more likely you were to get your contract through faster, which is totally ironic in terms of where you'd want to spend your energy. In this case, it actually has a journey right there in it.
The information's available. It pops up to help them be able to complete the order quickly. There's a knowledge base, which is, hey, maybe in some instances, I need to have a little bit more detail about some of these things because I do have to explain it to my customer at the time. Huge knowledge repository that's available to them at the click of a button. Finally, there are instances in which the customer says, "Hey, I got you, but I'd like to get on the phone with somebody who is an expert in this stuff." We took a lot of those resources that were doing, again, this manual, repetitive behavior, and we redeployed them in the service mentality so that they're now there and available to engage with our salespeople and our customers to give them more of the hands-on help.
Again, lots of support that then comes with this. In terms of the rollout, it's been amazing. Q3 is really when we went through and said, "Okay, let's deploy this," and we launched the project. In Q4, we already had the application changed and up and running. Part of that, we had revamped and simplified the ordering process, and we did some training. Now it's Q4, not a great quarter for us to do a lot of training. We'd like to close more business, as I'm sure all of our investors would agree. We held off on doing too much until Q1, where, again, we used another opportunity to add some more of these capabilities, and then we did a full-on training with our usual beginning of the year training processes.
Already through two quarters, two full quarters of closed transactions, we've gotten two-thirds of these cloud contracts to go through the accelerated buying experience. When I say accelerated buying experience, I mean, this is where there is no manual intervention. Okay? There's no lawyers that are going off and writing obscure terms and things like that. This is all online through the process I just described, with the target of getting to 90%. Now, how will we do it? We'll continue to look at improvements, ways to tweak things, et cetera, as well as education of not only our internal folks, but also our customers. They also need to be educated about how to rethink buying of enterprise technologies. The other thing we're working on is moving this to non-cloud transactions.
We're looking at this for our hardware business, for our license business, et cetera, because, again, we see huge amounts of value and benefit for both parties through this process. These are some quotes that have come from our field, from our sales teams. Again, these are words we're not so familiar with in the support organizations at Oracle. Things like, "That was awesome", "That was very efficient", "That was very effective" These are great things. The sales teams internally love this. They love it because it helps them do their job better. It gives them more opportunity to go out and sell more. Again, in the service mentality, we're spending a lot of time doing interviews. We do case studies with our sales teams to give to other sales teams. We do best practices around what have we learned.
They help us tweak the model, et cetera. Again, it's a much more engaged process. Our customers are ecstatic. Here's two quotes that I put up here, and I'm going to roll a video in just one second. This notion of being delighted with buying Oracle technology. Honest, I've been here, again, 16 years. I don't remember anyone using that word to describe Oracle buying processes. Anyway, that's the type of transformation. Let's roll. Oh, sorry. Let me go back. Let's roll the video now just to give you a little sense from the customers directly. Video? Maybe. Are we in charge of the video?
Slide problem.
Slide problem? Okay. I'll just talk a little bit more before the video because I think it's one of the most powerful things we have. Let me just summarize what the benefits of that buying experience is. For our customers, it's that notion of being simple, fast, rapid access to the cloud technologies. Anything that stands in the way from the moment they decide they'd like to take Oracle Cloud services and the moment they're live and having their users on it, that's all wasted time for us, right? Anything we can do to eliminate that is good for the customer because they want to get access to those technologies, as is good to us. How many of you, as you think about your iPhone or your Android, and you said, "I'd really like to listen to"-- "I saw this Sting concert.
I'd love to listen to another song from Sting." The response was, "No problem. We'll have it available for you in three weeks."? Probably wouldn't be that satisfying. It's the same thing from an enterprise technology standpoint. For our sales teams, the reality is for them, this is why they're so ecstatic because a lot of these folks were spending more time on internal processes about how to get a deal through than external processes. It was almost like if you wanted to, you could have built a small business around how to help a salesperson get a Contract through within Oracle, and you probably make a lot of money. Now, all of a sudden, instead of doing all that, they're out selling more. It means they can make more money because they can go out and more opportunity to sell the cloud services.
For us, I think we look at it as a leapfrog from a competitive standpoint, which is at the end of the day, we're trying to not just evolve, we're trying to completely revolutionize how we, again, think about the business model, as we go through this cloud deployment process. With that, let's get the customers. Are we able to Q&A yet? Yeah. We don't have the video? No, we have technical difficulties. Okay. Oh my gosh. Now I'm so bummed. All right. Let me just, my last slide before our questions is the summary of what we're doing. This for me is my big statement, and this is what we do internally with all of the constituents to help them understand. We've got the best technologies in the cloud, in our belief.
Now we've got to marry it with being the best company to do business with in the cloud. I've been here all week, been with customers, talking to them about this, and the amount of enthusiasm with customers is overwhelming because they've all have been desiring this for so long. You open this up to them, and it just unleashes the ability that has held us back for a long time. With that being said, that gives you a little sense of what we're doing from an operational transformation standpoint. I think we got, you said a couple of minutes for questions. Knowing you're going to this point, we do have the video ready. Oh. We got it. Yeah. Thank you. Please roll the video. We'll do a couple questions.
As we went through the accelerated buying experience process, the entire process took a day and a half from start to finish. That was from the time that we decided to purchase the product and notify our Oracle rep till when we actually could log into Planning and Budgeting Cloud Service. I received an email from my Oracle rep. In that email, there was a link that allowed me to launch a browser that gave me access to the document. I clicked on the document link, reviewed it internally, we approved that document, we clicked the accept button on the same webpage, and we were owners of a Planning and Budgeting Cloud Service. I was very satisfied with the fact that, in very little bit of time, we were able to efficiently purchase the product. We also were able to use our company credit card.
My experience was fast and secure. It is a refreshing approach to purchasing software from Oracle. It's safe, secure, efficient, the way business should be done.
I think the accelerated buying experience was excellent. The best testimony you can give is that I barely remember it. That's how quick it was. I think it was, "Hey, I want to get a couple environments," get on a quick call with a sales and tech rep just to clarify what I need. They prepared the documentation and gave me a link, and when I went on, everything that we talked about was there, and all I had to hit was accept. Next thing you know, I have access to a full Oracle environment with operating system and everything that I need. I think the fact that it was so quick and easy that I don't even remember it probably speaks volumes.
Now you know why I was so desirous of making sure we got that video up there. It's just that powerful. Yeah, I'll take questions, if anybody has any, before we move on to Mark and Safra.
Hi, Doug. Ed Maguire from CLSA. Have you been able to quantify in any way, to what extent the accelerated buying process is actually reducing sales cycles and also, whether in terms of full-time worker hours, what savings you've been able to realize from processes that were previously manual?
Let me do the first part, and then you can retry the second part, because it got confusing for me. On the first part, the savings, yes. We've been tracking the notions of cycle times. The problem with some of these notions is it doesn't fully encapsulate all the things that we care about. The reality is we went back and looked at year previous and looked at, "Okay, what were the transaction cycles for this type of deal?" Then we looked at the transaction cycles for an accelerated buying experience. You have to remember, customers can get involved where they just go quiet for a bit where they're not actually doing the buying. We saw over a 50% improvement, and we see, again, it's dramatically more than that in other cases, when we think about it on an aggregated basis.
Yeah. The second part of my question really had to do with process improvements and use an example of billable hours for lawyers, for instance.
Yeah.
Maybe your customers, their lawyers are billing them less, or are you using fewer internal legal resources. To that extent, that may be playing out in other parts of the organization.
Right
If you've been able to measure some of those benefits.
Yeah. Safra can talk to a lot more. We specifically decided that what we would be focused on is trying to increase the customer service as opposed to do operational efficiencies from it at this point in time. We redeployed all the resources because the idea was, rather than be in a position where we cut resources from it as a result of this, let's over-deploy all these resources so that we are always looking to make sure that the customer and the sales teams right now have the most unbelievable experience possible. We're sort of taking that approach. Again, when you look at it from a customer standpoint and you look at customers like that, obviously, to be frank, the first contracts that we're doing are obviously starting at the lower ARR values.
We've already had $1 million ARR transactions through the accelerated buying experience, so it's floating upward. When you talk to like a Lyft, they would never have wanted to get a lawyer involved to begin with. From that standpoint, it's different. Versus if we're talking about GE or Siemens, the reality is they do have these gigantic organizations that are built around this, and that's why I said it is a customer education process. The buyer doesn't want any of that stuff. They love it because the more they can disintermediate their support organizations who only slow it down, the better for them. As we go through this more and more, we're going to be able to quantify a lot more of the efficiencies. Yeah, I don't. Okay. All right. Sorry about that.
I'll be available later, hopefully, to answer questions. With that, I'll thank you.
Thank you.
EOG will bring you up.
Mark Hurd. I'm just going to go. You want me to do it, Mark? All right. Do I need to No, we're going to bring all this stuff up. Hi. I brought my iPad up because as Doug was talking, I think actually this process is actually a little bit more involved. Part of what got this started was, you go to customers that have signed a deal, particularly with AWS, and what it would be interesting for you to do, and I won't do it here, I pulled the agreement up. I encourage you to pull it up and just read it. It's a quick read. It's several pages. I'll let you do your own interpretation.
Part of what Doug was describing that I think is very interesting is these agreements that Doug was describing that people don't like, they're much more advantageous to the customer than this. What really is done when you read it is at the bottom of it, there's a click. It doesn't say Mark Hurd sign, it just says click. I always ask the customer, "Who clicked? Because what you committed to was far different than anything you've ever negotiated with us." This is actually an accelerated buying experience with actually quite differentiated terms at the same time. If you get time, please take a look at it because I bet you it's a little bit different than you expect. All right. Ken, you've gone through the safe harbor statements. I don't have to read this?
No.
Excellent. You've gone through the non-GAAP financial measures?
Thoroughly.
Okay. I don't need to read this?
No.
Thank you. All right. Here's what I'm going to do. I'm going to talk about SaaS. I'm going to start talking about the current state of SaaS, where we are in the market, and I'll do that not just with SaaS, but actually the aggregated apps ecosystem. On-premise, both license and support. Secondly, I'm going to take you through, based on what we have now, what the opportunity we think is in SaaS, and I'm going to do that really by taking you through details in one specific pillar and the magnitude of what we think is in front of us if we execute. Then I will talk to you about platform. Just as an opening, we don't have nearly the amount of runtime in PaaS. Excuse me? Just read on the left, your left. Thank you very much. It's very helpful.
This is very complicated technology up here, and I'm going to try and catch up to it. The learning experience, it's been a long week. This is the most exciting crowd I've been in front of all week. I'm so thrilled to be here. In PaaS, I'm going to talk to you about the opportunity. We don't have the runtime on PaaS that we have on SaaS. I'm going to tell you more than I ever could have told you last year, but not near as much as I can on SaaS, given how much history we've got now, and how much data we now have on the application business. I'll talk about that. Okay? That's the plan. Somebody's still got a microphone on. Okay. This is SaaS. This is a view of SaaS bookings ARR. I think we actually skipped the chart.
Can I go back one? This again is Yeah, this is the revenue. 23% growth, and you can see the acceleration to 36% revenue growth to 42%. I think you've seen these numbers, but the important part of this, as our revenue has gotten bigger, our growth rate has accelerated. This is something we talk about a lot in calls, but I wanted to make sure you saw it as it related to the past several years. This is what's happened to us as it relates to bookings. $247 million, call it a quarter of a billion in FY 2013 to $861 million in FY 2016. I want to make sure I'm clear because I get several questions about what is a booking. There seems to be some confusion that some companies have different definitions of a booking. We count new logos and expansion.
The definition of expansion to us is a customer had HR and had compensation and benefits and then bought succession planning. That would be an expansion to us. If we were to renew the original compensation deal, that is not in these numbers. That is a renewal. We don't double count. Anything in our booking is a net new logo or a net new module to us. Everything else would fall into a renewal. This is what's happened to our deferred revenue. Obviously, 65% year-on-year growth in FY 2016. We had some dialogue over the past couple of years about how we entered the market, how we promoted, how that related, that we had deferred revenue and bookings and revenue that didn't align perfect. That's why we started giving these forward-looking projections on bookings originally, was to be a surrogate for what our future period revenue would be.
Most of that has flowed through the model now. If you ask the question, are there still promotions in our model in several products? The answer is yes. Are they material? Not really. Now we're in a pretty good position where the bookings, deferred revenue, and the revenue that shows up on the income statement will all roughly align. I'm going to switch now, I'm going to take you through a couple of different ways to look at our numbers. What this is transactional, is code for license. That's the license as a percent of a total. The red plus the gray equals the total. 68% of the total in FY 2013 was licensed. That's how you read the chart. The total of SaaS revenue and license in totality is growing.
We have crossed the model where we had apps declines in revenue that where we actually now have growth. I'll take you through this in another dimension in a second. In addition to that, the recurring revenue as a percent of the total is increasing dramatically. This is at the end of FY 2016. In FY 2017, we will go over 70% as a percent of the recurring total, and we will get aggregate growth yet again. I've now changed the metric for you. I've now included support on-prem, let me redefine this for our horizontal applications. It's not GBU applications, these are the horizontal apps, SaaS revenue, and our support. 76%, 24, moving to 85 recurring now. That includes support, and you can now see the arrow that brings you aggregate growth.
I believe that SaaS bookings this year, let me make sure I say this you get it clear, SaaS bookings will exceed transactional revenue as it would've been defined on the previous chart. This is a very important year for us from an apps metric perspective, because obviously, when we book a dollar in bookings, the next year, we get a dollar. If we were to book a license fee, the next year, we would get $0.20. We now have bookings, aggregate growth in bookings against a base that's already bigger with a much bigger percentage of our total that is recurring. This is our pipeline. Last year, actually it was two years ago. I told you two years ago, I prognosticated that we would have a lot of bookings.
I'm sure there was lots of, "I don't know if that's right," and based on the behavior of our pipeline. This is what's happened to our pipeline. Our pipeline in apps is not 2 billion, it's not I get a lot of coaching on what I can and can't say. It's big. Big. Two, three, big. Billions. It's come from nothing. I'm talking about nothing. Growing 62%. Let me explain how our pipeline works. Our pipeline scales into about mid-November. What I've done here is I've mix-matched a couple of numbers. The 85% and the 68% are full year comparisons of how the pipeline behaved against the previous full year. The 62% is as of 10 days ago.
Likely, this will, although I can't tell you for sure, likely, you will see that the full year will surpass this 70% range, would be my guess. What we have during the year is we have a decline in pipeline as bookings close towards the end of the year. We rebuild the pipeline as we go to the next year. This chart in isolation is a big deal. This is the current state in conclusion of where we are. Bookings, that's a 52% CAGR in ARR. You see our performance in non-GAAP revenue. The deferred obviously exceeds our bookings. Steve, I think, touched on our go-lives, about our go-lives. Our go-lives, this has been a big deal because you sell something, we got to get it live, and then when they go live, they become a reference.
That is one of the things, if you ask, other than fantastic leadership and all the other variables I'm sure that you think is what's driving this. The fact is we have great products and great references. That's at the core of what drives this. We have good products, I tell you they work. Customers like them. At a conference, I hope you all got a chance to see customers. Customers tell other customers, this sort of conference builds on itself. We have now, in all the things we've changed amongst our sales organization, we've also changed our marketing. I know many of you know this, we now have actually customer events discrete to each pillar. Safra hosts a Modern Finance event. What did we call it? Modern Finance, yeah. What we did this year is something pretty interesting.
Safra was hosting one, I was hosting HCM, they were basically right next to each other. HCM, we have thousands of customers now who come to an HCM event. We can do an HCM event. Now, we've actually changed this whole process, not just from a sales perspective, but now a market perspective. At those events, it's basically customers talking to customers. You get the echo effect of what's transpired. Pipeline, I've talked about as well. I'm going to shift now. I'm going to talk a little bit about an opportunity. I'm going to focus on one pillar, I'm going to focus on ERP. I'm going to tell you a little bit about the size of the opportunity. Now, I'm cautious to show you this chart because I had a lot of help with this chart.
I had help from lawyers, I had help from investment relations experts, all participated, as you can see, it resulted in lots of question marks and a lot of X's. I would recommend to you to listen to me as opposed to look at the chart, you can decide to do as you choose. We have something like 15,000 on-premise ERP, EPM customers. Okay? Something like 15,000. You know we have 2,800 ERP, EPM SaaS customers. We report it pretty much every quarter. We talk about the incremental quarter to quarter. You know what we've got. We have roughly 8,000 in our pipeline. That is not meant to be, we're going to close all these. This is a qualified part of our pipeline deduped for the on-prem customers. Okay? That would tell you that in aggregate, we know of 25,000 names.
The marketplace between the two leaders and on-prem together do not have 50% of the marketplace. What I would also tell you is, much as you think that there's a company in Germany and then us, and together the two are the big bullies of the ERP on-prem market, it doesn't add to more than 50% of the market. It's still a very fragmented market. Our pipeline is growing at something like triple in terms of the scale of people entering the pipe. This is the benefit of being ahead from a product perspective and now having thousands, or in the process of going into thousands of go-lives. We will get a lot of customers. You can insert, and that's why it says question mark, you can start doing your math, whether you think that's 30,000, 35,000, 40,000, 50,000, you can make up your own.
There's going to be a lot of customers. Let me tell you what's happened to us, and again, I've had a lot of help with these charts. Our current ARR, I'm going to use that term, and I'm going to switch to a term called ACV. ARR is annual recurring revenue. The average contract we take is roughly $300,000. You sign up, and you give us $300,000 in annual recurring revenue. Let me explain who the profile of the customers are. Many of those customers are mid-market, upper mid-market, mostly as a percent of the total financials. ACV is the chance for us to sell those same customers more. You have financials, now you buy budgeting and planning. You now buy supply chain management. You buy manufacturing. By the way, we have a whole slew of large logos.
You see them, but I'm not here to tell you we don't. I'm trying to tell you that's not the predominant part of the base. Many of the large logos that are in our pipeline and also in our existing on-prem support business. That ACV will go up materially from the ARR. I'm not here to tell you whether that's 50%, but I've thrown out a number. Okay. Could be north of that, could be south, but it's a number, and it's not immaterial or else obviously I wouldn't bring it up. You have to multiply that ACV times the customer count to get a total. That's how the model works. We know this model well. One of the reasons you hear us talk about ERP as much as we do, and by the way, make sure I'm clear, none of that includes HCM.
The connect rate between ERP. By the way, HCM does not drag ERP. It's not a phenomenon we've seen. ERP drags HCM, and the connect rate continues to increase, but I did not include that in any of the numbers I gave you, nor did I include PaaS, Platform as a Service, infrastructure, anything of that type. If that were to occur, the multiple on our support would be very large. Our support is nothing close to a multiple of those customer base times that ACV. This is sort of what conclude on the SaaS pipe. Strong growth in FY 2017. We will have a good bookings year. This is not meant to include any numbers other than organic growth plus the acquisitions that we closed in the early part of the year. Revenue will accelerate in FY 2017.
I'll stick on the pipeline, this bigger base in FY 2017. My guess is we'll be in this 70%+ pipeline growth rate as we enter the Q3, Q4 timeframe. Doesn't mean we'll grow revenue, just where the pipeline is. Obviously, you have to convert from there. That's where we are on SaaS. This current state and the mega opportunity we have if we can just execute in ERP. By the way, the drag effect, even though I didn't quantify it, we believe to be material if we can get that base as I've described. All right, I'm going to talk a little bit about PaaS. First, I'm going to start on database support. If you're trying to figure out if the bars are the same, and you're trying to get up here and measure them, if they look the same, it's because they are.
We just haven't had any change of consequence in renewal rates. The renewal rates on database are, frankly, virtually identical across the last four fiscal years. This is a view of new software license in database combined with Database as a Service revenue that would be a segment of the PaaS revenue. I've now extracted Database as a Service from PaaS. I've added it to the database license revenue, and you see this decline. You're going to see a chart in a second on bookings. If I were to replace that gray of revenue with bookings, you would see a different result. Because of the way we booked, whether it be metered or subscription and the mix of those two, we have a similar phenomenon to what you saw in SaaS. You have a bookings number.
The license doesn't flow through the P&L the way a booking flows through the P&L, because the bookings doesn't go through the P&L. It goes through ratably, and you get the revenue outage that you get. Our FY 2017, as we look at it today, shows overall growth in the combination of Database as a Service revenue added to new software license revenue in database. The mix now 11% recurring, 89% transactional. This is a prognostication. If customer moves faster from the bookings perspective, you could see a change in the mix. This is the bookings. You can see $500 million. You can see what would have happened. This number is material, but this was a booking. This booking is what's driving this year's revenue, this FY 2016 booking.
You can see what happens as it relates to each quarter. This is a look at FY 2015 compared to FY 2016, almost a quarter of a billion in Q4 bookings. Within it, we have seen a mix shift from metered to subscription. Customers buying more subscription and less metered. That mix is now roughly three-quarters subscription, one-quarter metered. That is almost a flip from 75/25 the other way when we started. This is a view of PaaS revenue that actually trails the bookings breakout. You really have to, again, this flows if you follow the flow, the bookings ahead of the revenue. Listen, this is the way customers want to buy. We have this issue of the mix, what becomes important, the mix of the bookings in relationship to the mix of the new software license.
Once you get to the other side of aggregate bookings growth, which is what we're now at, you get a better performance on the P&L. It's a number we've not given you before. This is average Platform as a Service ARR. We have a couple phenomena going on at the same time. This is a PaaS view of it, but I could have just as easily given you a SaaS view of it. Let me explain what's happening here. At a show like this week, we have released a blizzard of new services. What happens now is as our portfolio grows, our ability to now cross-sell grows. We have customers that were buying Database as a Service now also buy Exadata as a Service. They will add to the service suite as they go. That's point one.
Point two, when we tell you we release a service, we usually mean that we are releasing that service in the U.S. Same thing is true for SaaS. Over the course of a release or two, that service suite moves its way across the world. Just for an example, in ERP, we released financials, Steve, two back, and we just went to Brazil this last release. Two releases, and even though we would've said financials was available, we released then manufacturing supply chain in the U.S. Manufacturing supply chain is not available yet in Brazil. There's about a couple of release lag as you work your way.
In some ways, we have new services coming in, services that you might think are already available, but they work their way across the globe as we localize them and get them available data center by data center as we go across the planet. Does that make sense to you? That's some of the reason isn't just growth, it's also the fact there are more services available, and they get released in more countries. We get more services on one dimension and then more countries at the same time, and we're going through that process right now. This 57% increase, while a fantastic looking number, it's driven to some degree by both the dynamics that I've described. This is really the same view of pipeline. We have lots of nearlys and greater-thans in all of this. This is a $2 billion pipeline.
This has come from zero eight to nine quarters ago. This is a material amount of growth. The market is bigger. We have a stronger market position in this part of the business than we did in apps. Obviously, we've had this revenue you've seen as not coming off the backs of any material amount of support. As you can see, that's why I opened with the chart that I did. Strong growth in FY 2017. I've told you, I think all about this. I talked to you about deferred revenue. I talked to you about the pipeline. That's sort of what's happening to us as we speak. I will continue to give you more and more transparency on PaaS as I have it and feel like I can represent it with clarity.
The pipeline is $2 billion, clearly, the opportunity is materially bigger in pipe. Our pipeline in SaaS is bigger than our pipeline in PaaS. The opportunity in PaaS is materially bigger than the opportunity in SaaS. That said, this is what it's delivered. I know it's been a while getting from a booking to revenue, but this is how it's matured to the 80% that we've talked about for last quarter, 82%, 68% in Q4. I would tell you that with all of that said, we're still hitting our stride. We don't have all of our services yet available and certainly yet available in all markets. Okay. I think we're going to do Q&A now. Safra's going to come up, and we'll take any questions you got. Do you want to go up, or you want to stay? Sure.
Whichever you want. I think you pick them.
I hate to pick. Ken, you want to? This is such a big decision.
It's hard.
Sorry, Mark, I got picked. Right here. Brent and I were talking earlier today. This could be our 20th Oracle Analyst Day, right, Brent? 20 years, congrats, Brent.
What a great opening. Yeah.
I know. Nice. Yeah.
Is this the best Oracle Analyst Day?
Of course. Of course. With a $2 billion PaaS pipeline. We never had PaaS back then. The reason I brought it up was back in the late '90s, the questions that we used to ask was, as Oracle became a big player in ERP, how did you approach the sales specialization? Because the skills that you needed for a CRM sales rep were considerably different than somebody selling ERP. As your product breadth has gotten wider, we're tempted to ask the same question we asked back in the late '90s. How are you approaching this from a sales deployment methodology? How does it not complicate sales cycles so you can still continue to close a lot of SaaS and PaaS business? Thank you.
Sure. This really goes back to four to five years ago. We didn't start this this year, Keshav. We made the decision. Historically, we had sold a bag of apps to the CIO. The CIO had the control point of the architecture and the data center and all of those things that really put the CIO in the center of any decision. As certainly the cloud and SaaS became the norm, it democratized the decision-making process. Now you had chief marketing officers, chief sales officers that all, with that control point gone, had a lot more vote in what get bought. We made that decision way back then to change our sales force. To be very blunt with you, it was painful for us. We have a sales force that really, in aggregate, is organized by product, by buyer, by competitor.
We have a sales force that goes to market that sells sales automation into the sales organization, typically into sales operations, and competes with salesforce.com. We have a sales organization focused on chief marketing officers selling marketing automation, competing against Adobe and others. We have a sales force that goes to market in ERP. Inside the context of ERP question, our sales force sells ERP and EPM together. We have some pre-sales specialists who are expert in supply chain manufacturing, and they're brought in by that team. It is a specialized sales force, specialized in the three dimensions that I described, and that's not new. That's been in four years. Four years. Yeah. Well, we're getting better at it.
We also, I know you all know this, Mark has talked about it extensively, is we hired a lot of salespeople.
You some remember that, don't you?
Yeah.
Hired a lot of guys.
Right. We hired a lot of younger kids out of college to really change the entire engagement and to make sure we had the coverage. We started in the U.S. then spanned out across the world. I know you all keep track of our employee numbers, I'm sure you've seen that it has grown massively because we knew that we had to have more coverage for all of these product lines, because that's what follows after the years of development effort. Once we had the products or we had line of sight, we had to get set up to distribute them broadly. That's been a big effort. Though some of you ask about our real estate and why we're buying and all of that. We've completely changed our strategy and the way we've worked with our different sales teams and training.
We're moving to large campuses and coordinated training. You saw Doug up here before, talking about the buying experience and all of that. The flip side of it is the selling experience and the productivity you have to have when you have a lot of less experienced salespeople, but who know their product line very, very well. All of these pieces, I hope, are starting to draw a picture for you about what we've been doing under the covers in this massive operation. It pops its little head up in different areas like real estate or things like that, or employee numbers. This is the biggest change in our distribution of our company, is how much more sales distribution we have than four or five years ago.
It is true. What Safra says, it's a lot of points to this. We've bought big hubs. We're building big hubs. We're building a big hub in Austin, a big hub in Santa Monica. We've built a big one in Burlington. It really is different selling. Our cost per seat, if you will, is declining. We're putting these people in hubs. We've implemented a brand-new training organization, ground up. We didn't have one. We now have a sales training organization that we bring talented people from the college campus. We train them ground up. Their productivity doubles roughly every year. We get them in, we train them, they double their productivity every year. We put them in these fantastic hubs that Safra was describing, all with the most modern technology, so they can actually demo across. They don't just inside sell.
They demo, they travel a little bit, as opposed to the old model of the field rep who was always on the road. We still have a mix of both, but our primary go-to-market organization in that mid-market we've talked about, we've had so much success, is exactly what Safra's describing. Big change for us.
Hi. Great. Thank you. Mark and Safra, clearly, your cloud strategy seems to be working, you emphasized again that the absolute dollar growth in SaaS and PaaS is going to more than offset the decline in license revenue this year. With that being said, with gross margins steadily improving for cloud, how do you think about the potential for earnings growth for Oracle as you look past FY 2017?
We think it's going to expand dramatically. Our whole goal here in getting through this period of last year and now our move through this year, we believe that we will end up with significant earnings growth. That's what this is about. That's our job here, we appreciate that some of you have been bearing with us through last year and a little bit of this year. Our goal is significant earnings growth. I give really full credit to Larry Ellison. He has dragged our sales force and our whole company through this transition as quickly as possible, which has caused us to give up a little bit of earnings because of license.
As a result, there'll be a lot of earnings, because simultaneously, not only are SaaS and PaaS and now IaaS revenues going, but you also have a situation where our gross margins are starting to improve very dramatically. We told you last year that they were bottoming out last year, and that is turning out to be exactly true. I don't think anyone would have believed they would've been up so dramatically quarter-on-quarter this past quarter. We expect that to continue as we get to leverage our economies of scale, and we also get the revenues to match some pretty fairly massive investments we were making.
Hi, it's John DiFucci from Jefferies. Mark, you had a slide up there where you had, I think, horizontal apps for SaaS and you added license. Actually, I couldn't see the tiny footnote at the bottom, but Zach next to me, he's 20 years younger, so he could read it. It said it didn't include vertical apps. I'm just curious, just roughly even, how much is vertical apps relative to horizontal apps in the SaaS business? Then a follow-up.
Interesting question. Yeah, we never say.
Oh. Okay.
It's interesting. It's not been a really relevant metric because they're sort of in their own unique markets with their own unique behavior. When you get into retail and communications, they have sort of different dynamics than the horizontal market for a lot of reasons. It's. Yeah.
The vertical apps, the amount that is SaaS and PaaS, if that's what you're asking. They're moving in this direction, but they are behind where we are in HCM and ERP, and they're smaller anyway. The impact is, this is not what you should be really focusing on. I do expect it.
In the long run
They are further back behind the Fusion world.
Got it. They have fewer competitors, John. The reason we don't do it analytically is it's just apples and oranges. We'd be taking a combination of a retail market and a communications market. They have their dynamics for obvious reasons, less people go into those vertical markets. They have less competition. They're going to go through it, they're just going to go through it at a slower pace across a much lower base.
It's also different between the different verticals, like the Primavera folks have quite a bit of SaaS and PaaS, moving to SaaS and PaaS, while a couple of the others are further behind.
If I could just a follow-up to that, and I think it's probably.
Well, this is a two-part question.
Well, it's not eight like Keshav's, but sorry, Keshav. I think this one's for Safra, but it maybe Mark too. I realize it's small, but you've made a couple of acquisitions. I get this question a lot, so I'd like to hear you address it publicly. It happens to do with how you define ARR. One more clarification on that.
Sure. I haven't disclosed this number. I think it's three points-
Okay
out of the 82.
Okay.
When I say insignificant-
Yep. No.
I mean insignificant.
I appreciate that.
We would've had Let me say that, be clear. We would've had significant organic quarter-to-quarter growth without the acquisitions.
Understood. Understood. When you give an ARR, let's say you're buying another company that is bigger than some of the others-
Like who?
assuming it goes through. Like NetSuite. When you give that ARR number, you said renewals aren't included. When-
Renewals
some renewals of some of your acquisitions come through, they're new to you, but that's not included in the ARR, is it?
No, John. John, listen.
Okay. Sorry.
I want to make sure I'm clear on this. We don't game the system. There is no renewals from anybody in that ARR number. Okay? There was not an anticipation of something in that ARR number. There is no renewals in that number, nor the anticipation of somebody else's renewals in that number. Those are new. I tried to define what's new. It's a new logo. We've never had them before. An expansion is a net new module we didn't have before. That's what's called booking an ARR.
Safra and Mark, it's Brent Thill, UBS. You've outlined a very aggressive push to IaaS. That's a really big departure from what Larry had outlined, saying that you didn't want to go to that layer. You wanted to move up to the app stack. I'm just curious if you could outline what tipped the decision for you to make the move. Then I think we're all asking the question, given the core database business has such a good margin, can you preserve that in this world? Maybe outline a little bit of rationale why.
I would encourage you, by the way, to ask Larry this question because he is coming here next. You know that, right? I want to make sure we are really clear on what IaaS and why IaaS for our customers. Because there is the, what I would call the completely undifferentiated compute world. That was never our whole mission on its own. Just sort of rolling around in that world. However, as we started to create it and build it up to help our own customers, our SaaS customers, what we saw as they were deploying, and we expected, was that they would need PaaS for their other Oracle applications. What we hadn't foreseen, frankly, sufficiently, was how much straight compute they would want.
How much non-Oracle workloads they would actually be interested and may prefer to bring onto our cloud because of speed of light and just wanting to have everything together. Securing it together and reducing the handoffs and all of that. While we were really experiencing that, our ability to execute in this area technically, and it sounds like by your question that you heard Larry, maybe on Sunday and Tuesday, which I encourage all of you to do, as attended. It became clear that we could bring a totally differentiated offering that was significantly more competitive than frankly, probably we even realized we could do. As Larry talked about, I am sure, I hope one of you has the sense to ask him about Gen 2 in IaaS.
It is now a truly compelling offering. As we were benchmarking it, testing it, and working with our customers, we realized the demand is significantly larger, stand-alone for that service from Oracle. That we could be so competitive and also profitable that we decided to really sort of, sorry, Sheryl Sandberg, lean in to that offering even more so. We are really meeting our customer demands. Was that a two-part question? Was there a part for Mark on that?
No.
No?
The profitability.
Okay.
Sorry.
In the database profitability. That was your question. There we go. The database has always been very, very profitable for us because we are at scale. We have a large development team, and we have a very, very large installed base of customers. We have, for the 30-plus years of the Oracle Database, continued to invest in R&D. We've never sort of leaned back on our laurels. New features come out all the time. I'm sure you heard about the excitement around Release 2. That is a cloud release that has features that other companies have not even considered. The question is: Do we want to stop selling licenses and sell only a Platform as a Service? No, we want to do whatever our customers want. Our job is to fulfill their requirements and meet their needs and offer for them whatever they want.
The Platform as a Service in all the different levels, whether it's database or other parts, because we end up providing a much larger service than just a license, and because we can do it at scale, it can be very profitable for us. As Mark said, instead of just selling a license and then getting $0.20 a year, we sell a complete offering where the customer actually pays us more and yet spends considerably less him or herself. You heard some customers, it's just very, very compelling. For us, the issue is to be at scale. At scale, the marginal cost for us is not very large. As a result, we can be extremely profitable doing it, and yet benefit our customers in having them spend so much less than they spend on their own.
By the way, you see it in the numbers already, Brent. Look at the DBaaS. There's no chance those margins went from 39% to 62% with the DBaaS contribution to the total being either unprofitable or 10% or 12% or something like that. Again, as opposed to just what we say, because I know it's a question, look at what it says on the numbers. We've known this for a long time now. To Safra's point, it gets yet better as we get more scale. It's what we said from the beginning. We pull more revenue on the top, most of the cost stays relatively fixed. We have some scales. We go to new countries and new pods within existing SaaS places where we open up PaaS, so there'll be some more.
You see it perform exactly as we've described, and DBaaS is now a material part of the SaaS PaaS revenue that she's flowing through the P&L as we sit today.
Is that on? Hi, Kirk Materne with Evercore ISI. Along those lines, on Database as a Service, Mark, I was wondering if you could talk a little bit about the wins you're seeing in that. Are those new workloads that your customers are deploying? Are they expansions of existing workloads? I guess somewhat relatedly, when you think about the maintenance and the ability for support around database to keep growing at a modest level, I assume as long as renewal rates stay at the level that you showed earlier, that sort of low single-digit support growth, you expect that to continue.
Okay. There's a lot of questions in there. First, on DBaaS, most of the use case we see today still sits in dev test. Dev test is by far the fastest of the movement of the infrastructure to the cloud. It's low risk to the CIO. Dev test, on average, is probably 20% utilized in most of our customers. It's roughly 30% of their IT spend. The opportunity for pretty quick arbitrage from their budget is high, their risk is low. You see them move. Now, there still is, I would say, a weakness of standards in most of our customers about the standard dev test environment that they're moving to. I think you will see more and more rigor from IT around that, about how we go do dev test and what the standard is, but that is the primary use case.
That plus sort of the movement of some on-prem apps that are moving, and they run on some sort of Database as a Service as you go. To one more point, you didn't ask this. I do not see, when we're out there, Redshift or Aurora in competitive deals for any sort of production-type application that I see out there as well. That's what we see. The second part of your question was?
Just in terms of confidence, as you make this transition on Database, your confidence that the support revenues-
The support is a direct correlation to the license fee part of the business. It'll be about the transition. We have a very high connect rate, attach rate, as you would imagine, and That drives growth in our Database. As we get more subscribers, it's simply a subscriber model. We have very few subscribers leave, and we add subscribers every year, and it's just simple math. That'll be back to that point I meant earlier about the conversion of pipeline moving to bookings as opposed to transactional. It's simply a case of math. Right now, renewal rates are what I described.
Listen, this is Keith Weiss from Morgan Stanley. Thank you guys for hosting us. We've been running around Oracle OpenWorld talking to a lot of partners and definitely sensing a lot of excitement around Platform as a Service. One of the elements the partners were talking to us about was starting to build out vertical solutions on top of that platform. Can you talk to us a little bit about that ecosystem effect of, do you have ISVs built into the platform? Do you have other partners building the platform? Where are we in building out that ecosystem to start to get some flywheel effects going?
Yeah. We have a team dedicated that only does that. They do exactly what you described. They go from partners basically moving the Oracle, sort of non-Oracle apps, partner apps built on the Oracle architecture, moving that from today's on-prem to our cloud. You can go down ISV by ISV. We have a path to move, typically a part of the app, and then all the way to the full app suite over time.
Hi, Safra. Michael Turits from Raymond James. Hi, Mark.
Hi.
I think this is a follow-up to Brent's question on profitability, on the Infrastructure as a Service side. I know you've been asked this in past years, obviously, it's been more of a focus as a product area this year. What can you tell us about how you think the infrastructure margins evolve over time and impact the whole? Also the same on CapEx, which your efficiencies have kept very low so far.
Yes. Well, this is going to very much depend on how successful we are as far as how much capital expenditure we've got to use. We will spend more in capital on IaaS than we did in previous years. Depending on how much uptake there is, will affect it. What we'll have benefiting is the fact that we'll have some significant revenue coming from it. On the margin side, I think we'll be in good shape. I don't know. Was there a third question in there?
That No.
In the scope of profitability, obviously IaaS is less profitable than PaaS. However, it also kind of depends what's deployed, et cetera, and what's the scale. PaaS is more profitable for obvious reasons. I don't know. Did I answer all three questions?
We know what the Infrastructure as a Service margins are now.
Oh, no. They should improve dramatically. Oh, yeah. No, because you see, what you see in our Infrastructure as a Service margins now is dominated by our Gen 1, which is Oracle Managed Cloud Services. In IaaS Gen 2, what Larry spoke about and what will take over and basically dwarf the current IaaS numbers, is significantly more profitable than Gen 1. It's a fully automated, software-defined operation. It's very cookie cutter managed in a very specific, highly automated, and generally inexpensive way for us. As we scale there, the margins increase. It really does depend on what the mix is over this next year as we transition from really Gen 1 to Gen 2.
It might be interesting for you to know we have a team of people at Oracle who try very hard to convince us that Infrastructure as a Service is going to be just as profitable as PaaS. We sit here today telling you, we'll see.
We'll see.
By the way, that's what we tell them. We have a lot of technical reasons. Larry will probably get into them if you want to get him down that path. That tell us that can be the case. We're not OEMing a lot of part. Remember, we have a different model than everybody else that approaches this market because of our technical position. I will say on a spreadsheet, on a piece of paper, as you play this out, what they say is true. We want to see that in practice before we make any statements like that in a meeting like this.
That would be why I didn't say that to you.
Yeah. I just want you to get a feeling, though, that we're not going into this thinking this is a 2% business. Yeah.
Hi, Ross MacMillan. I'm at the back.
Ooh. Hi.
Hi. Ross MacMillan, RBC. Thanks for taking my question. Mark, you had a slide looking at the database business where you had both new software licenses and Database as a Service revenue.
Yes.
In 2016-
Yes
your license declined, and you started to get some growth in DBaaS. Then in 2017, it looked like you expected a flattening of that database license while still getting nice DBaaS incremental growth. That's a little bit different than what we saw with the application transition, where the licenses.
It's probably a bit the way the size of the base because the total is big when you look at it. There is some decline in the license number. The bigger phenomena to watch, we've got that factored in, is really the turning of the booking into revenue. What I was trying to say is that you could see an increase in more bookings in 2017, which could be a trade-off from license. What we see right now is what I showed. We would not mind taking that trade-off. If that occurs, we'll take it.
Is there a structural reason, though, I guess, was really my question as to why it might be a different transition on database relative to applications?
Sure. Yeah. Let me try to explain. In the apps market, most of our customers that have an app, let's say I've got an HR app, and it's an on-prem app, and I use it. I say I'm moving to the cloud. At some time after a transition period, that app will go away, now I have the subscription for the cloud. Most of our customers have what's called a ULA. They have an unlimited right to database. I may have 100 units of database that I've committed to over several years. I may move 2 units, 3 units, 4 units, but I still have 96 units left. I still run the bulk of my applications on-prem, you don't get the relationship of the purity of this goes from here to there.
You still have a premise based on a material size. That's where you see agreement to agreement. Most of our customers operate on a different type of agreement for database. It's a utility type of agreement as opposed to the application environment I described. It's different.
Walter Pritchard from Citi. I guess one of the things we hear frequently, we saw a lot of your customers that were cloud customers, existing customers, not necessarily new Oracle customers. I'm wondering, as you think about the PaaS and the Infrastructure as a Service adoption that you'll see, how are you thinking about your share of the new applications that are being built versus being able to migrate your base? I think one thing we see when we look at what's going on at Amazon, they obviously don't have an installed base. It's mostly capturing new customers. Trying to get a sense as to how successful you need to be with the new customers and the new apps to achieve the goals that you're talking about.
Walter, let me make sure I understand. You're talking about the apps market, app to app?
Yeah. I guess what I'm thinking is if I look at what's being built on Amazon, people are building IoT and they're building engagement apps, social, mobile, this and that. What we heard a lot of your customers in the PaaS and IaaS space, and what you see over in the exhibit being shown is more existing stuff migrating to the cloud. I'm wondering, do you think you need to get the kind of share that you've had historically of new apps? Or-
In the cloud
Can the new stuff go to Amazon? I know you're not going to say that's going to happen. Could the new stuff go to Amazon, you migrate what you have, and this is how it plays out?
Of course, the answer is we'd love to do both. In straightforward-
I'm just trying to figure out what you're thinking will happen.
We've got a huge workload to move. A huge workload to move. Today, if you look at our workload on-prem, we'd love to push a button and move it over, if that's your question. If that's big enough share. Certainly, if we just did that alone, our revenue would scale dramatically. We would gain enormous amounts of market share because we now do all the work. Just for our ecosystem alone. If we just move the Oracle ecosystem from how it sits today on-prem to how it would then look in the cloud, our revenue would go through the roof, and so would our gross profit dollars and our net income as well.
Interestingly, when you look even at our SaaS business, you'd say, "Well, you must just be moving over E-Business Suite customers, JD Edwards, and PeopleSoft customers." Some, what, half of our customers?
More than half.
More than half of our SaaS customers are new to Oracle. That should just tell you right there. This is in the app side alone. Even though we would be thrilled to just get our own installed base, which is a 35, 40-year-old base, all the way over. First of all, not everyone will move and all that. Let's say we could push that button. As Mark said, our business would be enormous. What's actually interesting is not only are those guys moving, they're also taking paths, which you even heard here or we've been at this for a week, as customers talk about what they're building next to it, all new stuff. Yet, as Mark mentioned, more than half of our customers are brand new to us. I think we're going to get it all.
Ultimately, it is just about being so compelling that they just go with Oracle. A number of customers, that's what they're talking about.
I'd have to bet three or four years ago, nobody thought that 60% of our application SaaS customers would be net new to Oracle. That the conversation would be just, you're going to move your base from here to there. By the way, we're still going to do that. It's we're going to move our base from here to there. That was the purpose of my (EUP) chart plus. You may say, "Well, that's not a surrogate for what's going to happen in the infrastructure or the platform market." We'll see. My point is, again, I've said it multiple times, I'll say it again. We have a better position in that part of the market, a stronger position than we did in the applications market. We have more experience, more user base, and to be blunt with you, more differentiation than we had in the applications market.
I think our applications team has done a fantastic job. That's how strong I believe our position is in this part of the market.
Brian Malinchak from Barclays. Mark, the last few years, we talked about the journey you guys were on, especially on SaaS now with get the product right, get the sales force right, get the sales commission right, get the reference customers. That was a big focus last year. The last leg of the story to me seems to be the SI market, those guys are always the guys that they come when they see the money and smell the money. What do you see in terms of that bigger ecosystem evolving? We saw KBACE getting acquired by Cognizant because they didn't have anything.
Right.
I think Deloitte is doing a good job, where do you see the others in their journey towards the Oracle apps? Thanks.
Well, I don't want to get into it SI by SI. I always felt like you're really popular with the SIs once you're popular. As a result, you have to do it sort of pillar by pillar, geography by geography. While you mention brand names, it's not really about just the brands. It's about the individual partners and the teams inside those SIs. I agree with you that the bigger SIs have probably been more attached to their legacy business than the smaller SIs. The smaller SIs that have invested early, they've done a fantastic job. You can imagine, we are extremely popular with the SIs and ERP. We get lots of people that want to meet and invest. Of course, now we're almost 3,000 customers deep into the process, a lot of the repeatable processes have been built.
There's a lot now we do as implementation and interfacing. A lot of those boutique SIs have built a lot of IP around that process as well. It's a little different SI by SI and a little different geography by geography. You mentioned one big SI, but they all have sort of, frankly, different stories. Right now for us, our SI portfolio has broadened, mainly with a lot of boutique SIs that have become really strong, really fast.
Am I? Okay. I have two questions, one for Mark and one for Safra. Mark, has PaaS reached the maturation level where it is selling more as a standalone sale against the database customers versus attached to SaaS, and how do you think that changes over time?
More is going to non-application customers than it is to application customers.
That trend obviously continues, you figure?
Obviously, the market is
Yeah
materially bigger in that part of the market than in our application customers. Remember, our application customers do use our PaaS, but remember, most of our application, as Safra said like two or three times, I've probably said four, most of our application customers are net new to Oracle and are just going live.
Okay.
They will get bigger over time as they want to add capabilities using our PaaS. Both markets will grow, but it's hard to argue with the enormity of the non-application side of the market.
Makes sense. Safra, knowing Database as a Service is much more revenue than the traditional on-premise database sale, as Mark discussed earlier, which will in turn drive more earnings. Understanding Database as a Service is still sub-scale, how should we think about the long-term margins that Database as a Service business?
I'm going to stand up even on my sprained ankle to tell you this.
Okay.
I think what will be amazing, because as you can imagine, we run these models like in 500 flavors of this. Profitability is one of those things that we are very focused on. Just because we're in the cloud and no one else in the cloud makes money, we don't agree with that concept. All of the old Oracle rules still apply here. We believe that in the PaaS business, we can be more profitable than we were in the database license business. Okay? I am not kidding. I know it seems impossible, and yet so many other things we've been saying that seemed impossible are playing out exactly as we said. Are margins in running a PaaS, especially on the database side, more profitable than the software business?
Thank you.
Take two more questions.
Bigger.
Two more questions.
Now I'm sitting down again.
Right. Great. Sarah Hindlian from Macquarie. Just a couple of questions. I'll be quick and painless. It looks like we have a couple difficult tax rate comps coming up, and we got dinged in the quarter on that. Safra, when does that start to normalize for us? Is this just a nice setup for 2018, or should we get a little better? What are you seeing in bookings to help with that?
Very fair question. Obviously, Q2, which is what my guidance was, it's very hard to match the tax rate we had last year in Q2. Many of you see it was 20 point-
Four.
That was an unusual tax rate for us. I don't know, maybe there will be another tax rate question, do you mind if I go a little long on the tax rate?
No
Tax rate and what's going on in particular? Okay. We tried to explain it on the call, but I'm afraid I didn't do a very good job. There are a couple things. There are two things going on related to tax rate. First of all, just last year was unusual in Q2. That's an unusually low rate for us. Q1 was 25.5, basically. I had originally thought it would be about 25, maybe a little bit less, because historically, it's been a little bit better than that. Two things. One, the easy one, which is just a currency change, which means that money coming in from abroad is worth less. When I have my mix of earnings, my U.S. amount is more than my foreign, which is often at a lower rate. That's one piece. That actually was not the biggest piece.
It was equal to the fact that especially in Europe, new licenses were lower because there were significantly more cloud bookings. New license, which would've been a number I would've recognized immediately at the end of the quarter, did not fall to the bottom line. In fact, they booked a lot of cloud, which will not translate into revenue until as it's going to play out. By next year, we're going to finish this year, I feel like that will flow all the way through again, and we'll get back to the tax rates we were at. Now, again, assuming currencies don't do more than they have, and I heard Janet Yellen the other day also, who knows what's happening. Assuming that, we should normalize back into normal tax rates in about three quarters.
Unless, of course, there's a tax rate change, you guys can guess that better than I can. Okay?
Thank you.
Thank you.
Mark. Over here.
Over where?
Right here.
Okay. Yeah, over here doesn't really help. Go ahead. Hi.
Hi. Derrick Wood at Cowen. I had a question. We talked to some partners, and it sounds like there's a lot of excitement in what you're doing in manufacturing within the Cloud ERP Suite. Can you just give us a sense where you are with this release, when you think adoption's going to start to kick in, and maybe what your customers, your JD Edwards, EBS customers are saying in terms of interest in moving to the cloud?
Can I start?
Yeah.
Okay. We ourselves are moving to the supply chain products. We originally, my manufacturing team and supply chain team, as you can imagine, we're very, very happy with the E-Business Suite. Many Oracle customers are actually E-Business Suite customers, actually very happy with their supply chain planning and all of that. I initially thought that in reviewing it, that we would have a lot of gaps because our supply chain is so full-featured. It turned out that our cloud offering is actually phenomenal, and customer after customer are extremely enthusiastic about moving to it. I think that in my view, it's one of the last things holding back some E-Biz customers, JD Edwards customers too, but E-Biz customers from moving full out into the cloud. They have been waiting for supply chain, and also, I don't know if we've announced this. Warehouse?
It's always dangerous if you don't know.
We also acquired a warehouse management product, which was that last little box that had to be filled, and we did that by acquisition. That's LogFire. Yes, okay. I think that's going to be fantastic. The initial views are very, very, very positive, and I can tell you as what I consider myself usually customer zero. We look at it and we work it out first. Very, very excited about it.
I'd add to it that I don't think it's just about the E-Business Suites. Those guys are going to move. The bigger question is our ability to go gain share from other providers. As we mature this product, I'm surprised Steve wouldn't have described it, but we have some competitor, big accounts that are now thinking about moving. We needed this release to be able to go do it. If we can begin to get a couple of those to move, that's about all it would take to hit a tipping point. These are named companies now saying, "I want to move to the cloud. We waited on this release to have a full feature set. We're now ready to go." Yep. Okay.
Oh.
Right here?
Are we done?
Meaning we're handing it off to you?
You are more than welcome to take more if you'd like.
You want to take a break?
Any other questions?
Do we want-
Okay, we'll do one question, John.
We'll do one question.
We are going to break.
We'll go to break and then Larry will be here.
Larry will be here. Correct.
Is it this John you were talking about?
I think it's-
Yeah. You get the second shot? I raised my hand really quick.
Okay.
Really quick. Thank you, Safra. Mark, Doug was up just before-
Wait, you thank Safra, then you're going to ask me.
All right. That's crap.
That's great.
She's much better looking. Doug was up before you guys, and he's talking about the accelerated buying experience, which I think is really interesting. He acknowledged that that's been a sort of a complaint or an issue for customers for a long time in dealing with not just Oracle, but any enterprise software company. I'm just curious, and Safra just made a point about seeing even better margins on PaaS over time scale, presumably. One way you can actually do that is if you could sell everything that way, right? Even on-premise.
No, that's the plan.
That's what we're going to do.
Everything's pushing that direction.
Yes. Imagine. Just to be clear, I think Doug said this, it's like one of the things we want to keep repeating. We started this in March, the beginning of Q4. Of the times to have great ideas to implement, this was not it. We really did this with a minimal amount of training, yet we got about 40% of our transaction volume, not dollars, but transaction volume, mostly deals $200,000 and below cloud specific. We raised that. We did more training in Q1, we actually raised the bar in Q1, we got up yet even higher, closer to high sixties or something like that in transaction volume. Our strategy is to do this across everything we do.
Including license plus maintenance?
Yes, sir.
Definitely.
Yes, sir. We want to turn this into frictionless because this is not only good for the customer, it is great for us.
Okay.
This is clearly the direction that we are going to. I know Doug described this. We have a set of terms that we would always give on, but you had to ask.
Got it. Okay.
Now we've loaded those in, and we want to move through this as quickly as we possibly can.
Okay, great. By the way, you're not a bad-looking guy. She's wonderful.
Thank you.
We could compare you and me as opposed to me and Safra. You're better looking. It's better for me.
Old and withered appreciates it. Thank you. I do actually want to say one last thing about this, because I am amazed that none of you wanted to talk about NetSuite today. Okay. Let me just take a moment to talk about. I'm going to stay seated. My ankle's just killing me. If you don't mind. Can you see me? Yes. Okay. As part of this transition, the accelerated buying experience, mirror image is the accelerated selling experience, by the way. Okay. The massive expansion in our sales force has given us the confidence to go down market worldwide. Okay. One of the questions Doug got was, what about lowering G&A expenses and all that? On its way. Okay. Absolutely on its way. A more automated, simplified process. No one else is doing this in the enterprise software business, and that is our path.
As we reduce our selling expenses and our operating expenses at scale from just providing the cloud services, and as we benefit from G&A sort of per order going down, we can afford to go down market economically at scale and expand globally, which they have not been able to do before. For us, we are going to be small to large, global, automated selling experience at scale, providing it, and at scale in an automated way in the back office. We're doing all of that top to bottom. We view this as our opportunity to become very popular in the applications market. We have been fighting with SAP for decades, we would like to be top to bottom, multiple industries global, and that's our plan here.
We have cleared from an antitrust point of view everywhere except the U.S., which we hope to clear very soon. We're looking forward to getting this done. That's kind of it. I don't have really much else to say. Is that good? Are we good?
Greg?
Greg? Yes, please. It's been a long road. We are in the home stretch. If we could please reconvene here at 3:15 P.M. Larry's en route, and we'll start right around 3:15 P.M. Thank you.
in charge of this company. Okay. Whatever you guys want to talk about. God, this is a dangerous chair. Wow. Moves. Yes, sir. Can we turn the lights down? I'm sure we're broadcasting this to Ukraine and everything like that, but those lights are ridiculous. Oh, yeah. Thanks, that's better.
Hi, Larry. It's John DiFucci from Jefferies. You had some really exciting announcements today, and you talked about in your keynotes on Infrastructure as a Service. Just trying to understand a little bit better, because the way I've always looked at it in the past anyway, is it was there. If you're going to provide Platform as a Service, you're going to do Infrastructure as a Service, too. It just made some sense because you had to as part of platform, so also offer it independently. When I think of Infrastructure as a Service, I think of AWS, sort of the foundational part in Azure, and I think of commodity, which I don't normally ever think of when I think of Oracle.
I guess if you can talk to us a little bit more strategically on why it's such a big focus right now for Oracle.
Well, we have to do it. If someone wants to bring their existing workloads into the Oracle Cloud, they can't do it unless we have infrastructure as a service. Every custom application runs on compute and storage. If we want to capture most of the applications written on top of the Oracle Database, in other words, if we want platform as a service to be successful, we have to move not just the database portion of the application. We have to move the application itself, whether it's written in Java or Ruby or Python or Node.js or whatever modern language or an older language. It doesn't make any difference what it's written in. We've got to have infrastructure as a service.
If we want to move an SAP application that runs on top of an Oracle Database, and almost all of them do, to the Oracle Cloud, we've got to have infrastructure as a service. The SAP application runs on infrastructure as a service. The SAP Database runs on Oracle. You look at Oracle's foundation business, and I think I've said earlier, we haven't even begun to move our installed base. We're very different than Azure. Azure has moved a lot of Office customers to the cloud. If Office, you can say, is really Microsoft had two franchises, one is Windows. I've never thought it was Windows, actually. I always thought it was Office. The big Office franchise is moving to the cloud, and that's the bulk of what Azure does. Amazon's way ahead of Azure in infrastructure as a service.
The bulk of what they've done is move their customers, Office customers, to Azure. They also do infrastructure as a service, but that's the focus. This will make sense at Oracle. The bulk of Oracle Cloud should be the Oracle Database. We can't really start moving those people. We can't start moving those people until we have infrastructure as a service because the application, the custom application, runs on infrastructure as a service, and the database portion of the application runs on platform as a service. Okay. We've done pretty well in platform as a service. How are we doing that without infrastructure as a service? Well, you buy platform as a service with infrastructure as a service for custom apps. You also buy platform as a service with SaaS for cloud apps that we sell, ERP, HCM, marketing, so on and so forth.
A, we had no choice. If we want to move our customers to cloud. By the way, our big customers haven't really started moving much to the cloud yet. That's a huge opportunity for us. The next thing is we wanted to make sure that our infrastructure as a service had some significant competitive advantages over Amazon. We did version one of infrastructure as a service. We learned a lot, and now we've come out with Gen 2 of our infrastructure as a service, which we're going to now roll out. Why didn't we roll out infrastructure as a service version one broadly? Because we didn't want to build a bunch of data centers based on what we currently had. We thought we could do way better than that once we had Gen 2 done.
Gen 2 means you build a fundamentally simpler, cheaper, more reliable, more secure, and more powerful data center. We have significant advantages. In some cases, our costs are multiple times lower than Amazon's. Much lower cost in building our data centers. We deliver better performance. Why Infrastructure as a Service? It is a commodity. Commodity isn't necessarily a bad thing. If you can generate electricity at half the cost of your competitors, being in the electric business is pretty good. We have a high margin platform business. We have an even higher margin SaaS business. We will have the highest margin Infrastructure as a Service business because the technologies inside of our data centers are technologies that we primarily build, which is not true of Amazon, not true of Microsoft. It is just a much more modern architecture. It is a much simpler, flatter network.
We think we not only get performance advantages, which are very important and it makes it a little less of a commodity, but huge cost advantages. By the way, cost advantages when you are in a commodity business is where you make all your money. A, we went into business because we had no choice. That is how we move our base to the cloud. We think even though it is a commodity business, I agree with you, we have better technology, significantly better technology, newer technology than Amazon, and we can do it much more cheaply than they can. By the way, we have big advantages in security, which is going to be a very big deal, and big advantages in reliability. Amazon does not support what we call RAC. That is where multiple computers running a database.
If you are running on Amazon, you have to understand if that Amazon server goes out, you are down hard. Now, maybe you do not care. A lot of applications, Oracle Database, a lot of applications. Oracle Database does not go down for years. Years. No downtime. We can do that in our cloud. They cannot do that in their cloud. I am not going to go into all of it. I have made presentations on this in the past. We can detect all sorts of intrusions they cannot detect because our technology is different than their technology. We think, again, we can differentiate on reliability with RAC. We can differentiate on intrusion detection, all sorts of other kinds of security. We have a much more secure network. We can do all of those things and deliver a service that we sell to our customers for less, and we make more money than Amazon.
We are very excited. It is a huge business, by the way. It is a huge new opportunity. I will just stop there. Thanks. Over here. Yes.
Hi, Larry Cash from Bloomberg Línea. On the next Gen 2.0, can you talk to us about how you view Amazon Web Services revenue, which is certainly massive, $10+ billion, and Oracle's just getting started here. Historically, it's been tough to leapfrog that kind of revenue stream with just pure technology. Maybe you have some thoughts there. What's your plan to leapfrog from a revenue perspective, technology perspective with your next Gen 2.0? Thank you.
Hold it. I'm not sure I understood the question. It's hard to leapfrog Amazon with technology? I don't know what you're saying.
From a revenue perspective, they have so much mind share that their business is about.
Oh, mind share. No, I mean, they have the advantage of being the first mover. You know who uses these things? Computer programmers. People are running existing database workloads. Let me just talk about doing what Microsoft did, which is let's say, moving the Oracle Database into the Oracle Cloud. Let's say we can just accomplish that. That's most of corporate data. We won the database wars, what, a decade ago? You can argue Oracle hasn't been competing in database for at least 10 years. Has anyone talked about that? I know SAP said they're going to build this thing called HANA, it's going to put us out of business. The last thing that was going to put us out of business was open source.
There's PostgreSQL, there's MySQL, there have been so many stories of the next thing that was putting us out of business. Our database customers really can't run their mission-critical workloads at Amazon. They can't do it. Now that we have a cloud, are they going to at least try the Oracle Cloud? These are engineers who are running on Exadata today or running Real Application Clusters using the Oracle Database. Where are they going to move the Oracle Database? To a single node? The most they can run is one small four-socket server in Amazon. That's the biggest database load you can run. It just doesn't work very well. When you try Oracle, the biggest workload we can run is multiple Exadatas, 100 times larger, 100 times faster. You think they're not going to try? These are Oracle customers running the Oracle Database.
The bulk of corporate information is stored in Oracle Database. You think Amazon's got so much mind share, they're not even going to try it in the Oracle Cloud? I don't think so. I'll tell you where Amazon's done very well is the new generation of ISVs. I think the new generation of ISVs, a lot of them went to MongoDB and wrote, Uber wrote or Netflix, we would never have gotten Netflix anyway. Netflix is downloading masses of movies. It's a very different application. Most of the corporate world, most of enterprise, most of government, most of corporation, everything's stored in an Oracle Database. That's going to end up in an Oracle Cloud. Amazon just can't do it. The alternative is they can say, "Well, I'll tell you what, I'm paying too much for the Oracle Database anyway.
What I'll do is I'll move up off the Oracle Database into Redshift." Well, after spending a huge amount of money, you will run hundreds of times slower. You got to tell your programmers to redevelop their applications, retest your applications, move it to Redshift, and then run 100 times slower and less reliably? I think Amazon has as much chance of getting the Oracle Database workload as they do of getting the Microsoft Office workload. People have Microsoft Office licenses. You can bring them to Amazon. It just doesn't work as well as Microsoft's Office 365. Amazon's going to have a very hard time competing with Oracle against Oracle. What is Amazon's database that they're going to I grant you, they got mind share. They got incredible mind share. What is the database they're going to sell to their customers to compete with Oracle? Tell me. Anybody.
Microsoft got the Office workload. We're going to get the Oracle workload. These are early days yet. Again, I've heard this for as long as I've been here. I've always heard the next technology that was going to put us out of business. It was object database. There have been so many. We run the Oracle workload in the cloud much, much better than they do. I showed slides. OLTP, about 10 times faster than they do. Query processing or what are called analytics? 40 times faster than they do. Why would a customer do that? Mind share? We have a lot of mind share among Oracle customers. Amazon, yeah. As far as I know, they invented the name cloud. I thought it was called SaaS or something else, or network computing, and it had many different names.
I grant you, they have great mind share. Amongst Oracle customers and Oracle DBAs, Oracle, that's their title, Oracle DBA. We have great mind share among people whose title is the name of our company. Oracle DBAs will try the Oracle Cloud, and the Oracle Cloud is so much better than Amazon that it's not going to be a hard decision. It's also much cheaper. We'll get a chance at bat. I think we'll do as well with the Oracle workload as Microsoft did with the Office 365. It's an unfair fight. We control all the Oracle source code. We put what features into Oracle we want to put into Oracle. Amazon doesn't do that. Amazon puts storage and a microprocessor up and say, "You can use this. Pay me rent." It's a great idea. I'm sure the guys who invented car rental, it's a great idea.
It's a different idea than inventing the car. I guess that's my answer. I think we got great mind share among our own customers. Moving that workload is the majority of corporate data. I think the best years of Oracle are in front of it. The reason I'm still here is this next few years is going to be very exciting. You can say, "Well, I just don't believe you," but no one believed us when I said we're going to do well in SaaS. You can argue we're number one in database. We were only number two in applications when the SaaS wars began. SAP was number one. When people started moving applications to the cloud. It's over. Already, even with our late start, we are selling more SaaS than anyone on the planet Earth. We will sell well in excess.
We sold more last year, more SaaS than salesforce.com did last year. We'll sell a lot more SaaS than salesforce.com does this year. We're in first place. If you look at that business of ours is growing, the kind of combined, the SaaS business and the associated PaaS business. Because some PaaS goes with SaaS. The bulk of PaaS is going to end up going with infrastructure as a service. That business is now growing 82%. As we get larger in SaaS, our growth rate is getting higher. Who else is growing? Who else has scaled growing 82% in SaaS? We were not the incumbent leader in applications. We were the number two player. SAP is the number one player. We will be the number one player in SaaS. No question. In fact, we just measure votes last year.
Customers who voted last year, customers who voted the year before voted for us in greater numbers than any other SaaS provider. We were number two, coming from number two to number one. In database, we have a gigantic lead over anybody. In fact, it's interesting you're talking about Amazon as a competitor to our database because they're not a database company. It's fascinating. What database does Amazon have? Surely not Azure. Surely not Redshift. Surely not Aurora. I said Azure, I meant to say Aurora. You know who develops Aurora? Us. It's MySQL. All they did was open source. By the way, MySQL was the last thing that was supposed to kill us. Open source MySQL. Redshift is PostgreSQL, a UC Berkeley project.
In the database business, we beat IBM, we beat Microsoft, and we have more share than both of them put together, number 2 and number 3. Amazon doesn't make a database. They lifted Redshift PostgreSQL, and they lifted Aurora. They don't make database. How can a non-database company beat a database company at database? It's an interesting idea. No, I'm serious. I hear this, and it's okay. How are they going to do that? It's the same question as Amazon's going to kill Microsoft in Office. Amazon's going to kill them in Office. How? They don't make anything like Office. Google makes some stuff like Office. Google's got its own kind of suite of applications. Office is still doing just fine, thank you. It's just peculiar. Again, most of corporate computing are applications that run on top of an Oracle Database.
We do that 100 times better than Amazon. That's a workload we're going to get, and that's going to drive incredible growth. The margins we get will be the highest in SaaS, second highest in PaaS, but we'll have very strong margins. Better margins than Amazon has in Infrastructure as a Service because we have better technology. We also have a much more attractive mix that we have high-margin SaaS and, if you will, high-margin PaaS to go with our middle-margin Infrastructure as a Service. Next question. Surely someone wants to ask another question. Right here.
Mark Moerdler Bernstein. Larry, how should we think about the opportunity for analytics, AI, machine learning, whatever you want to call it, in the Oracle Cloud?
I love that you asked that question.
Thank you.
Analytics in the Oracle Cloud. Do you know who the last big purchaser of our analytics software was? Really big? Amazon. How good is the opportunity for analytics? Well, Amazon's buying our analytics. Amazon runs most of their retail stuff on the Oracle Database. They're buying our analytics. We think the opportunity for analytics is enormous. We're selling analytics to our big database customers. We sell a lot of analytics. How big was the Amazon deal? Do we want to talk about it, or we're not allowed to? Don't tell them. $ millions and millions. On top of their much bigger purchase of Oracle Database. You would think if anyone's going to get off the Oracle Database and onto their own analytics, it would be Amazon because you would think Amazon has great mind share at Amazon.
Everyone's business card, even the security guard says Amazon. I mean, it's a phenomenal mind share. Amazon bought our analytics. Amazon uses our database. Everyone uses our database. I'll keep going on. I feel really good about our database. SAP, I know it was a couple of years ago, people were talking about HANA. HANA is going to be an in-memory database. SAP is going to kill us. Before Amazon was going to kill our database business, SAP was going to kill our database business with in-memory HANA. Let me tell you how I know that's not the case. What do you think SAP runs in their cloud underneath Ariba, their procurement system? HANA? Anyone for HANA? They run Oracle. SAP runs Oracle, not their own. You would think HANA would have incredible mind share inside of SAP because they make it.
It's on the front page of their website. Incredible mind share. They don't run Ariba, and that's their biggest cloud app. They also have SuccessFactors, another really big cloud app for them. It's the center of their modern HR products. Well, surely SuccessFactors runs on HANA. Who's guessing HANA? Oracle. SuccessFactors runs on Oracle, and then Concur, the third party. They have three properties in the cloud, only three properties in the cloud. They all run Oracle. Nothing runs HANA. You say, "Am I overconfident?" Maybe. If SAP can't get SAP to use HANA, and Amazon can't get Amazon to use Redshift and Aurora, there must be a reason. It's not like the guys at SAP love us. They don't love me personally, which is hard for me to deal with at times. I'm getting over it.
They sure don't like Mark that much or Safra. The lawsuit was really bad for them. They hate us. Their entire cloud is based on Oracle, their entire SAP cloud, not their own technology, because we're so much better than they are. When your biggest competitors, when the people who are supposed to be putting you out of business or at least hurting you, not putting you out of business, hurting your business, really a big threat to your business, are among your biggest customers with their latest and greatest stuff, you should feel pretty safe. We're really good at this database stuff. We've been doing it for a very long time. It would be amazing if Amazon could build database. Or for that matter, even SAP build it. It'd be unbelievable. We beat IBM when IBM was IBM.
I don't know who they are now, but when IBM was a great company, they built Db2, we built Oracle, and we won. They had a lead, but we won. We also beat Microsoft when they built SQL Server, and we won. We beat a bunch of other people, Ingres, Informix, Sybase. You just survey customers, and what are they running? What do they store their data in? It's all Oracle. It's all Oracle. That's going to gradually move to our cloud, not Amazon's, because we do it so much better. Yes, ma'am.
Hi, Sarah Hindlian with Macquarie here. Two questions for you, Larry. One a little bit general, then one maybe a little technical. We heard about the release of R2 this week, and we've certainly been hearing of some pent-up demand for that product. Would love to know how you see that impacting the business. Then number two, if I read your Infrastructure as a Service architecture slides correctly, it looks like you're defining your cloud at the network layer as opposed to a lot of your peers that may be defining it or maybe using hypervisors. I'm wondering a little bit about the economies of scale you foresee there, in particular for data-heavy workloads. Thanks.
You're right. We have a truly virtualized network. With off-box virtualization, I think that the biggest difference between what we have with Gen 2 is a very different network. The magic we have with our Gen 2 data centers or our Gen 2 IaaS, a version of it is in the new data center. The cool thing about this is you can pick up an existing workload, an existing corporate workload, and move it intact into one of our Gen 2 data centers, onto our Gen 2. If you are using VMware and Red Hat, by the way, we'd love for you to move to our hypervisor. We'd much prefer that because we wrote the code, we can support the code. Actually, it's open source. It's Xen. We've contributed to the community. Our operating system is Linux or Oracle Linux 6.7. It's open source. We contribute to the community.
We can support you more easily. It's one-stop shopping, if you use our hypervisor and our operating system. You can bring your own, which is really what these other guys can't do. You can lift up an existing workload. In fact, you can lift up an existing data center, IP addresses and all, network and all, and plunk it down without changing anything, and plunk it down in our cloud. That is a very big deal. Right now, people are moving an application at a time, and they have to partially rewrite, at least reconfigure the application, retest it on the new OS, change the IO. They have to do a lot of things to move an existing application to a data center. They have to retrain your people. There's a bunch of things you have to do. Not so with our new bare metal experience.
Take your existing network, your existing network definitions with your existing range of IP addresses, your existing operating system, your existing hypervisor, your existing database, your existing apps, your existing everything, move it over, and it just runs. No one else can do that. We think that's very important. I think you'll see a lot of emphasis on the network because in these cloud data centers, your limitations in performance are largely the network. Your risks in security are largely the network. We've spent a lot of time re-architecting this and building this virtualized network in our Gen 2 data centers. It allows us to give you much better security as messages move around the network. Not only is everything encrypted in motion, it's encrypted at rest in storage, it's encrypted in motion on the network.
Because it's virtualized, each message is encapsulated with virtual addresses. It's almost impossible for people to interpret these messages as they're flying around. Even if they were to get into the network, which is, by the way, impossible because our control plane, everyone else does their software-defined networking on their computer servers. We do our software-defined networking in these network processors off-box with no access to the public internet. You can't break in. There's no door. Yeah, we did it very differently than everyone else did it. The big differentiator between us, because our storage is still SSDs. Now we think we have better software for storage, and we're smarter about it, blah, blah. The big difference in our Infrastructure as a Service is our network and how we do virtualization.
By the way, our storage virtualization is also completely off-box, so you can't get to a storage device. You can't remap any of our mapping tables through the public internet. There's no public internet connection. The way we establish public internet connections, even for users, is in a virtualized way, and we can turn those on and turn those off and control those much better than anyone else in this cloud business. You're right. We spent a lot more time on the network. I think we spent more time talking about it because our network is so different from anything that came before. When we decided to do it this way is when we slowed down in Infrastructure as a Service, waited till we finished all the software. There's custom hardware in there, by the way.
The custom hardwares are what we call the control plane, these network processors. We wanted to finish all that so we could deliver this much more secure, much more performant infrastructure platform. It took us a while to get all the software done. Non-trivial, get it all tested. We've been in test for about a year, and we're now rolling out in scale. That's the big difference. I forgot the first part of your question.
R2 release.
R2 release. Again, we developed This is back to how we're developing the Oracle Database with lots and lots of powerful cloud features. The multi-tenant release we think is very attractive. We're talking to big ISVs. They'd shock you, the name of these ISVs. We want to bring big ISVs into the Oracle Cloud. They were waiting, again, for our Infrastructure as a Service, but also the new generation of our database, where we went from 256 PDBs, pluggable databases, virtual databases, if you like, to over 4,000 per container database. Let's say you're a big SaaS company and you want to run on the Oracle Cloud, you want these database efficiencies. Most SaaS companies run on top of the Oracle Database.
We've been putting in lots and lots of SaaS features for ourselves, but also for this new generation, the largest ISVs out there, so they can lift what they have and move it into our data center. Our targets, we have a very large group of ISVs that haven't moved to the cloud yet, or you think of them, they're already cloud companies. They're already SaaS companies. It doesn't make sense to me for very many SaaS companies to run their own infrastructure or for that matter, run their own database. If we can do it, we should be able to do it much more securely than they can do it, much more reliably than they can do it, global presence and at a much lower cost. We should be able to save them money, get a better performance, better security, better reliability, all of those things.
We've aimed at 12.2, version 12c Release 2, at dramatically improving our capabilities in the cloud for serious cloud ISVs, serious SaaS companies. We think that's a gigantic. These are big contracts and big deals. I think that's one thing that no one is expecting. In addition, we take existing applications, existing ISVs, and because our multi-tenancy is at the database layer. Now at the Java container layer, we can take their existing applications, an ISV that maybe they were doing more like hosting as distinguished being a SaaS company. We can move those ISVs and make all of their applications SaaS applications and move them into our cloud. We think the ISV business is just a gigantic new opportunity.
Of course, we have a bunch of features, especially in the security area and the networking area, that all of our end user customers have been waiting for and are very excited about. We think more and more of those customers will discover the features and the functions and the advantages of 12.2 not by loading it on-prem, but by just logging on to the cloud. The Exadata service is now priced at $175 a month. I think it's extremely attractive from everything from an MIT graduate student who's starting her own company to a large company like General Electric, with thousands and thousands of developers who can develop on our cloud rather than develop in their data center. I think the opportunity for 12.2 is gigantic. I think the opportunity for us right now is staggering.
You look at where we are in the cycle. I know I'm repeating myself, we haven't started moving that Oracle workload out of big companies and into the cloud yet. We have started, we're so early. The size of the opportunity is just stunning. ISVs and end users. Okay. Yes.
Hi, Larry. I'm Michael Turits from Raymond James. I think you began to partially answer this, how aggressively do you go after the type of workloads that are non-Oracle workloads that Amazon has been successful with? You mentioned ISVs, newer ISVs that they captured a while ago. You just rolled out a container as a service offering. There are places that are non-Oracle they've been successful. How aggressively will you move in that direction?
Well, I think there's a natural cycle. I don't think you're going to see us making an offer to Netflix that they can't refuse. If naturally we can charge less, if we really do have a performance advantage and a security advantage in our Infrastructure as a Service. Here is where Mindshare is a big issue. You're a startup, and we really have these advantages. Engineers, it's amazing how fast word of mouth, one engineer talks to another engineer and says, "Wow, this thing is much faster than Amazon," or much cheaper or much better, it's much more secure or this or that. I don't think you'll see us aggressively pursuing them other than the fact that we built infrastructure with significant differentiators. Engineers are engineers.
Engineers notoriously are difficult to sell to because they don't like to be sold, but they like to go on the web and investigate and look at features and function, spend $100, $200 on their Amex card and try Exadata as a Service and convince themselves one way or another whether this really is better than what they were using before. I think you'll see our efforts around just this enormous software install base around Oracle databases around the world. That is such a gigantic opportunity. That's clearly what we're going to focus on. It's not we're going to pay no attention to these other guys. The other guys, we hope the quality of our technology will attract them. Our focus is moving that Oracle database workload into the cloud. If we do that, we're by far the largest cloud. There's nothing remotely close.
Certainly, if you compare the size of the Oracle workload, the number of computers, and the amount of storage associated with the Microsoft Office workload, I think it's enormous. Yes.
Hi. Kirk Materne with Evercore. Given that scale is somewhat of a prerequisite to compete at the Infrastructure as a Service level, there's a certain amount of capital expense that you have to be willing to spend to get into that market. Yet there's also companies like yourself, like you mentioned, that have a competitive advantage around your own products and Microsoft around the productivity tools. How many big Infrastructure as a Service clouds do you think the market will support over the long term? Is it two or three? In the apps market, it got down to sort of you and SAP. I realize we're early on, but longer term, how many do you think can sort of succeed?
Yeah. It's a great question. Who do I think can go into this business? Amazon's obviously already in it. Microsoft's obviously already in it. I'm very optimistic about, again, we're in it. We're the leader in SaaS right now, which depends how you count. If you count votes last year and votes the year before, we're selling more than anybody else. I think we can just extrapolate on those curves, and we win SaaS big time. I think we're in the process of winning in SaaS. I think our infrastructure leapfrogs everybody. Plus, we have this huge installed base, so I think we can play. That's Amazon, Microsoft, us. Google, if they want to. They print money. They certainly can play. Technically, they're a very capable company. Facebook, if they wanted to. Again, they have a lot of data centers.
That's one of their core competencies, is data centers. Another core competency is data. So far they've shown not a lot of interest. Let me tell you, as we look at the quality of their data centers, their data centers are a lot better. Facebook data center is better than a Google data center, and it's better than an Amazon data center. Let me tell you why. They're newer. They're just newer technology. They've got very high-quality data centers, so they could play if they wanted to. I can't think of anybody else. We were the largest, we still are the largest enterprise software company in the world. Microsoft is the largest software company in the world. Amazon is a very unusual company. They're obviously pretty good at a lot of things. They were the first mover in AWS.
Again, those four I can think of, but I don't know who else can do it. It's a lot of technology to build these things. The cloud is not just. People tend to think, it's very funny. You look at a salesforce.com and you look at an Amazon, and they're both called big cloud companies, and they have nothing to do with one another. They don't do the same thing at all. My belief is to compete broadly in the cloud at the SaaS layer and the infrastructure layer and the PaaS layer, A, you have to be a pretty big software company. I would argue Amazon's not a real big software company. Don't get me wrong. I have a lot of respect for Amazon. I really do. I think they've done an incredible job.
I don't think of them as a really big software company in the way I think of Microsoft as a big software company, or as the way I think of Oracle as a big software company. Number one database company, number two applications company, Oracle. Number one in middleware, Oracle. Oracle's got not a lot of number one positions. We're a very big software company. There aren't a lot of those out there. Then you have to either develop or have already a core competency in building data centers. Amazon had a core competency in building data centers for their retail business. They were a huge consumer of IT technology, and they were very good at rolling out these scaled data centers, as is Google. Search requires real competency in building large-scale data centers. Plus, they're a pretty good software company, Google is.
Those are the two competencies, right? You've got to be good at software, and you've got to be good at building, managing, architecting these hyper-scale data centers. I don't think there are a lot of companies that fall into that description. In terms of the cost of building these things, it's really not much. It's billions of dollars, but the size of the prize is hundreds of billions of dollars. The data centers are just not that expensive. If I look at how much we have historically spent on acquisitions, and we can build these data centers, and the overwhelming cost of the data centers are the computers in the data centers. We can start a data center small and one of the things we've gotten very good at is just-in-time manufacturing and shipping of our own stuff to our own data centers.
We build all the stuff that goes into our data centers. We build it in Hillsboro, Oregon, and I forgot what the early cycle was, but I think we're down to what, eight weeks, seven weeks from order to shipment? We're down lower. Okay. We keep on moving that down, but it used to be a long time, but from the time an order is placed to the time the hardware is provisioned in a data center, I think is now under seven weeks. Clearly, we have excess capacity. It doesn't mean when a customer orders seven weeks later, they're provisioned. When a customer orders two minutes later, they're provisioned in a data center.
I'm saying when we decide to increase the capacity of one of our data centers or open a new data center, it takes us about seven weeks to get the hardware there. From the time we configure an order, from the time the hardware is going to ship. We'll continue to get better at that. Those are the core competencies of an IT utility. You got to have good technology, good software, and the ability to operationally run these big data centers. There aren't a lot of companies like that.
Hi, Larry. Karl Keirstead at Deutsche Bank. Larry, I listened to your keynote on Sunday when you were talking about your top competitors. You've spoken a lot in this meeting about Amazon on the infrastructure side, but you identified Workday as your top competitor on the SaaS side. I don't want to read too much into that, but I feel like in prior years, you've been gunning a little bit more towards Salesforce, I'm wondering if you could explain that. Thank you.
Salesforce is a company everyone's heard of. The problem with Workday is a lot of people, I'm sure everyone in this room has heard of Workday, but a lot of places, no one's heard of them. Yeah, we have more competitions against Workday in ERP and HCM than we do against Salesforce in sales. We have competitions against Salesforce in marketing, that we are the leader in marketing. We tend to win those. We think the size of the prize against Workday is just bigger. We think the ERP-- if you look at the size of the ERP market, don't take my word for it. In the previous generation, SAP was the largest ERP company, and Siebel was the largest CRM company. The ERP business is an order of magnitude larger than the CRM business. Our focus is where the money is.
We think it's very important that we win in ERP and win in HCM. By the way, in the mid-market, we think. By the way, I haven't had a chance to talk about it, we think we're leagues ahead of Workday. Workday, I think, made a fundamental mistake. I know those guys don't like us. I have no idea why. Must've been something Mark said, I don't know. They decided not only were they going to build an HCM suite, okay? Which they did. They're going to also build an ERP suite. Much bigger job, 10 times bigger job, easily. They're going to build their own database. Everyone know that Workday built their own database? You kidding me? Sound like you're talking to guys at SAP, except you're a startup. Workday is splitting their resources between building database, HCM, and ERP.
I think that's impossible. I don't think it can be done. Database is ungodly complicated. If you look at our database presentation, you look at Oracle release 12.2, version 12 Release 2. They're going to build something that competes with that? With PDBs and in-memory database optimization, encryption at rest, encryption in motion. They're going to do all of this stuff? Real Application Clusters. It's impossible. Salesforce didn't do that. NetSuite didn't do that. Concur didn't do that. SuccessFactors didn't do that. Ariba didn't do that. Nobody did that but Workday. Thank God they decided to do it that way. They'd be much better off if they were running on top of Oracle database and middleware, they chose to build it themself. Therefore, they're our primary competitor in ERP, and we're doing very well. You can just look at our ads.
When we say we're selling 10 times as much in terms of customers, I think we're understating it. I think they count funny things as ERP, like expenses and payroll as ERP. They're having a very hard time making ERP work because they got to have a database. Making this database scale is very difficult. Just because you can get it to work on HCM, which is a much simpler set of applications than ERP. They haven't gotten to supply chain yet. They have procurement. They have purchasing. They don't have procurement. They don't have any of supply chain. They don't have any of manufacturing. We have all of manufacturing. We have all of supply chain. Our footprint in terms of modules is all the complex reporting and tax reporting and planning and budgeting. They say, "Well, you can use Anaplan." We have Anaplan.
We have one integrated suite for ERP, they don't. What they have doesn't work very well for a lot of reasons, not the least of which is it's their own database. They tried to build their own database and their own middleware. It's crazy. Their own reporting. Just about impossible. Why do I talk about Workday? Because if we win the ERP war in SaaS, then we're 10 times larger than Salesforce. I think that's about the market differential. ERP is about 10 times the size of sales automation, and Salesforce is still predominantly sales automation. In fact, salesforce.com is predominantly Salesforce automation, most of which is in North America. The amazing thing is how successful salesforce.com has been with really one application. Again, I know they got marketing. They've made a number of purchases in marketing, and they've built some interesting stuff in service.
I got it. If you look at the bulk of their business, sales automation in North America, look how successful they've been. Look at the size of that business. Look at the size of a SaaS business and compare it with the size of the last generation on-premise business. The SaaS business is much larger, and it should be much larger because you're not just buying a software license. You're buying software plus storage, plus a network, plus servers, plus service. Of course, it's much larger. Customer's getting much more. They're getting a data center and a help desk and so on. If Salesforce is that big with sales automation, how big is the ERP company that wins the SaaS war? Really big. Really big prize. That's why I talk about Workday. Yes. Yes, sir.
In the back here. Sorry.
Thanks very much. Ross MacMillan from RBC. Two or three years ago, I think, when we were talking originally about the cloud, you made some statements about data centers being real estate and power being 100-year-old utilities, and there was no value there. When you look at the big players in Infrastructure as a Service like Amazon and Microsoft, they've spent billions and billions of dollars building out designated data centers, and you talked about Facebook having more modern data centers. I believe, and I might be wrong, but I believe your strategy has been mostly co-location with your technology in co-located data centers. Does that have to change as you get more serious about Infrastructure as a Service?
I think we're pretty serious now. I think it depends. Again, that's also a very interesting question. Do you need to own the building that your computers are located in? It's an interesting question. In general, I like the real estate business. We own most of our headquarters. We're just building a huge super center in Austin, Texas right now. We bought all the real estate. We're building all the buildings. I don't think it makes a lot of difference whether you own the real estate or you don't own the real estate. I don't want to go into the details of our strategy, but in terms of where we're going to put all of our data centers. You can't, say, have one global strategy and say, "This is the way it is." In China, we're not allowed to own data centers.
We've got to have a partner. Networks are highly regulated. Maybe we're better off in some countries to partner with the telephone companies and use their data center. They already have an existing network. They have an existing relationship with the government. We might be better off in an environment that may be highly regulated. It's not highly regulated now, believe me. It's going to be highly regulated. It's a utility. It's a network. It's going to have huge influence on not infrastructure in the Amazon sense, but infrastructure in terms of water and power and all of this. If the control systems run on nothing, this thing has to work all the time, you bring your economy down.
As you look forward as what this looks like in every country and how much it will be regulated, back to the question, it's not unrelated to the question, well, who's going to run these things? Imagine we have four companies, or let's make it 10. Doesn't matter. 10 companies running all the world's data and data centers. You think governments aren't going to get involved? I think it was this week, The Economist had an interesting article. It was their leader, the first leader on The Economist, about hyper companies and some of the problems hyper companies are causing. They were looking at it quite differently. They were looking at it as just Google makes too much money. They have too much power.
It's creating problems in society because you've got these rich IT people or the rich people on Wall Street, whatever rich group you want to talk about, whether they're in San Francisco or New York. You have other people who are losing their jobs because manufacturing ain't ever coming back because automation is not the problem, not exporting the jobs. These hyper companies don't have competitors, and that's really the huge problem, they think. They buy their competitors before their competitors get scale. Anyway, that was the article. Imagine companies getting even bigger. You can imagine the size of one of these cloud companies. Well, it's quite breathtaking. I'll go back to look how big Salesforce is, and that's this one itty-bitty little app. Imagine what I said was true. Imagine Oracle moved the Oracle Database workload to the Oracle Cloud.
How big is that? A little spooky. It's big. It's big, and it better not ever break. It's not a matter, I don't think, of do we want to buy the real estate. We own the computers. We have access to a high-performance network. We specify all this stuff. If we don't own the building and we're not supplying the electricity, I don't think it makes a big difference. Whether we buy the buildings or don't buy the buildings, everything else being equal, I'd rather own the buildings because I think we end up making more money if we own all the buildings, own all the real estate. McDonald's real estate's worked out well for McDonald's. Real estate's worked out well for IBM. Anyone who buys all this stuff over a long period of time makes a lot of money.
I think it's going to be very important for us to have good relationships with governments all over the world, good relationship with telecommunication companies, start to learn some of the regulatory ropes so we can get around there, having the right partners might really help that. The most obvious place is in China, where we're simply not allowed. We've just been regulated out of the business. You can't start a data center. Google pulled out of China. That is not a simple question. It's really an interesting question. It is not a simple question. Has very little to do with The next, pick a number, we build 50 data centers. The cost is the bulk of it. It's not the building.
The way we populate those data centers, we put in capacity for existing customers, as customers order more capacity, we just try to stay one step in front of them with our supply chain, our very rapid supply chain. We make money. I hope we have to build 200 data centers. Whether we own the buildings or not, whether they're in partnership, again, we have to buy networks, right? We suddenly became a huge consumer. The cloud companies become huge consumers of network capacity. Therefore, partnering with the guys who own the network capacity in some way might not be a bad strategy.
I think the next round of build-outs are going to be pretty easy, I think you'll see us experimenting with certain partnerships, obviously, in China and a few other countries, which I will not mention right now, where we think we're better off with strong partners rather than trying to go it alone in those countries. Okay. Over here.
Hi, Joel Fishbein from BTIG. I want to change subjects for a second on going to the SaaS business. Mark laid out a lot of really good metrics around the SaaS business. The one thing that sort of stood out, though, was the go lives. Do you think the company Oracle's optimized the ability to take customers live? Is it efficient enough? Where are you with that? There seems to be a lag between signing the customers and having them go live. Any information there would be helpful.
Yeah. Again, I think it's a great question. One of the things we do a little bit differently in Europe and the U.S. The U.S., we relied heavily on boutique firms, new generation consulting firms to help get our customers live. We have an excuse for taking our time getting to the cloud. We had to rewrite all of our applications. It took us a decade to write the Fusion Applications. There's been tremendous resistance among the large consulting companies who are our partners in the last generation of application implementations. They have not been excited about the new generation of SaaS because the consulting engagements take one tenth the time and cost one tenth as much. Those traditional partners have been slow to adopt our SaaS applications, I'll say, in North America, and we've been relying on boutiques.
I think you'll see us use now what's a pretty large-scale university recruiting program to start building a very large. Our consulting business, if you've monitored it's one of those businesses that we've let shrink and shrink and shrink some more. It's a business. Certain parts of our hardware business we don't care much about. Some of it we care a lot about, like Engineered Systems, where it's high value. There are other parts where we're selling, if you will, commodities we don't care much about. We haven't invested a lot in our own consulting infrastructure, and I think you're going to see that change because the demand now for the SaaS applications is becoming so great. 82% is just the beginning. I think you're going to see triple-digit growth.
In order to meet the demand, we're going to see some of these boutiques bought up by some of the major consulting firms. That will be good because that will give us some scale on the implementation side, and then we're going to have to aggressively get in that business also. Not so much to be the prime contractor, but to assist these other companies with engineers. In the old days, the ideal situation was, let's say Accenture was doing an implementation, and they had 100 people on the project, and we'd have 10. The 10 people would be experts in the product they were implementing. The implementation projects might not have 100 people anymore. They might have 20 people from Accenture or 30 people from Accenture and four from Oracle. The same idea.
You'll see us starting to grow our consulting business because the demand around SaaS, and we want to get people live faster than we've gotten them live. In Europe, we've done most of our own implementations because there aren't a lot of boutiques in Europe to rely on. In Europe, we haven't moved aggressively enough. You'll see us move much more aggressively. We have no choice. You correctly identify one of the limitations to growth, which is our ability to implement our products, and we're making a bunch of changes there. You'll see, again, we have a huge training program for most of our engineers. Most of our engineers we hire directly out of school by now. Most of the guys at Oracle, we recruited out of school. In fact, most of them who are running Oracle right now, I recruited out of school a long time ago.
I used to interview every engineer. I do not do that anymore. In the old days, I did, and there are people like Juan Loaiza and Andy Mendelsohn who are kind of running engineering at Oracle. Ed Screven. It's a long list of people. I know Mark has put in a big recruiting program, university recruiting program for our next generation cloud salespeople that are coming directly out of school. We have a huge training program. They tend to be more computer literate than the generation that came before. The consumers are younger and more computer literate, and we've got a new generation of people that are selling to them. Our distribution is changing radically. It won't be so much that We'll have all of these field offices anymore.
We'll be selling out of these huge centers like in Austin, Texas, and in Santa Monica, California, rather than a million and one field offices. Again, you'll see our consulting organization start to grow again. Yes.
Thank you. Earlier, we heard from Doug about the accelerated buying experience, he talked about how that's a cultural change in the business. When I hear businesses talk about cultural changes and observe them, it always has to start at the top, has to start at the board of directors, the chairman, and then has to filter down to the lowest level employees and come up. Can you talk about that? Because it does seem like it's a change based on what we've heard from customers and partners about working with Oracle versus what you're talking about.
Yeah. Well, I think, I remember the meeting where we made the decision. It was at an off-site meeting in the desert. We decided that we were going to be as easy to do business with as any company. We're going to be as easy as Amazon. You could just take your credit card, click on this, click on that, and you're up and live. Why not for everything we sold? That's what we decided to do. Amazon was the benchmark. I think it's one of the cleverest things they did. I mean, look at the original idea of renting cars, very clever. How hard was it to rent the car? It was very easy. You just walk up to the car and get in it and drive away just about.
I think that would have been a huge impediment to our growth if we didn't make it very easy for you to buy a SaaS application or buy infrastructure or buy something. We got so excited about it, why not everything? Why not have every purchase experience, whether you're buying consulting services or hardware, why can't everything go through this accelerated buying experience? Sometimes customers want to negotiate Ts and Cs, They won't just click, click. They want to get five lawyers involved. Sure, that will happen some of the time, That's a customer demanding that we change our terms and conditions. We decided to redo our contracts to make them human-readable, to make them easy. We think to make them gracious in terms of fairness, that you could read them and those terms and conditions are actually reasonable.
Let's move forward, It lowered our costs of sales dramatically. It eliminated a lot of the friction and heat inside of companies who wanted to consume our services. Everyone's very excited about-- Safra, what % of transactions last quarter went through on the accelerated buying experience? Just under 70. How much was it? Just under 70. We're at 70% of transactions. I think that number is going to go much higher. I think it's the way customers want to buy. Again, I give credit to Amazon. I mean, the company that we were looking at, the company that we were studying, was Amazon. We knew that this was a different bar, and we had to transact business.
We had to sell Infrastructure as a Service as easily as Amazon sold a DVD or a download or a DVD stream, or for that matter, Infrastructure as a Service. We had to make it that same painless, instantaneous gratification without a lot of bureaucracy slowing things down and costing more money. I think we accomplished that, We'll continue to make it better and better. Thank you all very much.
Thank you, Larry. Thank you, everybody, for attending today. If I could just make a parting suggestion for you all. I always find the last hour here with Larry by far the most educational time. What I do personally, I'm just sharing my personal experience, is I take these transcripts of the session you just heard here for an hour, and I've been stitching together a running dialogue over the seven years I've been here. Before we come in every year, I reread it. I'd encourage you to do the same. I think this is when you really start to get a sense of the vision that Larry has. You start to get a sense of things that were talked about two, three years ago. You start to see the importance of them now. He's usually ahead of the curve.
If you don't do it, I'd really encourage you to do it. In closing, just want to say thank you all very much for coming. We know you come from far and wide. Thank you for being here. If you have any questions, please let us know, and safe travels home. Take care, everybody. Bye-bye.