Honored guests, please welcome Vice President of Strategic Finance, Treasury and Investor Relations, Phil Winslow.
Thank you for joining us today for Cloudflare's Investor Day 2023. I wanna just thank everybody for taking the time to join us because I know May, from my prior life, is always a busy time of the year, whether you're not an investor or an analyst. We appreciate you putting us on your calendar. Because we so value your time, we want to lead with the most important item first. What you came here to see, the safe harbor. At least that's what my legal team told me that you were here for. All kidding aside, we have a robust agenda for you today. First, we're gonna dive into products, and you're gonna hear from each of the product leaders driving innovation in our four main product categories.
You'll then hear from Marc Boroditsky, our new vice president of revenue, going into the changes in our go-to-market organization. After that, Thomas Seifert is gonna delve into our financials, followed by open Q&A with Matthew Prince, Michelle Zatlyn, and Thomas. Now, just to set the stage here, as investors and analysts, we think it's important for you to understand the depth and the breadth of our services and how they stack up against the competition, because at the end of the day, innovation is at the heart of Cloudflare. To do this, we're going to have the product leaders who are driving this innovation at the company walk you through each of the four main product cateories one by one.
To set the stage for these four product leaders, I'd first like to just step back for a moment and begin by restating Cloudflare's mission, because it really guides everything that we do. Our mission is bold but simple. We're helping build a better internet, one that is faster, more secure, and more reliable for everyone. Cloudflare's platform was built from the ground up with a full understanding of this audacious mission to literally help build a better internet. Our platform was purpose-built from day one to help customers of every size protect websites, people and networks, and to build and accelerate applications. We don't run separate networks for our different products, we do not use expensive proprietary hardware, and we don't start with one product and then attempt to Frankenstein on others over time.
As you heard from Matthew and Jen in the connect keynote just earlier, although a lot of people know us for our original set of application services, we've grown, as I've been saying, into four main product categories, all delivered on a single platform and all powered by one of the world's largest cloud networks. These encompass an incredibly diverse set of use cases. At the end of today's session, we want you all to have a better understanding of how our product portfolio has evolved, what makes us different, and how we add value to our customers. To start the journey, I would like to welcome my colleague, Patrick Donahue, to the stage to dive into act one, Application Services. Patrick.
Terrific. Thank you, Phil. Good morning, everyone. My name is Patrick Donahue, Vice President of Product at Cloudflare for Application Security. I'm here today representing application security and application performance. Together, these constitute act one of the Cloudflare story. The impetus for act one was a single fundamental change, where applications run. Companies used to build networks by buying a bunch of servers, putting them in a data center, putting a network rack in front of them. These were bandaid boxes from the likes of Cisco, Radware, F5, and Imperva. Each box had a different interface, a different contract to manage, and a different command line to operate. None of them scaled particularly well. When your company grew, you had to go buy bigger boxes, or you had to buy more boxes and put them in different data centers.
Worse, sometimes they shipped with security vulnerabilities, default passwords. These boxes actually open holes in your network when you put them in. Starting in the late 90s with SaaS applications like Salesforce.com in 96, applications started to move out of the data center, okay? In 2006, you had more general compute platforms like AWS opening up the possibility to run applications again outside your own data center. This presented some real problems for customers. First, you had minimal control over the SaaS applications. We'll talk about that a little bit in our Zero Trust talk. Second, you had insufficient capabilities. The cloud platforms were trying to sell you storage and compute and mark up bandwidth and make their money there. They weren't focused on delivering a really secure performing experience. They also had incentive conflicts there.
People wanted to run these virtualized boxes. They made money doing so there. This presented a real opportunity for us. How do we establish a really consistent security posture at the edge and allow customers to move their applications at their own pace? Most companies we talk to are in transition, especially older companies that have been building their networks for a long time, from the left-hand side of the screen to the right-hand side of the screen. They're forced to implement the least common denominator of their systems and manage different systems. Where we came in is we gave them a single dashboard to manage all of the functions that were in these boxes in a really easy-to-use platform, and we helped them modernize their networks. Let's look at some of the use cases we're solving in act one.
The first most important problem for a lot of companies is: How do you make sure that network is reliable and performant and sort of available for use for your customers? The first thing people think about is, how do I keep it online against DDoS attacks, right. Your network has to be hardened against them. The boxes that people are using historically simply can't keep up with the volume of attacks we're seeing today. They also were outdated by the time they arrived. They didn't have. They're not updated with the latest threat intelligence. We help customers make their applications performing and reliable first and foremost from keeping these networks online. Second, by accelerating the applications that pass through them.
Companies like Shopify, for example, that use us to protect millions of online storefronts, their website is their business. If you can't transact online, their customers are very unhappy. We run the fastest authoritative DNS in the world. You can check out dnsperf.com to see that for yourself. Nobody comes close to us on the DDoS side. We actually mitigated the largest attack ever at layer seven just a couple months ago, more than 50% larger than the biggest attack at 71 million requests per second. When companies like Akamai are kicking security researcher Brian Krebs off their network, we're publicly raising our hands, asking to take them on. The state of reason was too expensive. You'll hear about how we've built our network later, to understand why it's cost-effective for us to do so.
The next problem that companies faced as they moved from the cloud was essentially massive data transfer bills. They were used to in their own data centers, paying a fixed amount per month. They had a circuit, one gig, 10 gig, whatever that was. Very predictable pricing. When they started to move to the cloud, they were paying to move assets, whether images, video, static content, API requests, essentially paying multiple times along the way to deliver these assets. There was no predictable billing. If you had a good month, a lot of legitimate traffic, you'd have a higher bill. If you had a bad month where bots came after you and started using your resources, you had a higher bill. This is what Discord faced before coming to Cloudflare.
We're now serving two petabytes of content from our edge to them, for them, saving them over $100,000 per month in their Google Cloud bill. Companies love that our pricing is predictable, and we don't kick them off our network, and we don't charge them when they're down with excessive traffic. We're confident in our products like bot mitigation to stop attacks and ability to use new products that are built on our developer platform that Ali will tell you about, such as Cache Reserve, to use storage at our edge to further drive down cost and improve performance.
After reliability and cost, the next thing that customers worry about is keeping their sensitive data or their customer sensitive data where it belongs. Data is typically exposed to the world via APIs, right? That's for your web applications, for your mobile phones.
When this data ends up in the wrong hands, companies' reputations can be at risk as well as their finances. British Airways actually got fined $20 million by a regulator in the U.K. for losing credit card data and other PII. The last thing our customers want as they think about modernizing their networks is ending up in the news. We're in a really ideal position to help here. It's actually the number one focus for us in act one is API security. More than 50% of the requests that pass through our network go to APIs. That's over 23 million requests per second. As anybody in machine learning knows, the more data you have, the better models you built.
The boxes, the F5s, the Impervas, their virtualized equivalents, they just don't have the data that we do to build these security models to protect our customers against attack. It's also a great business opportunity for us as you think about expansion of the platform, and we think about this very much as a platform. All of the CDN customers we signed up many years ago have adopted APIs and are increasingly adopting APIs. That traffic is already in line with Cloudflare. It's passing through us. We can discover those APIs, we can help them build security postures, and we can do that much easier than any sort of upstarts. This is what we're doing with a large banking technology company here in the United States. They had to quickly produce a list of endpoints to provide to a regulator.
It would have been very laborious and difficult to do so and perhaps incomplete with the traditional solutions. We were able to produce this with no manual effort. After security on the API side, we're actually gonna turn our attention to the traditional API gateways, the MuleSofts, the Apigees, the Kongs, the ones that are even running on cloud providers. We're in an ideal position to start taking that management traffic as well. Beyond just leaking data, APIs can actually be used in a more direct way to abuse and drive up costs for our customers. These are fraud cases, maybe where you're sending an API call to initiate a wire, transfer cryptocurrency, maybe make a call to a company that uses a lot of GPU compute like an API. This can really drive up your cost here.
We help our customers identify and block these fraudulent requests using a combination of products like, our Bot Management, our API Shield that I mentioned, and the threat intelligence list that we produce and build directly into our product. Again, whoever has the best data builds the best models here. We're helping companies, payment networks like NCR block bad bots, reduce fraud, lower their cost to operate their service. The other key thing is here, we can do this extremely quickly, less than one one-hundredth of a second. The only reason we're able to do that is the design of our network. These machine learning inferences run all around the world in 285 cities, 100 countries. You have this kind of traditional trade-off between marketing and security.
I don't wanna lower my performance, I want to increase my security. We're able to help them do that. We're also within 50 milliseconds of 95% of the world, That's what enables us to provide that security with that performance trade-off. What do we do with all this data besides train ML models to build better products? We provide early warning of threats to our customers. What do you need to worry about? What's the guidance to reduce your risk? How can you partner with Cloudflare in any security question that you have? We have three key advantages over our competitors in this space. The first is that we operate really large-scale internet infrastructure. We want a public resolver that does more than a trillion queries per day. We have a certificate transparency log.
Every certificate for an encrypted connection that the browsers trust, we have a record of. We can see phishing infrastructure and other frauds spun up before anybody else. This gives us a really big leg up on the traditional threat intelligence providers that collect their data largely from after-the-fact breach investigations. First-party internet infrastructure, you'll hear more about this later. The second is we acquired a company called Area 1 last year. Area 1 had a tremendous threat intelligence team, in addition to email security, largely ex-NSA analysts and threat researchers, and these are people that are now accessible to our customers for any security questions they have. The third and probably most important point is our threat intelligence is actionable. It's not something.
It's really fun to, you know, get an email about an advanced persistent threat from some rogue nation state and, you know, read the story while you're drinking your cup of coffee. It's a lot more useful to have that automatically built back into your product. All of the detections that we find are built back into both our Zero Trust products, which you'll hear about later, as well as our application security products. Over the past year, we've opened up this offering to Fortune 500 customers and their SOCs. We're selling to new personas, and doing so has represented a new line of business for us, one you'll hear more about over the coming year. Lastly, from a use case perspective, we're increasingly helping customers adhere to complex and evolving regulatory frameworks around the world.
The power of running a homogenous network, not one that's cobbled together through acquisitions like in Akamai, for example, is that you can carve that network up and keep data where it needs to be. This is what we're doing with customers like PhonePe in India to adhere to evolving new data localization and residency requirements in 2022. The key thing about this is we take the pain out of it for customers. This is really a category we created. If you talk to AWS, they'll give you a really well-written PDF that talks about how you can install your own stuff in Amsterdam. We do this automatically. It's very, very easy to use, and it's a growing area for us. You can see a lot of products here on the screen.
This is the power of Act One is the platform we've built over more than one decade. Every one of these products you see here runs on every machine in every data center. It allows us to scale horizontally, reduce cost, avoid the performance security trade-off. What's more exciting is customers typically land in one particular area. They come to us because they had a DDoS attack. They came to us because their DNS provider is slow. They end up adopting more and more products over time once they figure out how easy it is. Instead of all those individual boxes that you have to learn, we give you a single pane of glass. At our IPO in September 2019, it was about 70% of customers use four products or more. That's approaching 90%. Really great growth there.
Looking forward to trying to accelerate that adoption and get people to use more and more of our products as they understand how easy it is. Let's talk a little bit about the competition. We break it into three buckets. The first group here you'll see on the left-hand side of the screen, these are the point solutions that target just one use case. Do an okay job of one use case, what we hear from our customers is they don't wanna manage all these different vendors. They don't wanna talk about all the different contracts. They also worry about the performance and the complexity of stacking these products behind each other. It adds latency. They come to us to reduce that complexity. The second group you can see here, these are the legacy networks and boxes that we replaced.
I spent a lot of my time meeting with customers. They probably put these in many years ago. They're looking to us to help modernize their application security and performance. How do you actually migrate to Cloudflare? That's an area we're seeing a lot of these days. These companies can solve a handful of use cases. The biggest problem with them is largely around the way they've built their networks from a performance perspective, their pace of innovation, which I'll talk about in a second, and the fact that they don't have a good compliance data localization story. Now that we're a leader in the key app security analyst reports from Gartner and Forrester, this has given companies the validation they need to start kicking out these incumbents.
10 years ago, when we were sort of providing additional security at the edge, we're now helping companies replace entirely what's downstream. The last group here is the hyperscalers. You're all familiar with them, AWS, Azure, Google Cloud. They address all of these problems, but they don't solve them in any way that we do. The biggest challenge with them is that they're essentially locking you into an ecosystem. If you go back to that first slide, the left-hand side, people are in transition to that, you know, applications leaving the data center. They want to maintain that independence and neutrality.
They wanna be able to choose the best product from each of these clouds and have a network that will put a uniform security posture at the edge and help them move workloads where it makes sense and at their own pace. We also win here. Our application security and performance products are superior directly, and we provide that optionality for customers to not be logged into a cloud provider. As I mentioned, they also have incentive structure problems because they're trying to get people to run these virtualized boxes, which pay for compute and storage rather than enhance their products themselves. There are a number of reasons why we win deals today. I'm not gonna read all of these to you.
I think the thing that I'm the most proud of is our product velocity and our pace of innovation. It's truly ingrained into our culture. You've heard us talk about this before. We ship products and put them in our customers' hands very early in the process. We have tens of millions of customers. That's also what powers our threat intelligence models, not just a small subset of enterprise customers. There's always somebody that wants to try a product early on, and so we get great feedback early. We stay close to our customers. We solve a problem for them. We build trust, and we come back to them time and time again.
Been at the company years years, products that I built eight, seven years ago that are successful with companies now, you get a lot of credibility coming back to them and saying, "Hey, this is on that same platform. You just need to check this new button. Let me tell you how it works. Let me get your feedback." Why we'll win in the future largely boils down to, you know, customers with the problems I described. They want to reduce complexity in their network. It's a conversation over and over again that we have. Our products work better together. As people adopt them more, they get smarter, faster, more secure, and this sort of accelerates that adoption.
The platform only exists because of the scale and the design of our network, you'll hear more about that later, and the data that it brings to us, that incredibly diverse set of customers. We reject that premise again, the performance versus security. The more apps that get added, the smarter it becomes, and I wanna tell you a little bit more about that. I'm gonna hand off to my colleague, Rustam, who leads our networking practice. Thank you for your time.
Thanks, Pat. My name's Rustam. I'm a VP of product focused on networking. I actually saw this morning, one of my cousins is in the audience also named Rustam. It's kinda cool. Let's talk about Cloudflare's global network, which underpins everything we do. You've already heard some of these stats, to repeat a few of the important ones. We're everywhere in over 100 countries around the world and over 285 cities, including mainland China. We're connected to every network that matters, we have a huge amount of capacity to support the world's largest workloads and to stop the world's largest attacks. Much of this is actually made up of our own dedicated private backbone. All of this adds up to incredible performance, no matter where you are in the world.
Our network has to be close to both the people and devices accessing the Internet, what we call eyeballs, and infrastructure, where customers are putting their applications. You're forcing customers to trade off between performance and security, as many of our competitors who have smaller network footprints force people to do. We're constantly focused on making our network even faster, more reliable, and more efficient. Using smart software and the data we collect on traffic flowing across our network, we're able to make things faster and more reliable, both in real-time, routing around congestion on the Internet, and over longer timescales, figuring out how to better interconnect with other networks and improve performance and reliability globally. In each country marked in orange, Cloudflare was the fastest network amongst our peers. This is a snapshot from September of 2021.
As you can see, we've been busy over the years, and our work isn't done until the whole world is orange. Initially, we built this network to be the foundation of our application services portfolio, what you just heard about from Pat. We did this instead of building on a public cloud provider for cost efficiency and geographic reach reasons. In act one, we're focused on protecting our customers' public-facing infrastructure, the applications that our customers expose directly to their end users.
Shopify to their shoppers, as an example. We initially focused on websites, but over time, customers asked us to provide the same protection we provide to websites to their entire Internet-facing networks. In response, we productize the capabilities we developed in-house to connect and secure our own infrastructure, allowing customers to use our network as an extension of their own.
Magic Transit was our first foray into providing network layer services to customers. Customers can bring their entire data centers to Cloudflare, enjoying best-in-class connectivity to the Internet and security from Internet-borne attacks. Transit plus security in one simple, magical product. Some of you might be wondering what the difference between applications and networks is. Let's go through a quick refresher. The OSI layer model is an abstracted view of how networks are built. Cloudflare offers products at all layers of this stack, from the physical cables, literal copper and fiber at layer one, all the way to the application layer products at layer seven that Patrick just talked about. Our Magic products cover any kind of IP traffic at layer three or the network layer.
Our network interconnect offerings enable customers to plug directly into us at layer one and two, providing higher reliability and a stickier customer relationship. Because we built all these products in-house and deliver them all from everywhere on our network, customers can combine them to suit their needs, composing products with others in layers above and below. This leads to natural expansion over time.
We see many customers join us at layer seven and then start adopting layer three products over the years and vice versa. Magic Transit competes with three categories of DDoS protection in the market. First, we still see a lot of customers with specialized hardware. These are Radware, Arbor, A10 boxes. These boxes are expensive, they can't block large attacks, and they require cumbersome software updates to add new kinds of protection. That's just not good enough to keep up with quickly evolving attacks.
Some competitors have developed DDoS protection as a service. They only deliver that service from a handful of locations. Akamai's Prolexic, as an example, lives in only 36 data centers around the world compared to our nearly 300. This means that customers need to backhaul traffic way out of the way and incur significant performance penalties when they come under attack and need to protect themselves. ISPs also offer DDoS protection, which is usually the worst of both worlds. They provide it via rudimentary hardware located at specific points inside their network. This gives customers incomplete protection and requires performance penalties when protection is turned on. In contrast, Magic Transit doesn't require any hardware or bad routing, and it's fully integrated with all of our other services for the best possible customer experience.
It's called Magic Transit because it connects customers to the Internet transit, and it truly is magical. Customers no longer have to worry about performance or security once they're connected to the Internet through Cloudflare. After DDoS protection, what else is left in our customer's network security hardware closet that we should go after and replace to help drive better performance, better security, better reliability, and better ROI? T
he next obvious choice is firewalls. Network firewalls are a $20 billion-plus market, and we've just barely scratched the surface of it. With Magic Firewall, we'll follow the pattern, the go-to-market pattern set by our WAF. We've started by augmenting customers' existing capabilities, stacking on top of them. We're advancing to taking out hardware in specific use cases like branch offices, and we're gonna expand to replace data center and branch firewalls entirely over time.
That's network services in act one. We secure and accelerate all of our customers' public-facing infrastructure, whether it's the application or network layer. We're doing very, very well in this market and there's lots of room to continue to grow. That being said, Magic Transit was a warm-up for the main event, allowing enterprises to build their entire corporate networks on Cloudflare.
We're taking on a brand-new market, SASE, in act two. As we listened to our customers talk about their existing network architecture and pain points, we learned a lot more about their internal wide area networks, or WANs, that keep their businesses up and running and often aren't exposed directly to the Internet. The tech landscape has changed a lot since these were initially built over the past few decades, and IT teams have tended to add complexity over the years with each new development.
Customers often draw us a picture that looks something like this. It's a really complex mess of point solutions that combines hardware with virtualization and cloud-based products that half solves lots of problems while creating even more new ones in security, visibility, troubleshooting, and cost. We're fixing this mess by consolidating all these point solutions into one unified platform.
Our customers connect their entire network to ours, and we add security on top and route traffic to its destination faster and more reliably than any other network in the world. This platform is called Cloudflare One. In a minute, you'll hear from Annika about how Cloudflare One connects and secures users and applications, allowing a seamless transition from a legacy castle and moat network architecture to one incorporating modern Zero Trust security principles. Let's talk about the network as a service that underpins that Magic WAN.
Our customers connect all the physical and virtual infrastructure that makes up their corporate network, whether that's offices, data centers, retail branches, manufacturing sites, cloud environments, IoT devices, and more to Cloudflare using Magic WAN. This completely replaces traditional forms of connectivity like MPLS and band-aid boxes like SD-WAN appliances. This directly makes networks faster to provision, cheaper to operate, easier to manage, and more reliable.
A new addition to this portfolio and ecosystem is the Magic WAN Connector, which Jen spoke about in the keynote this morning. This makes it even easier for customers to connect to us. Our customers have told us they don't want to spend hours configuring legacy Cisco, Fortinet, or Palo Alto hardware just to bring traffic to Cloudflare. They want a true plug-and-play experience and have everything just work. We're doing just that with the Magic WAN Connector.
It's a Cloudflare software package that our customers can purchase pre-installed on hardware provided by Dell and other partners. It connects them to Cloudflare's network automatically, straight out of the box, and makes sure their traffic takes the best path to us. Magic WAN is the next evolution of the enterprise network. It's easier to use, faster to provision, more secure, and lower cost than legacy MPLS or SD-WAN architectures.
It's fully integrated with Cloudflare One, allowing customers to buy networking and security from a single vendor with global reach. Customers are excited about these products and strategy, we expect a lot of growth here over the next few years. To drive home why we believe Cloudflare is differentiated in this space, the three components needed to win in SASE are Zero Trust security services. It's labeled SOC on this slide.
You'll hear more about these from Annika in a moment. You need a network to run these on, and ours is world-class, and a footprint within local customer networks or LANs. With Magic WAN Connector, we're now there. We've got all three pieces, and we believe we're best positioned to execute and solve enterprise customers' networking and security challenges. We're not stopping there.
There are several other projects we have cooking to set Cloudflare up for its next phase of growth, even beyond the acts we're talking about here. We're building a universal secure network that makes anything connected to it in the Internet more secure, faster, and more reliable. We're working on IoT and hybrid cloud connectivity offerings, as well as services directly targeting ISPs and telecoms that they can use to build and secure their own networks and operations.
In closing, we offer a global network that's always getting faster and more reliable. Our products are easy to use and combine. We never make customers choose between performance and security. Just like peanut butter and jelly, these two are always better together. I'm going to turn this over to Annika to discuss the Zero Trust side of Cloudflare One.
Thank you, Rustam. Hi, everyone. My name is Annika. I'm on the product team at Cloudflare. I'm so excited to expand on how with Act Two, Cloudflare is solving new problems for our customers. Rustam talked about how we're solving problems for our customers' networks. Act Two is not just about networks, it's also about our customers' employees, their devices, and their data. There's two big shifts that have created new problems for our customers' IT and security teams that are driving this move for us into our second act. The first shift is that our customers' applications left their data centers. They moved to cloud, they moved to SaaS, and that dissolved the concept of the corporate perimeter and really fundamentally changed how our customers' IT and security teams need to provide connectivity and security for their applications.
The second shift was that our customers' users left their office. Although some of us are back in the office, maybe some of the time, hybrid work is absolutely here to stay. Whether our users are working from home, from their office, from a coffee shop, IT teams are on point to provide secure and performant access to all of the applications that users need to get their jobs done. If you put yourself in the shoes of these teams, it's easy to understand the real pain that they are going through now.
Their entire world changed out from under them way faster than anyone ever expected, and the old architectures that they used to use don't make sense in the new world. Troubleshooting basic problems is now challenging. Organizations are dealing with data loss and other compromising events at a higher rate than ever before.
It's really challenging sometimes to even just answer basic questions with the fragmented world that our customers are left in with this technology. The work from anywhere model that organizations are expected to be able to work in now means that they also now have to support connect from anywhere and secure from anywhere. This is demanding a total shift in how companies think about keeping their workforce safe and productive. The new shift is creating this new architecture paradigm, which is called Secure Access Service Edge, or SASE. SASE meets the challenges of this new way of working by combining network connectivity and Zero Trust security services. Historically, customers bought products to solve these problems separately. They would buy MPLS from AT&T, and then they would go to Cisco and buy VPNs or firewall hardware.
As the lines between the corporate network and the Internet have blurred, it doesn't make sense to solve these problems separately anymore. Network and security teams can't keep working in silos. SASE is a cloud-based IT model that actually combines these things, software-defined networking and security functions, and then also delivers them from a single service provider. SASE offers better control and better visibility into our customers' traffic, their users, their data, regardless of where people are in the world.
Networks built with SASE are flexible, and they're scalable, and they're composable because we know that this change is not stopping now. Our customers are facing new kinds of problems every day, and they need an architecture that's more agile and more able to set them up for the problems that they're going to be solving tomorrow. Cloudflare One is our unified SASE platform.
By combining the network services that Rustam talked about with our flexible and powerful Zero Trust security tools, we can deliver amazing outcomes for our customers' IT organizations and improve the productivity of all of their users, no matter where they're working in the world. We do this by making things really simple. We have to deliver an amazing user experience. People have higher and higher demands for the performance of the connectivity to the applications that they use every day. We also need to make it easy to fit Cloudflare One into our customers' existing environments. They're not going to rip and replace everything in their existing architecture all at once and switch to SASE. We have to work within the tools and the context that they have today. Simple and then consistent.
We need to provide security controls that work the same way regardless of where our customers are connecting to us. Again, if you're in an office or a coffee shop, you should have the same security tools and filters applied to your traffic. We also need to provide consistent visibility. The analytics, the logs, the way that our customers understand the data flowing through our systems also needs to be in one place so that they can provide this comprehensive view of all of their systems as their traffic is running through our network. The simplicity, the consistency that we're able to drive with Cloudflare One results in really tangible business outcomes for our customers. It saves them time and money and reduces risk.
We're solving these customer problems as a unified solution. Let's dig in and break down some of the specific reasons that customers come to us and start talking to us about Cloudflare One. Rustam already told you about one of these, site to site connectivity. Previously, customers used to solve this problem of getting traffic in between their branch locations or maybe a retail store and a data center with things like MPLS and leased lines. It doesn't make sense to pay for these expensive forms of connectivity anymore because a lot of the traffic has now shifted to take the Internet. There's still some of that site to site traffic that remains, and our customers need a solution for that.
Cloudflare One delivers the solution with our Magic WAN product. On the employee side, the user and device side, one of the first problems that our customers need to address to enable work from anywhere is performance and reliable access to internal resources. Historically, customers have solved this problem with VPNs, but that's honestly never been a great solution. If you've used a VPN, you know how much of a pain it can be. It got the job done when remote work was limited and the resources that people needed to use using the VPN were also limited. Now users need fast and secure access to ton more applications from anywhere in order to stay productive, those VPNs are not cutting it anymore. Site-to-site connectivity, internal resources. Next, what about everything else? Access to the rest of the Internet.
How do you make sure that the applications that are hosted externally that your user are accessing aren't malicious? One of the ways that customers have solved this is by installing traditional hardware appliances, secure web gateways, and then backhauling traffic from customer locations or wherever their users are to those appliances. Hardware is hard to maintain. It requires customers to send their traffic out of its way to get access to that filtering, and it introduces performance penalties, which again, isn't acceptable anymore in this new world. Controlling access to internal applications, external applications, that's one part of the battle, just getting your users access to the things that they need to do their jobs in a secure and performant way. What about how the apps themselves are used and configured?
Our customers need visibility and control over their SaaS applications to detect things like misconfigurations, exposed files, suspicious activities. This is more important than ever. As we heard earlier today, more of our customers' applications are moving to SaaS, and employees depend on those to get their jobs done. We need to be able to keep them secure. Then finally, one of the scariest problems for customers in the landscape of their corporate data. They need solutions to view and control who's accessing their data from when and from where, and it's absolutely critical that our customers are able to prevent data leaks. This is the landscape of problems that we are attacking in Cloudflare's second act as a company. Again, we're doing this with Cloudflare One.
Our customers need to solve all of those problems individually, but they also wanna understand the roadmap to adopting a SASE architecture more holistically and solving them together. That's why we've built this as a unified and composable platform to help our customers transform their network one step at a time. There's three pieces to this platform. The first one is getting connected. We have a variety of what we call on-ramps that give our customers access to the closest Cloudflare network location to them easily and quickly and securely from any device, any application, even a physical data center. They can plug right into us if they're in the same building. Then once traffic is at our network, we layer on Zero Trust controls to modernize their security approach.
We've talked about Zero Trust a lot today, just as a refresher, what does that concept actually mean? It means that users are no longer inherently trusted to access applications just because of where they're sitting, maybe on a physical network or historically on a virtual network with a VPN. It also means that web applications, email links, and user interactions are also not trusted to reach companies' data or devices just inherently by being on net. That's Zero Trust controls, piece one, on-ramps. Piece two, controls. Piece three is the Cloudflare network, which means that protections are always delivered to our end users regardless of where they are in the world with one control plane to reduce their operational overhead. Who else is in this space and solving these kinds of problems for customers?
As we saw applications shift to that cloud, the first shift that we talked about at the beginning, we saw early signals of companies embracing remote work. There were first generation Zero Trust providers that emerged to start solving these problems in new ways. What we've seen over time is that these vendors weren't actually prepared to solve the really intertwined problems that our customers were facing. This is for a couple of reasons. One is that they made fundamental architecture decisions to only solve one problem at a time. They literally built separate networks for separate products. Now when customers have use cases that combine those, it's really challenging to do that.
They also often deployed solutions as virtualized versions of hardware appliances, which sounds great in the moment, when you think about the problems of traditional hardware like capacity planning, location planning, figuring out how to scale, it doesn't make sense. It doesn't work in the cloud-native world. Some of these competitors acquired companies to stitch functionality together, they're having the same problem of trying to figure out how to get all of the puzzle pieces to fit. Some also made bets on partnerships to deliver different parts of this SASE solution, which worked for them in the short term, it made it really hard for them to build out the rest of the pieces in a more integrated way because they were directly competing with companies that were generating revenue for them.
Very few of the companies on this slide invested in their own underlying network infrastructure, which again, is critical. Let's just dig into that why. What does this landscape look like now for customers that are trying to solve this problem? Well, if you have someone that's looking for SASE, they have basically two options. One is they can buy security and networking services from two different vendors and try and put those solutions together.
They can pick a vendor that has all of the pieces and has done that first step for them, but really delivered that same experience of just kind of piecing these different solutions together. These approaches don't work. The reason this doesn't work is because it leaves gaps in security. It leaves gaps in visibility. You have different tools that your data is stored and maintained in.
Deployment and maintenance is harder with all these different pieces. It's not just incrementally harder with each new tool you add, it's actually exponentially harder because every new tool means you gotta integrate with all of the rest of the components in your stack. It's also really challenging to troubleshoot problems. Customers report issues with backlogs of IT tickets that are just really basic questions like, "Why can't I get to Zoom in the office?" Even if our customers do manage to deploy SD-WAN and SSE solutions, Security Service Edge, that Zero Trust security part of this portfolio, and they figure out how to kind of piece them together, they're still missing the underlying network fabric. This either leaves dependencies on those legacy forms of connectivity that we talked about.
They're still maintaining MPLS lines, point-to-point lines in their branch locations, or it subjects our customers' traffic to the unreliable and unperforming conditions on the open internet. In many cases, our customers are now managing the cost and the headache and the overhead of the legacy network technology, plus all of these point solutions kind of layered on top of it. Cloudflare One is the best position platform to solve these problems for our customers holistically.
We have a natively developed Zero Trust security solution. We have a next generation WAN as a service that Ruslan told you about to get our customers connected and secure. We've delivered it all on the same network that already solves all of these adjacent problems for our customers that Patrick told you about: DDoS mitigation, web security, email security. Let's again click into that last piece.
Why does the network matter? Why is it better for customers to get these solutions from us versus kind of piecing them together? The first reason that the network matters is speed. Let's look at Cloudflare's network compared to a couple of our competitors. There's two things I want to highlight on this slide. One of them is that we are everywhere. Customers that switch to Cloudflare from Zscaler, for example, report immediate performance improvements, and that's just because we're physically closer to wherever their users are in the world. We're literally battling the speed of light, and we're winning. Second, we don't just have dots on this map. We also have our own global backbone connecting those dots. This backbone acts like a super highway for the Internet.
When there's congestion, traffic jams, potholes, package drops happening on other places on the Internet, our competitors are just kind of stuck. They're at the mercy of these conditions. For our customers, we can just get their traffic on our global superhighway and get it to its destination reliably and on time. The first reason the network matters is speed. The second reason is composability.
It's not enough just to have dots on the map or even the backbone connecting them. For our competitors, many of the dots on those previous slides actually only run a subset of their product or a product for only a couple of their customers. This means that in order to get your traffic running through multiple levels of security filters, it has to actually hop around between locations. This introduces more reliability, scalability, and performance trade-offs.
Because of this, many of the customers I talk to have chosen to sacrifice layers of security protection in order to preserve a user experience, or they feel like they're secure, but their users are having a really miserable time. With Cloudflare, because we run every one of our products in every one of our locations, those trade-offs don't exist. A packet that lands at a single Cloudflare server can run through those full stack of security tools before we send it off in an accelerated manner on its way to its destination. Speed, composability. The last reason the network matters for SASE is because it acts like a massive global sensor system.
We see every request and every packet that transits our network across an incredibly diverse set of users and organizations that are using us for our Zero Trust and SASE solutions, but also those Act One solutions that Patrick told you about earlier. The network sends signals back to us and helps us identify where we have network congestion or emerging security threats, like new types of DDoS attacks. Then we fold that intelligence back into our network instantaneously and back into our product to keep making it stronger for all of our customers. Of course, our focus is helping build a better internet for everyone and in conjunction with the community.
We also work closely with a variety of partners and threat intelligence providers around the world that help us augment the intel that we already get from our own network and put it back into the product as well. We have a team of industry-leading threat researchers that Patrick mentioned that are working on the broadest data set, which includes email, web, and network layer attacks. We combine our own intelligence and the intelligence from our partners to make our network stronger. We're proud of the foundation that we've built for Act Two, and we have lots of room to grow in providing security and protection and performance for our customers' users, their devices, their employees. We're also not done. Our speed of innovation, pace of innovation matters.
This space is evolving so quickly, and we'll continue to build quality products in this Act Two space at speed. That innovation is also how we excel into Act Three, by helping our customers and their developers do the same thing. With that, I'm going to turn it over to Aly to talk about our developer services.
Hey, folks. Thank you, Annika. I'm Aly Cabral. I run our developer platform team, the product team for the developer platform. I'm gonna first take you back in time. All right? Back when Cloudflare was just focused on Act One. This was a business, a portfolio of services that was about taking existing applications, putting Cloudflare in front of them, and making those applications more reliable, more secure, more scalable, more performance.
All of those things that really take something working for you on your local desktop and making it working at scale for production use. At the same time, we've always been an ambitious company. We were looking for the next act that could use that same investment we made in the network in different areas, different adjacent areas. Every time we started to invest in that new product development, we found ourselves solving the same problem again and again and again. These problems range from geo load balancing to cache optimization to disaster recovery. Every time we were building an application at production scale, we would build that application, and then we would think of these things. It just takes time, and you're revisiting the same problems again and again and again without much different data from application to application.
The ethos here was about taking those learnings that we already have and embedding that into a platform on day one. At the time you hit deploy, at the time you've actually built your application, we solve these problems for you out of the box. Now, at the same time, in our Act One business, customers were asking us for customization. That customization could not come with the price of sacrificing the value of Cloudflare, right? That value is security, scale, reliability. It needed to still have all of that magic that Cloudflare had baked in, but with custom logic directly deployed adjacent to those services. Now, with that, those two goals, us building for ourselves and that customization request of Act One, building Workers felt natural to us.
This is our serverless platform with opinionated ways to deploy at scale performantly, reliably by default out of the box. We also talk about our pace of innovation. You guys know it's really hard, even internally, to keep up with the features and products that we ship as a company. This is the magic sauce that our customers also wanted for themselves, whether those customers are startups or enterprises trying to keep up that pace of innovation with startups that are challenging them. This is really the genesis of the Act Three business. What's great about it is it enabled the Act Two business. It really enabled us to move as quickly as possible. All right. Let's talk about problems application developers have regardless of what application they're building. The first is production readiness, right?
There's an adage in software engineering where it takes 20% of your time to get something working for yourself locally, and then 80% of your time getting something deployed at actual scale for hundreds of thousands and millions of users. Why is that? Well, it's not actually because your differentiation changed at all. It's because infrastructure management is hard. Every time I'm asked to build an application, I'm asked, "Oh, well, what if this data center goes down? What do I need to do in that case?" Well, those problems aren't different across company, across application. It actually makes sense for the system to take on some of that decision-making, some of the intelligence Cloudflare already has, bake that into the platform, and ultimately both make better decisions and save developer time. We always talk about pace of innovation.
We know that this is something that companies feel stronger now more than ever with a lot of noise in the market. You've got to keep up with the pace that your users expect. The third one is vendor lock-in. I actually like to think of vendor lock-in as architectural freedom. The downside for many companies of being locked into a single vendor is that you are kind of locked into the future of your application a bit. Like, let's say I built an application that's an e-commerce business, and in five years I didn't predict this when I first picked my first cloud provider, hyperscaler, or whatever platform. I want to use a machine learning product.
Well, if I'm locked into a single provider, I may be forced to use a lesser product than would actually make my product better and differentiated in the market. When people talk about no vendor lock-in, they also want to retain flexibility for future development in the future to choose the best tool for the job and not be locked into lesser tools just because they happen to be in the ecosystem that they chose years ago. Finally, the last two, global audiences and high user expectations.
Application software has never been more global than it's been before today, right? When you deploy applications, those applications can be used and served across the globe. Then higher user expectations. Like in the mainframe days, maybe it was okay if a website didn't load for a bit, and you'd come back in a minute.
Those aren't the user expectations anymore, right? The expectation is that you have reliability baked in, you're planning for disasters as a development team, and your job got harder on the infrastructure side. At the same time, the system hasn't yet got smarter. Cloudflare sees that as an opportunity. What's great from our perspective about these problems is these are not unfamiliar problems. We talked about the Act One business, sorry, taking existing applications and making them more production ready. This is what we mean when we say more production ready, right? Scale, reliability, you get all of that baked in with the Act One business.
With Act Three, the idea is taking that expertise that we've built up in Act One, the foundational technology that we have a lot of data around that's hardened, and bringing that earlier in the development life cycle. Around that plan, build, deploy stage, instead of once I've built my thing, then I'm going to go to Cloudflare to scale it out. All right. What I love a lot about the Act Three business is it's not just a great business on its own, it also amplifies our Act One business. Why is that? Well, we talked about vendor lock-in, right? Let's say that I decided to build my application on AWS, and I get to that scale and secure stage.
Well, their networking products, the things that are actually gonna take that application that I built and scale it out for good user experience, it's actually, like that in ecosystem option is actually a lesser product than the Cloudflare product. We have the opportunity when people are building on Cloudflare from day one to also get more business on the Act One side. All right. Let's take a moment to talk about the aspects of application development and how they kind of break down at a high level. You need these parts of, like, development in order to service really any kind of functionality. The first is compute. Compute is the brains of the application.
It has all the instructions to send that information to and from users, locations, wherever in the world. Then an important overlay to those foundational building blocks that application needs is developer experience. That is really around products and tooling that allow developers to save time and ultimately make better decisions. The good thing is we've been focusing on these core tenets across our stack for many, many, many years. On the compute side, we have our Workers serverless products. You actually heard on the data and storage side, you heard MLB talk about Durable Objects, which is a distributed system state paradigm that we actually created at Cloudflare. And Durable Objects now powers those live updates you see at MLB games, right?
That kind of innovation for real-time collaboration and updates is something that we're really proud of on that state management directly on distributed system. We have our developer services business. This is about taking an opinionated approach to specific use cases built on top of the foundational compute and storage products. We talk about Pages. Pages is an opinionated way to build Jamstack applications. Stream is an opinionated way to stream and store and serve videos wholly built on top of the foundational compute and storage products. They're opinionated ways to build for those specific use cases, saving developers time for those specific use cases. It's not enough to just have the foundational technology. Let's talk about the opportunity in the space.
Break it down into these four categories, where the foundational bits are compute, networking, and storage, and then there's that overlay of developer experience. We all know the hyperscalers play in the space, right? Around compute, networking, and storage. They have services across the many, with many offerings. We also see challengers in this space that are development platforms, but wholly built on top of the hyperscalers with the high value add of developer experience, right? Allowing developers to move faster. If that's not a business case that there's a gap there in the hyperscaler developer experience, then I don't know what is, right? That these successful developer abstractions are in and of themselves good businesses means there's an opportunity for Cloudflare to do all of these things and sit in the middle of this Venn diagram.
If you think about the incentive structure of the hyperscalers too, you have these companies that are wholly built on top of the cloud, like the developer abstraction companies like the Netlifys, Vercels, PlanetScales, they're actually not incentivized to directly solve the developer experience problem, 'cause if those platforms win, they win too. We think that if we have a focus on developer experience and a focus on the foundational technology, we have a real opportunity to solve both. Awesome. I wanna talk about the movement and investment into being more of a data storage company. We, I think, can't do that without touching on R2. R2, the model is innovative, right? Zero egress fees. That leads though to architectural freedom. One of the things we talked about is how AI companies are being built on top of Cloudflare.
R2 is a very popular product with generative AI companies. The reason that is for many reasons, but Like, let's talk about one of the core tenets of generative AI right now. There's actually a GPU scarcity in the market. You guys might know this. We're seeing that these generative AI companies are needing to go to either multiple providers or multiple regions in a single provider to get the GPU capacity to run expensive models and do expensive training. Every time you're sending large amounts of data, even multi-region within a single cloud provider, you're incurring a aggressive tax, a commercial tax, right, in the form of an egress fee. R2 is an offering that opens up and gives these generative AI companies the flexibility to find those GPUs wherever they can on any provider.
This is not just a cost efficiency problem. This is a, how do I make my business tenable? This is definitely a core component of that's bringing companies that are starting up today onto R2. Other highlights of R2, region automatic. We wanna get. You've heard me talk about getting these previously developer decisions and making them system decisions. We're also trying to do that when we talk about data placement. When you create a bucket in R2, you're not selecting a region. We're taking into account application heuristics and deciding on where that data is placed for the user, for the developer. You can overlay things like jurisdictional restrictions for compliance needs if you have information we don't about the compliance needs of your data. We also, I just wanna touch on Workers integration.
From an R2 perspective, you do not need to use R2 and Workers together. They just work really well together and amplify each other. You can decide to use Compute somewhere else if you so desire, and still use R2 for your object storage, or if you use them together, you can get these high-performance application use cases. Finally, just wanna touch on S3 compliant. We wanted to pick a familiar API and make sure that it was as easy to come onto the R2 system as possible. Now, briefly touch on just some use cases for R2. Static assets power the web. Those are a large percentage of data stored on object storage today. Media files, large files, of course, take up a significant percentage as well. Logs.
Every tool that you use as a company generates event history about how that tooling was used. Cloudflare is one of those, Tools and software that creates a lot of logs. Of course, it's a no-brainer to store that information, those events, and do trend analysis on top of Cloudflare. Finally, this data platform use cases, working with data marketplaces to offer R2 as a way to give their customers, architectural freedom and liberating their data is something that we're highly focused on .
All right. Across the stack, developers are choosing us for many different reasons. On the compute side, this is the infrastructure management, saving developers time, being scalable, reliable, secure by default. On the storage side, that opening up that, architectural freedom, data liberation, all of that is incredibly important when deciding where your state is going to live.
On the networking side, this is our bread and butter. This is the best-in-class offerings we have. They're the best performance, most reliable, most scalable offerings on the market. The developer experience side, that is gonna continue to be a huge investment for Cloudflare, whether that comes in the form of Wrangler, which is our opinionated CLI for deploying and managing applications on Cloudflare's network or Pages, right, which is an opinionated service for website development.
We're gonna continue to invest in products like that. Now, if we talk about how we compare to the hyperscalers today, right? It's not an apples to apples comparison, right? We have the strongest networking products, that's for sure. On the compute side, we are specific and opinionated about how to build. We're not just playing apples to apples on the low level compute side where it's a commodity business.
We are making sure that scalability, reliability, all of that is baked in by default. With that, we are asking developers to come onto the platform and change the way they build. On the storage side, we're gonna continue to invest and become as build out more and more types of models to support all of the different types of use cases. On the databases side, we have D1, but we're just getting started. Our goal is to have a fully global database on top of the Cloudflare platform. In developer experience, we're gonna continue to invest and be strong. Data localization, this is always a high priority for us, and we offer tooling like jurisdictional restrictions and data location services to make sure that application developers can take advantage of it out of the box.
Our whole ethos, right, is to avoid that vendor lock-in by choosing models that are friendly to developers and allow them to build applications that make the Internet better. Now, we all end on this slide by really talking about the ethos of Cloudflare and the foundation we're all building on top of. That really first starts with the network. Then ease of use is the same ethos across all of our different apps. Shared intelligence, we're taking data that we have and running 20%-25% of the Internet and feeding that into better decisions than developers could make for themselves. Finally, no trade-offs from a cost and performance perspective because we know that the Internet is better if we make it more performant. With that, I'm gonna pass it off to Marc, our CRO.
Thank you, Aly. What a great presentation. I mean, Aly, Annika, Rustam, Patrick, amazing products. I wanna start by just saying that. That's why I'm here. Okay? Just sitting through that presentation, it reminded me of my path to Cloudflare. By the way, as you just heard, I'm Marc Boroditsky. I'm the new president of revenue here at Cloudflare. Before we get started, I wanna take a moment and address a misunderstanding that took place on our earnings call last week.
You may have heard that fish are jumping in our boats, and we discovered that some of our team members don't know how to fish. We also adjusted our guidance last week. This was due to the fact that we were seeing softening in the market. It was showing up in the form of elongated sales cycles, specifically in Q-Q one.
We took the step to adjust our plans for the year, assuming this is gonna continue for the rest of the year. I wanna be clear, our go-to-market, pardon me, our guidance and our announced go-to-market talent changes are two separate issues. Since I joined in Q4, we have been looking at performance and, more importantly, non-performance of our go-to-market organization. We have been actively implementing a plan to improve the performance. That's why I'm here. In Q1, we saw a widening between our high performers and our low performers. It's been difficult to get the data right, but it's irrefutable, and we recognize that now is the time we need to start evolving our talent strategy.
With that out of the way, I'm now gonna dive into and share with you what I have seen over the last six months and what we're doing to improve our performance. The next presentation by Thomas will go into the details regarding our guidance. Let's start with the positives. All the presentations we just saw, including the earlier keynotes, that's a great framing. I've known Cloudflare since 2014 when Matthew and Dane were my customers. I was then at Authy and then subsequently at Twilio. Over the years, I have marveled at Cloudflare's extraordinary product stack and the amazing pace of innovation. This innovation engine has been continually increasing the ways Cloudflare helps customers.
Every time I go to a customer meeting, I'm surprised by the kinds of things that we can do, whether it's in the box or an idea that we're having right then and there, and oftentimes opening up to new opportunities. The customer value of our solutions on our platform is significantly more than any combination of competitive point products in the market. That's great for salespeople.
When you're looking at this chart, this is an amazing opportunity. Clearly we are not afraid to enter new markets. We might not start in the upper right-hand corner, but we do move quickly up and to the right and ultimately lead the Leaders Quadrant. Today, Cloudflare is a leader or a strong performer in nine analyst reports. We should be able to say and agree that we produce world-class products. Again, an amazing opportunity for sales teams.
This expanding product and customer value has increased the opportunity in the market, and you could see it in terms of the pipe that we have multiplied over the years. You should be able to see that it's growing on average 50% per year, representing the interest in our technology in the market. Again, great for sellers. This growth and momentum is not just isolated to the SMB customers the company started with. This is really important. Our platform is being adopted by high-growth digital natives as well as major established legacy enterprises. Like a big pharma company that is using our Zero Trust solution to today protect 5% of their applications, and then met with us recently about how we can actually expand to the other 95%.
The group of LatAm customers that we hosted at RSA recently that were excited by what they could do with our platform. Like a major retail energy company and a major consumer bank that were riffing off each other and got excited about the possibility of moving their networks to our cloud and removing a ton of unnecessary overhead. A leading fitness company that's in the planning stages of releasing new online products, and they're looking to us to help them to protect against fraud using our edge. Our top salespeople get this and are regularly delivering deals like the ones I just described to generate outsized results. They're incredibly consistent at bringing in new logos and expanding existing customers. Our top 15% performers on average have achieved 129% of their target for the last five quarters.
They've also done that without experiencing the level of degradation that we saw in the broader market as a result of the lengthening sales sales cycles. These are not OG old-timer sales team members. About a quarter of them actually joined the company in the last 18 months. I am extremely proud of Cloudflare's consistent high performers and excited about the model they represent. Okay. When I joined, Matthew, Michelle, Thomas, and I all knew there was an opportunity to optimize Cloudflare's go-to-market.
This is not an uncommon situation for high-growth, product-led companies. Actually, it was what I was looking for when I left my last gig. Earlier this year, Matthew gave me the privilege to come to the board and share my unvarnished insights that I had gathered on opportunities for improvement, and he asked me to share them here today with you.
By the way, I love this kind of transparency, so consider this a privilege. Let's set the context. Cloudflare is not limited by TAM. Across our four main product categories, there is $146 billion worth of addressable market. Obviously we are less than 1% penetrated. The argument that we're running out of market is wrong. We are not limited by opportunity. I just shared with you the march of increasing pipeline that we've been generating over the years. Clearly there's opportunity in the market. We are also not limited by capacity. If we were, the conversation about our sales team wouldn't have happened. They'd all be crushing their numbers. Let's take a closer look at capacity.
As I mentioned earlier, there is a strong cohort of high-performing AEs that have routinely been crushing their numbers, and there's also a group of low performers. This graph that you're looking at is our recent distribution of AE performance. If you were looking at this just based on the average, you would assume everything's okay, but the distribution is not. This is classically known as a bimodal distribution. It happens when there's two things happening or being observed in the data.
We can assume that our offerings, market opportunity, enablement, and management are fairly uniform. Where there are idiosyncrasies, we've accounted for them in presenting this distribution. What's the difference? The difference is talent. Talent tops the list of opportunities for improvement that I shared with the board. I observed that hiring, enablement, performance standards were relaxed or nonexistent. We also didn't really understand our customer journey.
We were not measuring it in a way that was actionable. Teams had different metrics, and in some cases, measuring them differently or incorrectly, and there were data issues in understanding what was taking place. Our sales coverage was a series of snowflake models unique to many different situations, making it very difficult for supporting teams to plug in. When it came to actually assigning accounts, there wasn't great data to figure out the value, and there was more of a legacy ownership, illogical approach to account retention. Didn't mean that we were consistently assigning the best person for the opportunity or creating an equitable opportunity for the next hire that we were making.
Pipeline was foggy in the way that we tracked it with our tracked it in terms of our attribution model. There were some challenges associated with the way we were routing and getting opportunities to account executives. Also compounding this is the fact that Cloudflare is one of the most complex places for a seller that I have seen. By the way, this is not uncommon. Many problems had to be solved over time. Point solutions were applied against the problems that were being faced. This is not a, an artifact or an outcome because we have a lot of products. This is an artifact or outcome because the systems and tools just haven't been built yet for the scale that we're operating at. As a matter of fact, an AE needs to use more than 20 different systems to do their jobs.
We have a range of pricing metrics, overlapping packaging, different discount and margin guardrails, all of which the AE is arbitrating. Creating an order can take over 100 steps. I have seen this before at other big, much bigger companies. There's an opportunity, obviously, to improve this and make it possible for the AE to get that time back and spend it with customers. These issues are all easily fixable. These are solved problems. The high performance I shared earlier, we all should be proud of because it was actually accomplished, generating $1 billion in spite of these limitations. Imagine what's possible when we correct these issues. Let's dig into that. Let's talk about what we're doing. A lot of times, us in the tech industry, focus on the features and functionality of our products.
We think of the speeds and feeds and the performance characteristics, but when you really dig into it, you can't make the products, you can't sell the products, and you can't support the products without great people. I believe, at the end of the day, we're really in the business of finding and keeping great talent. If you do nothing else but hire great people, you will crush it. For sales, this is a massive opportunity. As you heard Matthew say on our earnings call, our AEs in Q1, our low-performing AEs in Q1 only delivered 4% of the results. If you take a look at our A-AE productivity, our low performers came in more than 80% lower than our average performers.
When you combine these two data points, you can see that by moving our performance up to the mean, there is a big step up in opportunity. Actually, a bigger step up in opportunity than moving our average performers to high performance. Not to mention, hiring high-performance people is really expensive. What we're striving for is a more normalized distribution, with 80% of our AEs attaining 80% or more of their targets. This performance means we're actually aligned to the population and provides the greatest potential efficiency. Think of it in terms of hiring hundreds and eventually thousands of people. You need to match the population so you can make average humans highly successful instead of always expecting you're gonna get the top performer to show up.
It requires hiring great people, it hires enabling them effectively, and it requires that we have great coaches and managers and leaders supporting them to be successful. By achieving this normalized distribution, we have the opportunity to generate a 20% bump in average AE productivity just by moving to this distribution. This will give us confidence to continue to invest in the improvements that we're gonna make, but more importantly, to be able to deploy added capacity knowing we're gonna get the results that we're expecting. These improvements are not crazy. I've actually seen in previous experiences a 120% increase in a five-year period. How are we gonna do that? You can't let me get away with just saying productivity is the answer. You need me to explain how we're gonna drive productivity.
Great productivity is i t isn't easy, I'm not gonna pretend, but it's also not rocket science. Go-to-market activity can be broken down into a series of processes that can be organized, measured, and continuously optimized. In the past four months, we have already been executing our plan to ensure that all aspects of our go-to-market organization are working together to deliver a powerful and productive revenue engine. We are adding leaders that have experience at executing a playbook of this nature with other companies at scale. We have five key priorities that I'm now gonna take you through. Our first priority is building the regional revenue machines. We have implemented a standardized structure based on global region, country, and segments led by regional GMs.
We've added coverage with an AE as a quarterback, and if you're not American, AE as a conductor model that has them responsible for all account activity, including new up-sell, cross-sell, and renewals. They are being supported by a team based on a consistent ratio-oriented model. We are expanding the specialist coverage that we're providing to the sales team, for the personas and products that they're supporting.
As an example, security specialists that support our Cloudflare One sales to the security organization, our network specialists that are supporting the sale of Magic Transit, and our developer specialists that are gonna be supporting R2 and Workers. By the way, what good looks at? What are we tracking? What's the metric we're using? Increasing productivity per AE. Number one priority is to get organized and methodical and continually improve the productivity that we're generating.
Priority two is sell the way the customer buys. Sounds simple and straightforward. Well, we are working on evolving to be a more customer and market-driven company in the way we execute. By using data to understand the customer experience and help drive sales and adoption. By also mapping our solutions to customer pain points and aligning our product categories to the key categories the customers are looking for. It's like the way that our Zero Trust team today is now positioning our solution, Cloudflare One, against the work from home requirement. We're also elevating our partner program so that we have clear guidelines about where and when you engage a partner to be successful. Success in priority two is meeting our pipe coverage targets. three, become trusted advisors to our customers.
We are leveling up our talent with clear expectations, enablement, and performance standards as trusted advisors, so they'll understand what's expected of them to be successful at Cloudflare. We're also evolving from selling tactical point products to positioning the Cloudflare platform and solution categories. Our partners in marketing have already delivered a great set of solution pitches, and we've already rolled out to the sales team a certification program where everybody is now certified on our platform pitch.
Success in priority number three is high customer satisfaction for all touch points. Number four, partner with product. Our product brethren are phenomenal at what they do. We have built bridges between our organizations by creating solution leadership teams that now has us operating a shared ownership model. What does that mean? We're actually pursuing the same outcomes, revenue results on the products that we're selling.
I gotta tell you, it is extraordinarily clarifying to be sitting with our product counterparts and discussing pipe gen, pipe coverage, AE enablement, and how we strive to exceed targets over and over again. Success in priority number four is exceeding our product revenue targets. Number five, operate as one revenue team. Every team in my organization now has a clear set of priorities that is based on the five that I just went through. It includes targets and expectations so that every team knows what they're driving for and what the results need to be. We've established a global single rules of engagement. Think of this as the operating guide in terms of the way that we operate in order to ensure that our execution is occurring the way we expect it to.
We've also established one consistent cadence throughout the entire organization, and we've centralized communication and engagement with the sales team. I gotta tell you, we're already seeing great results. Like when we just went through this talent improvement planning process, the people team, the finance team, our enablement team, key sales managers that we already know and trust, came together automatically to organize and execute on the program with great effect. Success in priority number five is great engagement scores. To wrap up, we are upgrading and investing in talent. We are becoming more systematic in our processes and operations in order to be more efficient and more productive. We are adding leaders with experience at executing models like what I just described at scale. There's no question there's a lot of work to do.
I'm confident that in addition to being world-class in product innovation and world-class in network stability and reliability, we will be world-class in go-to-market. We don't deserve the trophy yet. There is no better time than now to join the Cloudflare sales team. With that, I wanna pass it on to my colleague, Thomas Seifert, to share that extra insights. Thomas?
Thank you, Marc, and thank you everyone for joining us today. Now, if they're honored to keep your energy level up at the end and try to tie it all together, what you heard about products, about market opportunities or about different acts, and how we think this all comes together in a story of growth and profitability. Of course, after last week, I also have to touch upon guidance. I want to start this discussion with getting us all on the same page in terms of what we have achieved so far. I want to talk about growth and about our profitability on our journey moving forward. I think the best reminder of getting started is really the showing the path that we have come along.
An incredible innovation engine, as you have heard now in multiple presentations, is driving us to attack a TAM that has become massive. $146 billion, up almost 5 times since we IPOed. We've been on this journey acquiring larger and more enterprise customers, finishing currently at north of 2,100 large customers. Just a reminder, a large customer for us is a customer that makes more than $100,000 of revenue. That brought us to a run rate of $1.2 billion in revenue, up more than 3 times compared to our IPO. We are able to deliver this revenue at a gross margin of 76%. I hope after today, you get a good insight what drives this efficiency.
It really comes from the network we have built, standard hardware in many locations, one homogeneous software stack that allows us to offer all the products we have and all the service we have in every location. That is a really important statement to make when we talk about unit economics later. It's also quite an achievement to keep this margin at this level and north of 75% despite the increase in traffic, despite the increase in products, despite the increase in locations we are serving.
We've turned profitable and we've delivered record operating margins and record operating profit in the last three quarters, showing you that we can use the efficiency that is in this model to pivot rather fast to profitability when market requirements require it and when the market gets tougher. With this being said, the opportunity is significant.
You heard a lot about the four categories today that we are addressing application services, Cloudflare One, developer services as a new addition, you heard this from Ali's presentation, and network services. A huge market opportunity, almost five times up in terms of the addressable market that we are going to disrupt from where we were five years ago, all delivered on a single platform. Addressing this market, we have delivered consistent growth now over the last years of close to 50%. Durable growth, after today, I think you get a better understanding of where this durability comes from. Matthew always talked about winning this war of attrition, putting our products in front of on-premise infrastructure and taking upgrading and investment dollars into our budgets over time.
We still have only penetrated 1% of the TAM we address. An incredible journey so far. What made this revenue growth so durable is also it's driven by multiple independent growth vectors. It's the number of products we ship. More than 50 generating products at this point. Larger customers that adopt more products. I will give you a lot of insight over the next slides, how the face of a large customer has changed in terms of size, in terms of products and services such a customer consumes, and how we are doing selling our new products, our Wave Sets, two and Wave three products into this customer base. As such, we remain really confident about the opportunity that is in front of us. We've also to acknowledge that something's changed in the first quarter.
That negative sentiment came up after SVB. We saw significant impacts in customer behavior. They are becoming significantly more cautious, which impacted a couple of areas, as we mentioned in our earnings call last week, but a significant acceleration of our sales cycle. In typical Cloudflare fashion, I want to be rather transparent in what we've seen last year, what we saw at the end of last year, and what changed at the beginning of this year, what, you know, what made us change guidance and what the facts are that we looked at. What you see here is sales cycles over time last year and how our pipeline growth developed over last year. The first thing you see is this big dip that started in the fourth quarter of 2021.
You remember Matthew was, I think the 1st CEO who stood up and said, "Something has changed. Our pipeline is slowing. There's a new sentiment out there." We were 1st company to talk about it, got a lot of headwinds for being transparent early. What that does is it, you know, it didn't really impact the opinion we had for 2021. It made us really careful to think about what would happen at the beginning of this fiscal year, because pipeline moves into ACV moves into revenue. If this pipeline moves in one or the other bucket, you just shift revenue up. We reacted early. You heard us talking about changing leadership, not only on the sales side, but also on the marketing side. We changed our campaigns.
Business turned around from a pipeline generation perspective. Started talk about it in the third quarter earnings call that pipeline started to recover, that pipeline started to build again, not only third quarter, but also in the fourth quarter. Pipeline built into the first quarter of this year was also really strong. When we entered the year giving guidance, we looked at extraordinary good pipeline. We looked at sales cycles that were ticking up slightly over the course of the year, but nothing dramatic. We assumed that based on not even improvement, but slightly worsening, sales cycles and the pipeline we had, that we were, remember my words, we were more nervous about the first and second quarter of this year at that time than the second half, because we understood what the pipeline would carry.
The first quarter happened, based on the events that we just described, and we saw literally an explosion or in sales cycles, 27% increase on average in the first quarter in parts of our business, especially when it came to expansion with large customers, north of 40%. If you do that, the math is really easy. You have ACV that was supposed to happen in one quarter, shift out. With ACV shifting out, revenue is going to shift out. Under the assumptions that, you know, our win rates have not changed, this revenue is moving forward. It's not so much a question for us if this revenue is going to happen. It's a question of when this revenue is going to happen and when we are going to catch up.
For guidance, we have assumed that the increase in sales cycles we saw in the first quarter is not really going to shift and impact in the, over the rest of the year. It's almost like a bow wave in front of a boat that you push out, and it only starts to come down if either you know, the speed of the boat doesn't accelerate anymore or it gets, you know, slows down. For the time being, we've assumed that this bow wave is going to continue. We will not see an improvement over the course of the year, and that is literally behind the algorithms that impacted our guidance. Also there are near-term headwinds.
The company, we as a team pivoted really hard, and compensated external headwinds really fast. This is the third quarter now where we delivered record profits in terms of margin and operating profit. Really proud of the team to make this hard pivot in light of different headwinds. It's a sign for the elasticity in the business model. I think it's also a sign that we can show extraordinary operational rigor to manage to those results. We're extremely proud with what the team has achieved here. This sets the stage now of what I would like to discuss moving forward. How do we think about growth and profitability moving forward? How is this path to $5 billion going to look like?
How do we bring together what you heard in the presentations before regarding new products, where customers are, how we expand? It all starts with a platform that delivers innovation. Expanding our product portfolio is at the heart of it. This allows us to attack extending adjacent markets, and allows us to not only get to customers, but grow our customer base terms, both in terms of size as well as in terms of revenue we have with the customer. I want to give you insight in how this journey is coming along. Here you see this rapid enterprise expansion in our numbers. You heard from Matthew that 30% of all Fortune 1000 customers, 30% of Fortune 1000 is a customer of Cloudflare, and six of the top 10 are customers.
Our large enterprise go-to-market additions are really thriving and in incredible momentum. The largest cohort, customers that make more than $5 million with us, in terms of revenue is up by a factor of 18 over the horizon of four years. That's quite incredible achievement. You see the larger the revenue cohort, the faster the expansion. What is driving this? How has the face of our customer changed over the last four years driven by this momentum? If you want to be a top 10 customer of Cloudflare today, you have to be north of $6.2 million of revenue with us. This is up 5x from where we were four years ago.
If you want to be a top five, 25 customer, north of two and a half million dollars. Even to become part of the top 100 customer cohort, you have to get now close to $1 million of revenue with us, up 4 x compared to where we were four years ago. How do customers grow, and how do we expand? Can we say something in terms of the numbers of products they consume? You heard from Patrick already that one of the stats we used at our IPO was more than 70% of our customers are using four or more products. I'm really proud to share, you know, these stats now with you, how this has changed.
We are now in the good situation that, you know, 22% of our customers are consuming more than 10 products. All cohorts larger than eight, larger than nine, more than 10 products, have significantly and continuously expanded over the last four years. This is a testament of, you know, delivering products from one platform, all the opportunities we have to consolidate spend in front of customer's budgets. 10 products is not the limit. This gives you an indication, I think, of the white space we have. I think we can comfortably get to $5 billion just selling existing products to existing customers. This is the amount of white space we have in terms of expansion opportunity based on the pipeline and the products that are already built today.
More importantly, if you look at product adoption, you can also see that the bundles that make up more products are expanding the fastest and are delivering the highest rate of cohort revenue today. Today, the cohort of products in terms of revenue contribution that is more than 10 products is up, more than hard to time for me to read it, more than 15x over this period of time. The more products we sell, the more revenue they generate, the faster they grow. It's, I think, a good sign of the adoption we have and the soundness of the expansion strategy that is in place. This is before all the opportunities that Marc indicated. We've come to this point with the limitations we have. Now is the time to fix them.
Everything that Marc presented is not part of any guidance we gave earlier, at the end of last week. Attach rates are also interesting. How do all the new product waves that you heard about today, from my colleagues, when it comes to act two and act three, are impacting? Do we see traction in our numbers? The good stats to share is that attach rates for the new products are up quite significantly. You heard from Patrick that today 90% of our contracted customers consume our first act, our application services. Attach rate for Zero Trust today is north of 20%. Network services is north of 15%. We are now already attaching developer service to more than 20% of our contracted customers. This is reflected in the ACV structure.
There's a chart we added after some feedback from last week. If you look at ACV contribution by product wave, you can see that the momentum of the second and third act is reflecting in the numbers. Last year, already, more than a quarter of ACV was contributed by act two and act three products. You can see the trajectory. You can see where this is going. You know, Michelle, if she were up here, would say, "This is just the beginning. We're just getting started." An incredible opportunity, a huge opportunity to deliver revenue on many independent growth vectors. How do we combine this growth now with our trajectory towards profitability? How can we stay, not only stay profitable and deliver cash flow, how can we increase where we are going?
I would like to talk about unit economics first, give you more insight, really, on how the elasticity of the business model works, and then connect it to the long-term model that we update in one or two parameters. If you look at us and how we build products and how we sell products and understand the unit economics around it, you know, we have to deep dive. There are the cost to book revenue.
This is Marc and his team and marketing. You can see this, despite this journey of larger and larger accounts moving up the enterprise stack, we have been quite consistent keeping the ratio almost flat to slight kick up in the end. Quite an achievement. We've seen operating leverage, but the big opportunity to gain further unit economics in our cost to book were really described well by Marc.
It's a talent topic. It's a process and structure topic, but there's significant opportunity to continue to achieve scale in our cost to book. The blue line really shows you the cost to serve our products. Here, the efficiency of the network really comes to play. All products-On every server, in every location, the complete surface of the network is a degrees of freedom to manage costs in terms of demand and supply. Despite the increase in traffic, despite the increase in locations, despite the increase in number of customers, despite the increase in number of products up to 50 revenue-contributing products now, our unit economics have really scaled well. What we have to keep now in mind is the ACV chart I showed you. The increasing portion of especially Act Two products. These are Zero Trust products.
I think we mentioned before that adding these products to our platform almost come at zero additional cost. They take advantage of the infrastructure that is in place already. The pipes are sized by the application services, pushing traffic out. Now we can use Zero Trust products literally to use the capacity, bring traffic and data back without incurring additional cost. That is why our cost to serve will continue to scale with significant opportunity just by the fact that Act Two products will become a bigger part of our revenue moving forward. We've been quite efficient to not only keep a very high or very low attrition rate from a customer perspective, we've been able to improve. With this, our unit economic margin here was about 42% at the end of last year, a significant achievement.
If you combine this with what I showed you before, it makes us super comfortable with the operating model we have put in place. We see that we can keep cross margins in the range of 75%-77% with what we have built, and especially with the new products taking a higher share in terms of revenue and load on the product. We have seen significant opportunity on the sales and marketing side with the improvements that Mark has been talking about. We have been quite well already generating significant G&A leverage in our cost, coming down from 18% a year after the IPO now to slightly below 12%, and there's significant room. We are highly comfortable in a long-term business model that gets us to 20% operating margin.
The big update is we've turned free cash flow positive at the end of last year for the second half. We are committed to being free cash flow positive for this year. How the unit economics and the growth path looks like for us moving forward, we are highly confident that we can achieve north of 25% free cash flow margin in a long-term operating model. If I summarize everything I've addressed, innovation, I think that's a big takeaway for you, drives a massive and growing total addressable market that we can disrupt. We've already seen significant enterprise traction that yield large customer growth and large customer expansion with north of 10 products now being the largest contributor in terms of revenue. We will continue to invest in go-to-market, upgrade our talent pool, as Marc explained very clearly.
A highly efficient business model, highly elastic, incredibly good unit economics that allow us to profitably scale moving forward and digest significant revenue, customer, and product growth without really impacting the operating structure of the business model significantly. With that, significant and multiple levers to not only continue to deliver free cash flow, but also to increase free cash flow performance. A really exciting story and a massive opportunity despite some of the economic headwinds we're seeing in the first quarter. With this, I would close. We have a slight break because we get the stage here ready for Q&A. Thank you.
A little bit. Okay.
All right, there we go. Now it's on.
Okay. Yeah.
Yeah.
That sounds better.
Oh, there we go. Hi, Sterling Auty from, let's just say MoffettNathanson.
I'm sorry.
There was some really nice talk through the day about some of the incremental product and product trends, and two really stood out to me. One was firewall, and one was kind of around the WAN, SD-WAN portion. I wonder if maybe you could describe what kind of conversations you're having about the interest out of larger customers in particular to finally eliminate much more of the boxes at the edge of the network, so the service enabling. Is this gonna be the point where we get the tipping point or is there still, you know, more to go before we get there? Thank you.
Yeah. You know what? I think that, we have had a network that spans the globe for a long time, and I think that what our team over the course of the last three years and really accelerating in the last 12 months has done is spend time talking to customers and understand what it is that they need. I think we've got the pieces now put together where you can truly eliminate the big, expensive boxes that are there.
One of the things that was super counterintuitive to me at least was when we were talking to customers, they actually said, "But you still need a box." I remember Annika and Rustam on our team came to me and he said, we, you know, "We have to build a box." I was like, "This isn't some episode of, you know, Silicon Valley. Like, we're not in the box-building business." They said, "Yeah, but at some level, the traffic has to get from the branch office to us, and, like, there needs to be that connectivity that's there." I think that was the last piece of the puzzle that is coming online now, and it's, it was just a small piece of Annika and Rustam's presentation.
Personally, I was actually quite resistant to it internally. We had all kinds of rules about, like, we're not taking inventory in boxes. We're not actually going in the manufacturing business ourselves. You gotta go, you know, find other folks. We've now got multiple different partners who can give you a very inexpensive, you know, piece of hardware, effectively a server, that can go into your branch office and now gets you directly onto the Cloudflare network. For me personally, you know, that in my home now is my firewall. I connect to it. What's magical about it is not only is it significantly easier to manage, the performance goes up massively, the cost goes down significantly. I think that that's the piece which has come together.
If we hadn't really focused, if we'd been just dogmatic on, you know, that's not the business that we're in, if we hadn't spent the time listening to what large customers had, I don't think we would have come to that conclusion on our own.
Mm-hmm.
I think that that's an evolution of us as a company maturing, speaking to much larger customers, listening to what it is that they need. Then sometimes actually then going out and saying, "Okay, well, it's not about eliminating all the boxes. There still needs to be a box." There's a Wi-Fi repeater somewhere in this room. Like, that still needs to exist, but we should take the intelligence in that. We should drive the cost of that down as far as possible, put more of the intelligence in the network, and then build that in. I don't know, I'm, I have several bets with several people in this room on, you know, how long box manufacturers are gonna exist.
I think there have been sort of three head fakes, that, you know, the Trump tax cuts pushed sort of a bunch of box manufacturing earlier 'cause there were. You could depreciate it a lot faster. I think COVID dramatically accelerated that. Then I think supply chain shortages caused a lot of people to just sort of say, "If we're ever gonna need some equipment, we've gotta put an order in now." You have these giant backlogs in the hardware companies. I don't know that there won't be a fourth head fake.
There might be, I mean, an asteroid might hit the planet tomorrow, and all of a sudden everyone's like, "Oh, I gotta go buy more Fortinet boxes." But it, we're definitely on the right side of history here, and I think customers are telling us that. At some point, you know, you run out of head fakes.
Just, the only thing I will add is yesterday we had our customer advisory council here, one of them, we spent the day with customer, we had a whole long conversation about this exact topic. To your point, the stars are aligning. Actually, Chris from Long Walk and Annika and I were talking before the session started, she's like in the weeds, talking to the customers, building the products. In her words, like, 'I can see all the stars aligning.' I've been in many conversations with customers over the last quarter where it's like really deploying, running pilots, case studies. Like, people are really excited about it, but it's early innings. It's like the stars are starting to align.
We are not building boxes, no inventory, no supply chain issues, so.
We have a lot of partners building boxes for us.
Just making sure. Alex.
Great. Alex Henderson over at Needham. Thanks so much for a great presentation today. I did wanna go back to the quarterly call and clarify what exactly you're saying about how much of the pressure on the outlook is coming from the sales reorganization, which I don't think is the primary issue.
Yeah.
-or even really the driver of it, but rather more purely macro-
Yeah.
which is not a Cloudflare issue.
Mm-hmm.
Rather a broader macro issue. Can you clarify, you know, how much is coming from what and how the timing of the sales organization
Yeah.
changes will feed back into.
Yeah.
a positive trajectory in that category. Thanks.
Yeah. Let me get started then. You know, in hindsight, it might have been good to keep it separate. The reason for the change in guidance is literally the two charts that I showed you. An extension in sales cycles by 27%. You have to keep in mind that our sales cycles are short. Normally we close deals within a quarter, so 27% pushes ACV out by a lot, and that means you have immediate revenue impact in the year, even by a pushout in the first quarter. Just to reiterate, so the win rates are not changing, so getting to this ACV is a question of if, not of if, but of when. The transparency we share with you is also something we share internally.
We wanted to be in a position to talk to employees fast and not have rumors and information and misinterpretation out there. We thought by making it public on a platform of an earnings call would give us the ability to do what needs to be done with speed and transparency in front of the employee. That's why we put things together. Little did we, you know that it would be combined. If we had spent more time thinking about it, we might have come to that result. That is clearly on us. It has two reasons. We wanted to be fast. On the execution, North America will happen a lot faster, that might be executed by the end of next week already.
Europe might take a little bit longer. I think if you look at the model and how we built it, ramping up new AEs is very efficient within Cloudflare. We normally get a mid-market AE to full productivity in four months, an enterprise AE in six months. That was another reason why we wanted to be fast, because if we do a good job in managing through this upgrading talent, we might see impact, positive impact before even the year is over.
Just to be 100% clear, change in guidance is 100% because of sales cycles. I think we did a bunch of work to start reporting a week earlier than we usually would. I think what you've seen with every company that's reported since us, is everyone is seeing a lengthening of sales cycles. Now, I think it impacts us more than others because our sales cycles are so short, so we see it faster. We're a little bit the canary in the coal mine. I think, again, I have always tried as CEO to be 100% transparent about what it is that we're seeing in the market. In, you know, I was the first CEO that was out there saying the economy is slowing down. Lo and behold, it's slowing down.
I don't, I think as companies have longer sales cycles, they are going to not see the impact on revenue as quickly as we do. If you look at companies that have high velocity, very fast sales cycles, they are the ones that are out there saying, "We're seeing the same thing that Cloudflare is seeing." The changes around the go-to-market organization, I think there are two things that are notable about it. The first is, like, this is something. It wasn't something we just decided to do last week. This is something that had been in plan from the day that we hired Marc. If we're completely and totally honest with ourselves, we knew that our go-to-market organization was not as good as it could be, and we needed to change out the leadership.
That, that was a process that we started very beginning of 2022. We found someone who I think is incredibly detail-oriented, is in the weeds, knows the numbers, is about how do we make a really efficient machine. That's what Marc is doing. That's what Brent, who runs our marketing team, is doing. Again, the leading indicator of success there is pipeline growth. We're having 4x pipeline coverage, which is great, right? That, that's a, that's a great stat that was sort of up there and not a lot of indication on it. We're seeing record pipeline achievement, which is there. Again, that's the predictor of the future.
I look at that and I go, "I have never had more confidence in Cloudflare's business over the long term than I do at this exact moment," because The philosophy on the earnings call, why do we do what we did? Again, we learn from all these things, is we thought, listen, the same way that Netflix looks out and goes, "Yeah, we had a tough quarter, but we have this really big lever that we can pull, which is we can crack down on password sharing. Here's the opportunity for us to do it." We looked at this and said, "Listen, sales cycles are extending, but we've got 100 people on our team that are achieving less than 4% of the results.
We can rotate those folks out, and there's an opportunity." Even though we're not building any improvement into guidance, we knew that that was something that we could do. In our heads, what we were thinking is, yeah, sales cycles are extending, but we're not fully optimized as a go-to-market organization. If we just get to a normal distribution curve, that significantly will re-accelerate revenue growth. That seemed like a very clear thing to do. In retrospect, I think that it got all muddled in terms of the message and again, I think it's one of the reasons why I was happy that we had Investor Day today because we could be radically clear about it. What we are seeing is sales cycles extending. We do not expect that to improve the rest of the year.
We know what our productivity is across our team. We're making changes on that, but we don't expect that to improve for the rest of the year, right? With that's why we changed guidance. All of those things, if we do get improvement in sales cycles, if we do get improvement in productivity, those will all be upside to what it is that we're looking at today. Is that clear? Okay. I apologize for not being clearer before.
Yeah. Michelle?
Thank you so much. Thank you for being very candid and all the color you've provided. Maybe one for Michelle, Matthew, Mark, just given one of the slides that he's gone through, and maybe Thomas, a quick one for you. On some of the metrics that you've provided us, will that be just an analyst day update type of data point, or would we see that on a quarterly basis, maybe? What's the thinking?
Well, let's see how the quarter progresses. You know, I understand if we give a little finger, then we end up giving an arm. I think since we've been a public company and probably already before, we have been super transparent trying to explain our business, whatever it takes. I think that is not going to change. We also have to bear in mind that, you know, there are competitors out there, and we cannot share everything that wants to be shared. We continue to try to find the right balance of being transparent and explaining how the business works, but also protecting, you know, some confidentiality on how we are doing.
Fair enough. On that, maybe go-to-market. It was interesting to see that slide about, you know, all the many products that an account executive he or she may sell and how it might be confusing at times. My question is, do you think it is more the land issue where you, I don't wanna say struggle with, but maybe slightly more complex? It would appear that the expand is easier once you are within the customer. How would you think about it?
That slide is true, but we still have a lot of account executives who are doing a great job on both the land and the expand. One is not better than the other. It's really not. It's actually a great source of strength in our motion on the go-to-market machine that we continue to land because no one has to adopt all four areas on day one. That is a huge strength, where you find some use case, there's a lot, you heard a lot today. There's even more, like, we could do a lot of things. You find a new use case, get started, and then over time, we expand. It's really both work really, really well for us.
It's interesting, the sales cycle, like lengthening in Q1, they got lengthened by almost twice as long from the expansion, the new logo. Our current customers are like, being like, "We love you. We know exactly, but we're just, we gotta check with our CFO five times. We're not even sure the budget's there. We don't have the people. Like, we gotta wait." It's this idea of well, why now versus in one month or two months? We're not losing to a competitor.
Mm-hmm.
Which is really important. You can only push a customer so much. Marc and I have these conversations where it should never take a day longer to sell our service than it needs to, but you cannot make a company buy. Like, I've just said, if you spent any time with companies, it's like you have to go at their pace.
Mm-hmm.
If you push too hard, they don't like you. There's other vendors out there that push too hard, and they don't like them. You gotta kinda be when you're ready, we're ready, and then we pick up the phone or the email when they're ready, and we have a really good reputation for being that. One's not harder than the other. We have teams that are very, very focused on both, and I think we've got the right training and enablement. Now, is it fully world-class? It's not, which Marc went through, and that's why we're really happy Marc is here and the leadership changes we're making.
The one thing that didn't come up in Marc's presentation, I can say this, he can't, as someone who's been at Cloudflare for a long time, there has been no tissue and no organ rejection from all the changes in the plans that Marc put up. We've gone through that with our go-to-market organization. People are excited. They're like, "Thank you." People understand we have great products. They believe Cloudflare. Awesome. This is gonna help make our jobs. Everyone knew that we're hiring and spend so much time to hire, be more productive. I'd say there's been very, it's been, people are welcoming Marc and Brent with open arms, even though there's been lots of changes. That does not always happen. The integration's been really good. That's what I would say.
It's more about how do you make this easier, how do you make it smoother and continue to make progress on both dimensions.
Yep. Yeah. Maybe one more data point that I try to make sense out of this, because as Michelle said, you know, expansion was especially elongated in expansion with larger customers. There seems to be a shorter reaction time in large, disciplined, organized organizations to react to whatever we saw in the first quarter, react faster and push on the brakes faster. I think that is a big part of why we saw expansion, especially in large customers, elongate so much. They reacted significantly faster than everybody else.
They have a good CFO.
Yep. Not as good as the one at Cloudflare, but yes.
Adam.
How you doing? Adam Borg with Stifel. Thanks so much for the question. Maybe Matthew on Workers. It's great to hear about the early traction. The presentation with Ali was great. Where are we in the journey for building out a developer ecosystem? I know we talked about it on the earnings call. We talked about it last year. Maybe part one, overall journey. Part two, you know, you had this $2 billion funding program, that Launchpad program, and just thinking about how could that help accelerate kind of a developer or broader ecosystem? Thanks so much.
Yeah, I think that the thing that surprises me about Workers is that we didn't build it as a product. We built it for ourselves, and I think that's how every great developer ecosystem starts. That's how AWS started. That's how Apple started. That's how Microsoft started. You build for yourself, and then you, and then you turn on. Virtually every product that we build at Cloudflare today gets built on the Workers ecosystem. Customer zero is Cloudflare. Our team is really hard on Ali and the Workers team at pushing them. The reason we have an object store is because we needed it.
The reason we have a database is 'cause we needed it. The reason we have Durable Objects, which is this whole new kind of thing, and you heard how MLB is getting you game score. They, t here's no way. There are whole companies that are now getting built. Sequoia just backed an entire company building real-time data built entirely on, around Durable Objects because that, t he companies are building around that. I think that it's hard for me to say exactly where we are in that journey, 'cause I don't, I don't know how long the journey is 'cause I don't think it ever stops.
What I will say is that I think that looking at places where developers can start with a just blank white canvas and pick what their tools are is a great way to see how much you've built something that has a real differentiated developer experience versus how much somebody's just locked into kind of the decisions that they made before. AI, which is a space that everyone's excited about because it's exciting, but where I think it's exciting is not just because, you know, there's an opportunity that these things are gonna grow like crazy, but it's a place where people are starting with a blank white canvas.
If you look across AI company after AI company after AI company after AI company, almost all of them are using Cloudflare as part of their overall solution, and they're using Workers specifically in order to drive a lot of what they're doing. When we talk to them, they say the reason why is 'cause it's a much more nimble, much faster to develop, much easier to get things to scale. You have all these security benefits and neutrality and all the other things. Fundamentally, we can develop faster on this platform than we could if we were trying to use one of the other hyperscale public clouds.
I say that, you know, if you're not locked into the decisions of the past, we're seeing developers very much bet on our ecosystem time and time and time again. And what's been interesting is the, you know, the $2-plus billion that we had from, I mean, almost every one of the leading venture capitalists that are out there, like, those VCs are saying. They're the ones calling saying, "Hey, can we do more?" 'Cause we are actually becoming this incredible deal flow source for them, and especially in the AI space, that has been, that's been amazing. Again, if you look at every single AI company that has had an announcement around, you know, VC funding over the last little bit, they are all.
A lot of those actually came through that developer program. Time and time again, they're using our platform from the earliest days. This is a journey that doesn't have a definitive like we're done point, which is different than something like firewalls. Like firewalls, it's sort of like, "Oh, okay, we're done." You obviously have to continue to invest, I think the developer platform is something that naturally takes time to develop. The early signals when developers get to start with a blank white canvas, and again, the AI space, before that, which I think is very sustainable. Before that, it was, you know, crypto and blockchain and all those things. Everyone in that space, again, started with a blank white canvas, all were using Cloudflare Workers.
I think that's a much less sustainable space. What we see is that whatever the new thing is, Cloudflare becomes a part of it from day one.
Thanks.
Stay tuned for Developer Week. Yes.
Okay.
Yeah. Again, not next week, but the week after is Developer Week, and the focuses are developer productivity and AI.
Thank you for hosting us. This is Yi Fu Lee, Imtiaz Koujalgi with Cantor Fitzgerald. My question revolves around the pillar three R2 storage, right? We find very good value proposition in terms of the no egress fee and really like the analogy with the Hotel California. Could check in, but you can never check out. Just wondering, what's the pushback, the major pushback? Why isn't more copy adoption ramp up, right? Just like, kinda like interest rate, right? When a bank goes up with interest rate, you'll go to the better competitor. Second part of the question on the same thing is, if the hyperscaler were to go away with the egress fee, right? How would you know, envision that dynamic would change? Thank you.
Yeah. I would actually guess that if interest rates go up at banks, the majority of the people in this room, even though you're incredibly financially sophisticated, don't change your bank.
Mm-hmm.
It actually takes a ton of pain to change that over. The same is true with object store, where moving. You've got a whole bunch of stuff that's working over here, we can save you a bunch of money, but going and just saying, "We're gonna lift and shift everything over overnight," it's just not how organizations work. We see incredible demand for it. Again, when people are starting with a blank white canvas, more and more of those first use cases are starting on R2.
Moving that over is actually, it doesn't happen overnight for the same reason that, again, the majority of you, even though you could probably get a better interest rate moving to another bank, you're like, "Yeah, but, you know, my ACH direct deposits are all wired up here," and all, you know, these things. Same thing is true in this case. We're winning new use cases. That's thing one. Thing two, I think the pushback comes less from customers and more from our own internal teams. There are times where people will say, "We wanna move, you know, petabytes of data over to you." Yeah, that would generate revenue for us, but not all revenue is good revenue.
I think we're being very strategic about saying, "What are use cases that make sense?" AI, perfect use case, makes tons of sense. That's a use case we wanna have on it. If Cantor Fitzgerald has long-term log storage that you're, you know, storing, you know, and only looking at every five years, that's not a very good use case for us. We're not the best place for that to live, and that's actually a pretty low margin opportunity for us to go after. I think we're being very disciplined about what we're going after. It's important to remember why we built. To answer your second question, it's important to remember why we built R2.
If you go back to when we announced it in the earnings calls, and people would ask, they would say, "Well, what will you do if Amazon waives egress fees tomorrow?" I said, "I would throw a party, biggest party I possibly could," because that then means that the one thing that keeps people from naturally using us as that fabric that connects all the clouds together goes away. All of a sudden, we can move data from one cloud to another cloud to another cloud and be that central processing unit for the clouds in a way that's not. Egress is the thing that keeps them from doing it. By launching R2, we kind of created a problem for them. They know if they waive egress fees, it really empowers us.
If they don't waive egress fees, over time, R2 captures it. We win in either way that the game gets played. It actually would probably be a slightly better outcome because being in the data business, in the object storage business, it's not a high margin business. It is not margin accretive to us, whereas like Zero Trust is massively margin accretive to us. We balance these things out and it can hit our long-term margin targets. If all of our revenue growth all of a sudden came from storing data, It's a commodity product at some level. It's a strategic lever to give us leverage over the clouds to get them to drop their egress fees, and that is one of many things we're doing.
Same thing is happening, again, watching what's coming out of Europe, where Europe is starting to see egress as an anti-competitive measure. Like, if that happens and Europe suddenly bans egress fees, again, awesome. That is a wonderful thing. It might hurt our R2 business, but it's a commodity business. Like, we use this as a lever to change the industry more than using this as a way to drive, you know, revenue from storing data. Does that make sense?
Heads or tails, you win.
Heads or tails, we win. Again, I think we've tried to position ourselves in that, exactly that situation.
Go Jonathan Ho.
Thank you. This is Jonathan Ho from William Blair. I just wanted to maybe go back and take a look at sort of, you know, the lessons that were learned from some of the identification of the Salesforce productivity issues, and perhaps maybe understand, you know, as a management team, you know, what were some of your takeaways, and are there other areas that you potentially see for greater efficiency in the company? Thank you.
I'll start. Then Michelle and Thomas might have additional things. I think the right question to ask is why didn't we do this earlier? I think that it's a combination of a number of things. You know, one was COVID. I remember specifically having conversations with employees during COVID where they're like, "I don't know how I can, you know, raise my kids, keep my marriage alive, and work at Cloudflare." I was like, "Listen, if you have to prioritize those things, prioritize 1 and 2, and then make time for 3." I think it was the right thing to say. I mean, that's the human thing, and I think it's why we've had, you know, the loyalty in our team that we have. I think it's not unique to us.
I think every company that was out there said those same things. We definitely relaxed performance standards. And it's hard once you, like COVID was the impetus to do that, but when was the impetus to flip it back? So that was thing one. I think thing two was that on average, our productivity looks pretty good, but it's bimodal at that level. I think that we were okay with the averages, but we hadn't again tightened the screws around the sort of low performers that were there.
I think the third thing was that if you went back 1 year ago and you thought about what the tech talent market looked like, the opportunity cost, even for getting a low performer off, were substantially higher because the replacement cost of a new performer was extremely high. We looked at it, and again, we're like, "We know we have to do this, but is this the exact moment in time to do it?" I think the fourth thing is, you know, the person who Chris Merritt ran our sales team. We owe a ton to Chris. Very few sales leaders go from 0 to $1 billion in revenue. If we're again, super honest, Chris got to the Peter principle and was not making those changes.
We needed to change the person in order to change the team. We spent a bunch of time, you know, again, finding the right person with Marc. To this, in the very first event, Marc, you can talk to Marc about it. Very first thing that both Michelle and I said to Marc was, "There are gonna need to be substantial changes on the team. It's gonna take you a bit of time to kinda get your hands around it, but I want your kind of 120-day plan to be like, how are you gonna execute it?" Almost on his 120th day, we pulled the trigger on executing that. I think the, you know, the. I don't know if this is.
I think that the awareness on our team, especially with sales cycles lengthening, that the business got harder, was the perfect opportunity for us to say, you know, we're gonna start putting back in place real performance-oriented culture. We're gonna go back to what we looked like pre-COVID. We've got the right leader who can make this change. We've got incredible talent which is coming on board. We're on track to have over 1 million applicants at Cloudflare this year, 40% of which are for sales roles. Like, our ability to bring people on in a much more cost-effective way than it would have been one year ago is there. We had this opportunity with this quarter to be able to say, you know, we're gonna make that change and straighten everyone's spine.
What r eflecting what Michelle said, that my inbox that day was not full of people saying, "How dare you?" My inbox was full of people saying, "Thank you, and how can I step up and do more?
Great.
I have nothing to add because it was very feature complete right there. So.
Go to Gray.
Thank you very much. Gray Powell with BTIG. I want to follow up on Sterling's question earlier. Like everybody else, we've been hearing a lot about appliance refresh delays in the network security space. I'm just curious, does that give you sort of an opening to get more aggressive or maybe more creative with customers to take out appliances in six-12 months? Like, are there any special promotions or anything that you could be doing now to sort of get ahead of the curve?
Undoubtedly, the answer is yes. I don't think we have something to announce right at this particular moment, but what I think is if you can have a very dumb physical piece of hardware, which again, we've gotten partners to build for us now. It was really interesting. I mean, when the team came and said, "We have to build a box," and I was like, "We're not building a box." They said, "Well, you know, can we get a partner to build a box?" I, you know, immediately, you know, called Michael Dell and called, you know, a bunch of the vendors we work with, and every single one of them said, "We want to partner with you. We wanna work with you.
Tell us what the spec looks like, we'll build it." I think that that is again indicating that people see the direction the world is going. They wanna have some part of that. If, again, we can move more of that intelligence into the network and deliver that network with the network that we have, that that's a real opportunity. I think the other thing is, this is interesting, and this is also not an answer to your question, it's really difficult for a new entrant to build the network that we have today. Figuring out how to get a location in, you know, Baku, Azerbaijan, or in Lahore, Pakistan, or in, you know, I mean, in all, in Kigali, Rwanda. We're in all those places.
We're there because 20%-25% of the web sits behind us. Now that we're there, then that gives us the ability to efficiently run a network in a way that is very different. Again, I know that we're on the right side of history. I don't know how many more head fakes the hardware space has. I think that these are all extremely high margin, extremely attractive markets for us. I think we'll run a bunch of experiments to see what resonates. As soon as we figure something out, I think you'll see us be able to continue to accelerate in that space. We're trying to be very non-dogmatic.
You know, I was the one saying, "There's no way we're building physical hardware." Again, we're still not building it, but we partner with people to build it. Our team said, "No, but that's what customers want." I think you'll see us run new plays in that space. I think there's just still a huge opportunity.
Gabrielle. Gabrielle. Right there. Mm-hmm.
Good afternoon. Thank you. I wanted to follow up on some of the AI use cases that developers are using Cloudflare for. How do you think about the use cases that make sense in a distributed edge compute storage model versus those that may require either for data gravity issues or compute and storage issues, more of a centralized use case that might be better served by the hyperscalers? Thanks.
I think that I think there's a difference between training and inference. I think that you're gonna see training happen much more in centralized models. We're not the right place probably to do training, but you're gonna see inference happen much more close to where you are. I think the question will be, does that get pushed to your actual physical device? Some level we're competing with the actual physical device manufacturers or does it make sense to run, you know, within a few milliseconds of where that device is? We think over time it's gonna make more sense for that to actually happen, 'cause you can drive the unit cost of the device down further and further and further.
If inference is running, you know, within a few milliseconds of wherever that device is as opposed to running on the device itself. But also inference can be very power consumptive, which means that for battery power devices, it's actually quite expensive to run in that, in that world. And we can do it much more efficiently, and we can deploy much more efficient usage of whatever the silicon is that we have to do to drive that inference. I think training is probably gonna happen in big clusters of centralized models. Inference is gonna happen at the edge. What I think will be the interesting bit of that is how many models, how many trained models do you actually need versus how much inference do you need?
I think that I don't know what the answer to that question is, but I think that there will be much more inference than training overall. I think that positions us extremely well for what the future of AI is. We're, I mean, we're making this up every day 'cause this is just such a new space. I think we're trying to learn from our customers and our developers and then build the network that can support what they need.
One thing that I'd actually just layer onto that, just looking at Ali and Dane at the back, and they know this is my one of my favorite use cases of R2 is an AI company that uses virtual machine instances in Google Cloud but stores the data in R2. Because it's in R2, you can arbitrage out the clouds based on price and availability of GPUs. If it was still in Google Cloud, you lose some of that arbitrage because you have to pay egress fee. Then you can say, "No, actually Oracle gives you best pricing. You know, Google does, Azure does." That's just an example of Cloudflare as a foundational layer of a multi-cloud architecture.
It might be that, and I mean, just to, but to tie the two things together, it might be that training still happens in multiple clouds, but the data, the training set lives in Cloudflare. We think that is a exceptionally good R2 use case and strategic use case. Whereas again, like, log storage is not nearly as interesting.
Cool. Next question. Haley. Steve. Sorry.
Thanks. James Fish with Piper Sandler. I won't make a Pied Piper reference given the Silicon Valley stuff before.
The crazy thing about that is we launched at TechCrunch Disrupt. We came in second, I mean, there's lots of this. We were invited to the screening, and we're sitting in the Moscone Center, or we're sitting in the wherever it was, and all of a sudden our phones start lighting up like, "How did you get product placement in Silicon Valley?" We're doing, like, the screening of it, but it had aired on the East Coast already. We're just panicked, like, what's the product placement gonna be? They referenced us on it. Cloudflare actually gets worked into that. I bought a bunch of the props 'cause they have a diagram where Cloudflare is sort of a piece of what it all worked on. Anyway, there's a long history there.
Daniella in the back is, if you need, if you ever want a Pied Piper shirt, we have time.
I think as a Piper Sandler employee, I'm gonna have to take you up on that.
Yeah.
Twofold question. On the go-to-market side, Thomas, appreciate all the details around now eight, nine, 10 products. I mean, it was only a year or two ago that I felt like we were only talking four to five. You know, that still only represents 20% of your product set, though, when you think about those, you know, 10+ products. How are you guys thinking about tinkering with some of the packaging and bundling potential here? I know you guys do it kind of, behind the scenes, but is there a way to actually drive that better vendor consolidation with better packaging and bundling, especially when you start to think about acts two and three? Additionally, around that, appreciate the color around the 25%, coming from acts two and three.
In terms of that $5 billion goal, is there a way to think about what you guys are thinking about for what percentage that could be within that 2027 timeframe?
I start with your last question first. I think I mentioned it on the presentation. There's a path to get to $5 billion just by selling what we have to the customers we have. I think that is a really powerful statement. That is a lot of white space. Now if you connect this with what we said before, the flywheel goes the faster, the more products the customer already have. Expansion in general, apart from the macro impact we just discussed a minute ago in the first quarter, should be a easier go-to-market for us. It accelerates. You saw that also. The more products we have, the faster the growth rate. It drives itself.
That I'm really impressed by this. There was one question that we didn't completely answer, you know, are there other opportunities for efficiencies? I think what we learned is, if you continue to grow north of 30%, 40%, 50% and you compound this growth rate, staying ahead of the scalability topic is the biggest challenge. It points you in many improvement directions. Is the team growing fast enough? Are the processes smooth enough? Is the organizational structure the right one? One certainly is in this, how do you deal with the complexity of 50 products? How difficult, you heard this from Mark, is it for an AE to close a deal? How many touchpoints does it have? A big part of this evolution is us finding our way to bundles.
All the companies that we admire went there, whether it's Microsoft or Salesforce. We've been working on this initiative now for more than a year. It's a lot of work, right? Because it's not just the bundle and the pricing and the enablement. It's do we need a new billing engine in order to put this out? How do we meter the bundles? Is the billing engine that we have able to do this complexity at scale? You know, there was a lot of work involved, but we've made good progress, right? We should get ready by the summer.
I think what I would add is.
Yeah.
You know, over time, revenue should approximate TAM in terms of the TAM %. You know, I think about Salesforce, and, you know, once upon a time, I still think of Salesforce as a CRM company. Less than 15% of Salesforce's revenue comes from a CRM. There are customers that think of us as a DDoS mitigation company. I think certainly by the, you know, over the course of the next five years, that will be a very small % of our revenue, and I think you'll see a much larger % of revenue come from things like Zero Trust and Workers. You know, internally, we see ourselves as how are we on the path for those act two and act three companies to get to more than 50% of revenue.
If you just follow the growth curve that Thomas presented, you can see that that's not in the too distant future.
Okay. In the back.
I think Thomas also.
Oh, we'll do one real quick. Steve.
I wasn't getting ready. I'm a little grumpy today, so this may be my last question ever. Look, there's been a lot of spaghetti thrown at the wall to see what to stick in terms of product and delivery launches over the last year. A lot of it seems to be stuck in sort of development. D1 database isn't out, Super Slurper isn't out, and I'm a little worried we have the Google effect here, where we launch a lot of stuff and then neglect it. It's there, but if I'm a sales guy walking in, I'm not confident we can actually deliver. Is this the year of delivery and/or where do we kinda catch up on the back end of that? That's sort of question one. I just have to.
The word everybody cuts guidance, they cut three times. That's our rule of thumb out there. Is there a recession forecast in that guidance cut, or are we looking for another leg down when macro steps back? Just to put the question bluntly.
I'll take the first one, I'll try to resist answering the second one so Thomas can answer it. Our strategy is if you think of Gartner as running the world, which is a horrible way to think of the world, but here we are. I tell our product team, if we launch a product and we're top right when we launch it, we waited too long to launch it. From the beginning of Cloudflare's history, 14 years ago, it was get something out, get it out as fast as we can, get as many people to use it as possible, and then iterate more quickly than anyone else. You can measure this. If you look at our WAF, I remember early on, somebody at Gartner said, "There's no way you will ever catch Imperva.
They invented the category. You will never catch them." We got 'em, right? We just moved up into the right. You can actually measure the hypotenuse of that. Our hypotenuse is longer in any category that Gartner, Forrester or anyone measures and faster. The slope is steeper than any other vendor that's out there. That's the strategy. Yes, it is about throwing spaghetti at the wall and then seeing what sticks. I think we go through kind of TikTok cycles, which is the Intel terminology of new product and then iterate, new product, then iterate. This is. It turns out that odd years end up being the talk cycle, where we're actually, where we're actually optimizing. You will see fewer de novo new products this year.
You will see more improvement on those products. For things like databases, or things where usage matters, those are especially important for us to get out in early users' hands, and it's part of the reason why we have low-end users. When we roll out a new product that's there, we can roll it out to our free customer base and say, "By the way, if you use this product, your whole infrastructure might implode because it's brand new. Who wants to try it?" 20,000 people reliably will raise their hand and say, "Yes, I'd love my infrastructure to implode.
I will totally try it." The way that we can innovate as fast as possible is because if you're a Zscaler and you want to roll out a new product, you can't get 20,000 people to sign up to test something which is there. You have to kind of do that. I think we're. That there are aspects of how Google develops products, there are aspects of how Facebook develops products, there are aspects of how consumer companies develop products that I think we want to be like. What I think is different is we're not, we're not running a, you know, a company that has this giant revenue engine that we can run a university on the side as sort of a labs.
We see a path for every single one of these things to be useful, and our product team, I mean, you talk to our product team, they're more obsessed with revenue than any product leaders that I've seen. They care enormously about that. I, I think that, again, it is by design for us to release things early. There's cost to that. Sometimes people are like, "Oh, it's a toy," and, "Oh, it doesn't work," and all those things, and we have to overcome that reputation. I think if you look across our history, we have been extremely good at getting products out, innovating faster than anyone else, and becoming leaders in brand-new categories, time and time and time and time and time again.
You're gonna see exactly the same thing happen in Zero Trust, where, you know, there are Gartner analysts that are saying, "You'll never catch Zscaler." Wanna bet?
I'll take that bet with anyone here. I think you'll see the same thing with databases, and you'll see the same thing with the compute platform that we have as well.
Over the last three and a half years, we tried to be thoughtful when we gave guidance. We put a lot of math and data science to work to get it right. There is risk in guidance, and if you want to take all risk out, then, you know, that would not show that we want to have the hands on the steering wheel. In the simulations we did, and a sales cycle explosion by 30% was an extreme case. I have to be very honest, I'm old enough now that I remember the last banking crisis.
I would not have thought that I see, over the course of my remaining career, a U.S. bank fail, and now it's three failed, and one of the largest banks in Europe needed to be merged into it. There will be always events like that that are outside of what we control that might impact our business. That's why I think it was really important to say this was not a guidance adjustment because of the other topic. The other topic was contemplated in the guidance we gave.
In the guidance we gave at the beginning of the year.
In the beginning of the year. you know, there might be events that impact us that are outside of our control. I think even in the most difficult times, when COVID started, when we had very little visibility how the year would work out, we said we are not giving up on guidance. We want to show what we can own and put rationale around how we put it together. If it doesn't work out, we rather stand up here and explain to you where we made a mistake or where we got surprised. Guidance implies risk if it's done thoughtfully. I think we have done a good job rebalancing it. This guidance is, if it can be, a touch more conservative.
We said on the guidance call last week, we, especially for the current quarter, we did not bake in revenue recognition from in-quarter ACV just because we have no idea where sale cycles would go. I think we tried to do our job as best as we could, but guidance has risk. We're thoughtful about it. I'm very much convinced it's better to give guidance than leaving it to interpretation of what the business might or might not do.
If sales cycles hadn't increased, we would have been very comfortable with the beginning.
Mm-hmm.
What I don't know is our sales cycle is gonna increase more. Right now, they're holding at what we've seen. They have not gotten worse. They have not gotten significantly better, though. I think barring an exogenous effect, which again, I think a 27% increase in sales cycles in a quarter qualifies as, barring that, we feel very comfortable with where the guidance is today.
The interesting observation I want to also make is, you know, this journey to larger and larger customers, more and more revenue with the largest customers, also makes our pipeline and forecasting it much more lumpy, right? We've probably in this pipeline moving forward, some of the largest deals we ever had. One or two of those deals moving in a, in a business model where the sales cycle time, even after the elongation, is still mostly within a quarter, has huge impact. Getting all right has become more difficult. You could say it's our job to figure that out, and we try that as best as we can.
All right. There's zero, there's zeros on the clock. Thank you everyone for your time today. For people that are in the, in the audience, we do have lunch. Some of the execs will be mingling around. I know this is a little longer analyst day than we've done in the past. I hope you found it useful, thank you for your time.
Thanks, everyone.
Here we go.
Feels downhill.
Feels downhill.