Hi, everyone. Thanks for coming. I'm Ian Lee, Head of Investor Relations at Atlassian. Thank you for joining us for our first investor and financial analyst session. Thanks for coming. We hope some of you were able to make it to some of the events this morning, including the keynote. Before I get started, just a few things I need to cover off from our legal team. Today's session is being webcast and is available for viewing on the IR section of Atlassian's website at investors.atlassian.com. A recording of today's webcast will also be made available on our IR website. We will post presentations from today's session on the IR website after the session has concluded. Statements made during today's session include forward-looking statements.
You should not rely upon forward-looking statements as predictions of future events. These statements represent our management's beliefs and assumptions only as of the date such statements are made. Further information on these and other factors that could affect the company's financial results is included in filings we make with the Securities and Exchange Commission from time to time, including the section titled "Risk Factors" in the most recent Form 20-F and 6-K. I'll do a quick overview of what we'll be going through today. After my intro, we'll have Jay Simons, our President, go over a quick go-to-market model overview. Our CTO, Sri Viswanath, will do an R&D overview. Murray Demo, our CFO, will do a financial summary. We also have, in terms of executives, we have Tom Kennedy, our General Counsel, and Helen Russell, our Chief People Officer, sitting in the back there.
Have a short break after that. We'll move to a broader Q&A session. Please hold any questions you have till that session rather than at the end of each person's session. We'll move to quick cocktails just in the informal cocktail area just outside of this room here. We'll aim to have you out of here by about five o'clock at the latest. Before we move into the main presentations, I'd like to just cover off some of the key takeaways we hope you come away with from today's session. Firstly, our go-to-market approach continues to scale. As many of you know, we have a unique go-to-market model, which is probably quite different from most of the other enterprise software companies that you're familiar with. We don't rely on a traditional enterprise Sales force.
Instead, we have a highly automated model that's already been able to reach 90,000 customers. That number continues to grow quickly. Jay's going to touch about more on that in his section and also how we're helping that drive our growth. Our expansion paths are humming. Jay's also going to touch about how we're growing with teams of all kinds and sizes. I know many of you are also interested in how we're growing with our larger customers in enterprise. Jay will touch on that. R&D is the lifeblood of our company. Our model works because we build great products. Underpinning that is the engineering excellence which is at the core of Atlassian. Sri's going to touch on that when he talks about our technology platform. The forecast is cloudy, but in a good way.
The cloud is already an important part of how we do business today and how we serve our customers. We think it's going to be an even bigger part of our business going forward. Sri's going to touch on what we're doing with the cloud today and also some thoughts for the future. The price is right. Since our inception, Atlassian's focused on delivering great products but also delivering great value. Our value-centric approach means that we think about pricing very carefully and thoughtfully. Murray's going to touch more on our pricing model and our pricing strategy. Finally, we have a financial model. Our unique go-to-market and our R&D focus has helped create a financial model that has delivered consistent revenue growth and free cash flow for many years.
Murray's going to touch more on the financial model and also some of the longer-term financial model considerations. Just a few housekeeping items and logistics. This session is being recorded, the recording will be made available on our IR website after today. The slides that we'll be presenting today will also be posted to the IR website after today's session. If you don't already have Wi-Fi access, there are pieces of paper on your tables which have the Wi-Fi information. Restrooms are just outside there to the right. With that, I'll pass it over to Jay. Thanks.
Thanks, Ian. Hello, everybody. How's everybody doing? We've got an hour of content about go-to-market. I mean, there's a good chance I'm going to have to take a bathroom break in the middle of this. It is a pleasure to talk to you about our sales and marketing, how we acquire and grow customers, and to double-click a little bit into the detail around our go-to-market. There's really three things that I'm going to cover. The first is some of you might be new to this story and the company, I want to make sure that I quickly level set all of us in terms of the problems that customers hire our products to solve, our perspective on the market, sort of like what we're focused on.
Secondly, I want to describe the attributes of what makes our really unique business model work and function, how we think about high-velocity customer acquisition and growth. I'm going to spend the bulk of the hour, hopefully less time if I can do it, double-clicking on land and expand and sort of explaining how that actually works, how we prosecute the customer journey through that cycle, and the things that we do to both improve and grow over time. There's really three things I want to take away that echo some of the points that Ian made, most importantly, that our model is a really unique advantage. Sales and marketing combined with R&D work in concert to create what is a very, very unique B2B software company. I'm going to leave most of the punchline and the results for Murray's section.
That our sales and marketing approach continues to scale. We drive a significant volume of new customers that begin their journey with us, and we then grow them effectively over time by expanding their usage, selling them more products, and eventually growing to scale and standardization. Finally, those expansion paths from one product to the next, from one product to many, from a small team to many teams, are humming. There's a couple different dimensions that I'll talk about. One is the expansion across the diaspora of teams inside of businesses, from software teams to HR, to marketing, and everybody in between. Also, the premium upgrade paths for our largest accounts. That includes upgrades to our data center editions, enterprise services, and then enterprise editions within our cloud family, continue to help us grow in large accounts. Let's get on with it.
Chapter one is just a really quick primer about us and Atlassian. Maybe it doesn't bear repeating that we just live in sometimes it can feel like a scary world, but it is a very incredible and magical world that we live in, where we're experiencing unprecedented levels of change and human advancement across industries, countries, economies. The entire planet, I think, is ripe for a pretty magical, and epic era in front of us. It seems like the limit of what's possible is really only the limit of what people can dream and people can imagine. I think there's big dreams at play. I think there's even bigger challenges that require dreamers, in sort of a world that's changing around us.
The one thing that doesn't change, that we believe is constant, is that most of the things that people dream they can accomplish are really only made possible by groups of people that work together, by teams. Teamwork is really the currency of our future. You get it right, I think, and big, incredible things happen. You get it wrong, and I think the consequences, potentially to business and to all of us, are real. Teamwork is hard. It has been complicated since we crawled out of caves and tried to hunt big animals together. Inside of modern business, it is very complex and daunting, and it can feel like a journey sometimes without a destination. There's a lot of really important work that happens in the cracks.
Oftentimes you might start with something that you want to accomplish, and you need to rally a bunch of people around that idea to bring them together in pursuit of what you're trying to get done. You need to plan things together, organize and structure sort of the activities that are going to accomplish that goal. Create the work product that actually informs whatever you're trying to accomplish. Constant levels of communication and iteration. Sometimes you get it wrong and you have to restart, back the truck up, and then go forward again. Once you actually get there's oftentimes needs to improve or opportunities to make it better and improve and kind of iterate over time. It's a mess. It's getting harder.
Today's knowledge workers and teams face an increasing number of hurdles that they need to jump over in order to accomplish what they're trying to get done. There are more products and tools inside of workplaces that, in some cases, increase the fragmentation, and the siloing of information that people can use and leverage to move forward. There's new technologies like artificial intelligence and machine learning and virtual reality that are and will continue to change the way that people connect with each other, and also change the way that we connect with data that can help us make decisions or move forward. Finally, there are new ways of working. Having people be more distributed and potentially remote. The requirements for intra-company and inter-team connection and communication is vital and increasing, which makes sort of the pursuit of something together a little bit more difficult.
Teams sit at the center of all of that complexity and change and constant pressure. Atlassians wake up every single day with a religion and a relentless pursuit around helping them. Our mission as a company is to unleash the potential of every group of people that is trying to do something together. We do this uniquely. If you heard the keynote, you heard Mike talk a little bit about this, and if you saw Scott's presentation last night, a little bit more detail. Our perspective on teamwork is grounded sort of in a belief around openness. We believe when teams work in the open, then you unleash the full potential of what they can accomplish together. We take a pretty broad perspective around openness. It begins with our products, which are open by default. Mike talked about that.
Simpler, more open ways to connect people around the structured work and activity and tasks. Simpler ways to open up the content that people are creating for discovery and sharing and iteration and improvement. From the very beginning, our products began with this principle of open first and then secure second, as opposed to a single thing that only I can see, and then I have to figure out how to let people leverage and use it. We also believe in diversity of thought and the diversity of perspective of teams and how they're composed. We believe our products help create cultures. They help imbue personality and life into teams.
It can seem like a small thing, but it makes a huge difference in terms of the output and the outcome of what people are trying to accomplish together when they can bring their full selves to work. Even our own business model, which I'm going to spend some time talking about, is grounded and rooted in this idea of openness. Okay, how specifically do we help? What do we do? At a high level, Atlassian helps teams better plan and organize and track what they're trying to do together. It helps people create and share information and content. It helps them communicate through all of the iterative cycles and restarts and reboots and sort of teamwork moments that they have. It helps teams better service and support both their teammates and the other parts of the business that are there to help.
We offer a distinct portfolio of products across the bottom that really do specific things for teams and what they're trying to accomplish. I'll talk about them in a little bit more detail here. If you listen to the keynote, you heard us highlight the Atlassian Teamwork Platform, which is really represented in the center of the screen. You should think of this as a shared set of collaborative capabilities that are imbued or sort of bring collaborative life into each of the distinct products at the bottom. They allow people a common way to do things like react to an object inside of a particular Atlassian product in a very common, similar way. Whether that's a task, or a deadline, or a project, or a page, or a message, or a team channel, they can have discussions and conversations about those same objects.
They can @mention different teammates to bring them into those objects to work on them. They can share files and iterate on other files in the context of that particular object. Meetings, decisions, tasks, activities, there's a whole bunch of those elements that you saw in the keynote that effectively come together in a different context around a tool that we're providing to different teams. On a whole, our products are really generic. They're used by a pretty wide variety of different teams to accomplish similar things. The organization and planning and content discovery and creation and communication that I mentioned earlier. Really quickly, to highlight the products themselves, I'm going to start with the Jira family, which includes three separate Jira products, Jira Software, Jira Service Desk, and Jira Core.
These are all built on the Jira platform, which you should really think of as providing work management and workflow. Each of these products basically take that workflow and work management and apply it to a particular team or a team pursuit. Jira Software, it's really dedicated to the painstaking art of taking and managing and organizing a technology project, which is very difficult. It's important to point out that the users of this system are not just developers. In fact, I think people that know how to code tend to be the minority user of Jira Software, because software and technology projects, if you roam the halls and talk to customers, are now a full team activity inside of businesses. They include designers, product managers, people in QA, people in marketing, people in finance, business stakeholders.
Jira brings those diverse perspectives and people together and allows them to connect around the thing that they have in common, which is they're trying to bring technology innovation into their businesses or bring it forward. Jira Service Desk adds a specific set of capabilities on top of that workflow platform for teams that are servicing other teams. Capabilities like SLA management and different escalation paths, agent queues. If I'm providing a service as part of IT on an IT helpdesk, I have a very simple perspective on the things that I need to get done and the requests that people have of me within IT, as an example. Detailed CSAT reporting, different things that are relevant to that use case. Important to know that I think our perspective on service desks is that every team inside of the business is a candidate for this product.
Not just IT, which is the canonical one, but marketing, finance, sales, legal, they're all on one level supporting the other teams and the other parts of their business in some way. Most of the work that is requested of them is brokered through email, which has no workflow at all. It's just an inbox with a bunch of things that says, "Please help me." Jira Core simplifies and trims down some of the software and technical specific aspects of Jira Software into a collaborative project platform that's really purpose-built for business teams. Since introducing it a couple of years ago, we've seen significant adoption expansion by business functions from finance and HR to operations, just around that aspect of collaborative project management. I'm not going to cover every product. I'm going to try to go quick. The major ones here worth highlighting.
Confluence is a system that helps teams create and share content, and that content can literally be anything. It sort of depends on who you are. For software teams, it might be a specification document or a business case. For marketing, it's a launch plan. For HR, it's an employee manual. For operations, it's a run sheet. You get the idea. That's sort of the versatility of the product. Trello is a product that we acquired, as many of you know. Michael described it, so I won't. Simple scaffolding. You can organize cards on lists and boards, and its versatility is also its primary superpower. Everything from wedding planning to a real estate team managing their facility site selection. Bitbucket is the one product on the list that's purpose-built for people that write and manage and share code together.
Where Confluence is a collaboration system for the code that all of us in the room know how to write, English and words and sentences and phrases, Bitbucket does the same thing for actual code, which is a different kind of language. It's a collaborative place for developers to share, store, iterate code that they're writing. It competes with a product called GitHub and GitLab in the open source space. Stride, we introduced last week and then talked a little bit about on the main stage, I think a really exciting new entrant in the communication space. It combines, for the first time, I think, three really important capabilities around team messaging, real-time communication through video, screen sharing, and other real-time connection facilities. Then capability around action, moving conversation towards outcomes that are then trackable.
It would compete with products like Slack and Facebook Messenger and Microsoft Teams, but it also is an alternative to things like Webex and GoToMeeting, or Google Hangouts, where you're having to context switch to some other thing to connect in real time, which people need to do. Statuspage also came to us via an acquisition. It offers teams and companies a really reliable way to communicate to their end customers what's going on with their service. Think of it as the signal bars on your mobile phone. When you're talking to somebody and you do the classic, "Can you hear me now? Can you hear me now?" You usually look at those little signal bars to understand whether or not you've got a signal. That's what Statuspage does for users that are end users of some other service.
Because we now live in a service-driven world, even the simple things are constructed with a lot of other things that have their own levels of reliability and uptime. So having that communication endpoint to customers is vital. Finally, the Atlassian Marketplace, which is an online store for thousands of apps that extend our own products, building new capability on them, applying them to specific use cases inside of businesses. I think if you get a chance to walk the expo floor, you'll be blown away by just that small sample of applications and integrations from small companies in Finland to PagerDuty and InVision and bigger software companies that are building on top of the richness that our products provide. That's a quick primer on products. Why do we win? Why do customers choose us?
For starters, you're going to hear a lot about this from Sri, we invest a lot in making them absolutely awesome. Because our products are our sales organization. They are our most important influence to conversion and selling. If our products can't convince the customer that they're going to do what they need, we really have no backup plan, and it's very hard, by the way, if your product doesn't do the job to convince the customer that it will. We also win because we have multiple cracks at bat. We can land with almost any of the products that I just described and expand with any of the products that I just described. I'll talk a bit about how hard we work to smooth the acquisition path for customers. Really, removing friction is a religion at Atlassian and sort of a central theme of our model.
Simplicity is really our North Star. The easier that we make our products to use, the more that they are likely to expand. We use pricing as an advantage. Our pricing is very disruptive. I'll talk about the philosophy of our pricing pairs the highest quality, best product with the most incredible best price. Finally, the extensibility that our ecosystem is providing and the value they're building on top of our products is another reason why we win. Who we compete with. We compete in markets with a lot of different categories and a variety of different alternatives and competitors. In a lot of the categories, we have legacy incumbents that we're displacing. So a company like BMC in the help desk space or maybe even Lotus Notes for team collaboration. There's a variety of new upstarts and pure plays from Slack to Quip or Zendesk.
We win by building a better product over time at a better value that customers can see and feel. There's open source alternatives in some of the categories, we won't compete there on price. We'll compete based on our investments in innovation, the quality of our products, and the level of service that we provide customers. Most importantly, I think the biggest category that we're displacing is the white space that exists around just the old way of connecting and creating and sharing information, which is a bunch of documents or blobs of information that you attach from an email, you fling to 10 or 15 people, and you keep your fingers crossed. Mostly you go home at night complaining about how hard work is.
There's just a lot of opportunity for our products to, when teams use them, our products are a revolution, I think, compared to that. It's interesting to note that I think every company that I mentioned, it's hard to think of a competitor that doesn't use one or more of our products, which I think is an interesting point. That's our advantage. I hope you get to meet customers as you walk around the show. I think what you'll see is a pretty wide range of teams and use cases of companies and industries of all shapes and sizes, from Virgin to The Daily Telegraph in the U.K., to right in our backyard here, Tesla. That's chapter one done. We're level set on who we are and what we do. Beyond our products, what makes us unique is our model.
Our model is pretty straightforward and easy to understand. The math is simple. Great products, low pricing, and a focus on automation and removing friction, constantly removing friction. They all combine together to acquire a large number of new customers that we grow with over time. The cornerstone of what makes our model work is a focus on building incredible products, as I mentioned before, Sri's going to talk about it. We pursue a really big market. There's close to 1 billion knowledge workers and countless teams that are trying to improve how they work together. To reach as much of that market as we can, we keep prices compelling and competitive, we make it really easy for people to get started. We pursue a large volume of that market consistently. We sell online, right?
To reach teams in Zimbabwe, as well as teams here in Silicon Valley, we use the internet, we allow customers to self-serve and get started and grow with us. To do that means we're open. We share pricing, we make our products easy to try. Because the product is the cornerstone of how this model works, it has to be awesome. It has to constantly improve. That brings us all the way back to the beginning of this virtuous cycle with investing in making the products ever greater, ever more awesome, ever more compelling. We're different, I think you know by now, than most enterprise software companies that you're used to or you cover. Not to disparage them too much, most of them scrimp on product, right? It's virtually impossible. They don't want you to try the thing on your own.
They want to hold your hand through it because they haven't built the product to do the heavy lifting. They have to compensate, I think, for a lack of investment with a bunch of other mechanisms. As a result, products tend to be pretty expensive. There's heavy discounting and negotiation. All of that means that they begin top-down, their conversion cycles are measured in months, sometimes years. They're cumbersome. Their business models are offline. Some of these work really well, by the way, for certain categories of products and markets. We are very different. We focused from the very beginning on do it yourself and enabling our customers to self-serve. I think, by the way, if you just think about how you interact with things that you're evaluating or looking for, that's what customers want.
That's what I want when I go to the web. I actually don't want to have. Even when I go into the Gap to buy a new pair of pants, I just want the person to leave me alone. Just let me find the pants, put them on, and if they fit, great. I think online or with technology that we're evaluating, it's very similar. We really only seek help if we need it. If we don't need it, we won't. That's what we're focused on. We price to fit into every budget. We're predictable, so customers understand from the beginning of their journey what things may look like as they grow with us. There's no hidden pricing. We enter bottoms up through the team.
By the time we register on a C-level executive's radar, we have recruited, through our products, legions of fans and champions that are doing whatever convincing may be required to continue to grow and expand. Our cycles are measured in days. Our growth cycles are measured longer, naturally, but the entire model is online. We put a lot of energy towards helping the customer help themselves, and I'll talk about how we complement that by being available whenever they need and wherever they need. We're always going back to understanding whether or not we can fix future questions that customers had for the next customer without forcing them to say, "I need your help." We do all the things that I mentioned. Pricing online. Products are free to try. There's no forms or gates.
We're not harvesting a bunch of email addresses to give you something that might be useful or valuable. If we've written a white paper on how to be a better team, we want everybody to read that white paper. If it's really effective, they'll come back to us for either additional content or more importantly, our products. No custom contracts. No hidden pricing. Everything is designed to be straightforward and simple, and we're constantly making it better. Land and expand models seem to be par for the course. I'm sure in most of the companies you cover, they talk about land and expand. I believe ours is more multidimensional than most that you'll see. For starters, as I mentioned earlier, we can land with any number of different products. We don't care.
Team can start with Jira, or they can start with Bitbucket, or they can start with Confluence or Stride or Trello or HipChat. That gives us an opportunity to both plant that product more deeply in their organization, expanding the usage of it by the initial team and expanding other teams. Also complementing that with other things that we know will make a difference to their teams. We begin our journey with either someone who's looking for something that we have that could fix a problem that they know about, that they've identified, or what is very common for us, we start with either somebody who has existing experience with our products or someone that started who has recommended us strongly by somebody who has existing experience with our products. That's sort of point one.
That original team starts cranking. The built-in network effects of our products, things like simpler ways to invite, to share, reports and notifications. We benefit because we built team software in the features and capabilities of our products to basically connect and draw more people into it. People also help us spread. Check out how we're doing this, said from one colleague to another colleague, is an incredibly potent expansion trigger for any company. It especially is for us. As more teams onboard and as teams grow from one product to another, usage and products naturally spread, so too does account spend. Over time, we win more and more hearts and minds of people inside of our customer base. In the process, we're creating people that join the Atlassian movement, that become part of our word-of-mouth engine, that become fanatics, evangelists, promoters, champions.
If you walk the halls, you'll meet a lot of them. The reason that they're champions is because they have first-hand experience with how our products make their lives at work better and easier, and how our products make their teams more effective. These champions either roam around their own organizations, talking about how our products can make a bigger difference for their companies, or if they move from one company to another and we're not there, it's typically the first thing that they do, is figure out a way to bring us in. It's an incredibly vital part of how our model works. Gone through two sections, we're going to spend a little bit of time here on land and expand, how the whole thing works in practice, and what are the things that we do specifically to improve and accelerate it.
This is the customer journey, which is probably common. It is common for nearly any software product. The customer is trying to learn about what the thing does and whether or not it can fix what they need fixing. Ideally, they could try it, without a lot of friction and pain and time and cost and energy. If it does the thing that they thought it would do and has the potential to fix the problems that they need fixing, they'll buy it. If it really works, they might expand it or buy other things. If they believe in the company, ideally, they become promoters in the way that I described. This is a simplified view of how we prosecute the customer through this journey.
We spend a bunch of money, time, and energy building awareness for the things we do, for the problems we solve, and also for our brand, for our company, and what we believe in. We make it really easy, as I mentioned, for people to discover us, to learn about the stuff that we make and how it's used. The product and our smarts and investments around data, and I'll talk about some of these techniques a little bit more deeply in the deck, does most of the heavy lifting around conversion. We are constantly focused on improving that and getting better with it over time. Underpinning this, is a constant energy around delighting the customer and providing additional value, not just in the context of the products, but in the context of maybe practices or plays or guides or thought leadership around teamwork.
We organize customers together in contexts like the one that you're at, Atlassian Summit. We invest in offline and online community and user group programs. We put a lot of energy into being a company that demonstrates how much we care about customers and what they're trying to accomplish. I think that you'd see that reflected and echoed by customers that are here. I'm going to go through each one of these stages with a bit of detail. I'm going to start with stage 1 and the work that we do to build demand and attract new customers. We have a world-class marketing organization, absolutely world-class. We focus on generating demand in all the smart places that you would expect us to, and I believe in a lot of smart places where you may not expect us.
Search and discovery is a big part of our demand generation attraction because people are going to look for things that are related to what we do. We heavily optimize around search engine optimization to be at the right place at the right time when people are looking for things that can help. In addition to that, you might hear us on NPR or see us on TechCrunch. We'll do smart out-of-home awareness building. This is an example of a creative campaign that we ran specifically targeting developers with Bitbucket when Microsoft brought herds of developers to town for the Microsoft Build Conference. We have strong places where teams and people congregate to learn more about how to advance the state of the art for teamwork, industry trade shows, and events in places like New York to Tokyo.
The result of all this demand generation and word of mouth is a tremendous amount of traffic driven to our web properties every single day. This is atlassian.com over the 6 months between March and August. We averaged 2 million monthly unique visitors that came to learn about what we do and how our products could help. That is just one of many consumer scale funnels that we operate. Bitbucket.com is a completely separate high traffic funnel, slightly larger than Atlassian itself, just focused on developers that are learning about that one product and how it can help them. Trello is nearly 3 times as large as either one of those.
HipChat, Stride, Statuspage, they're all additional examples of high traffic funnels that we're using to aggregate new users every day, with tools that they can use where we can begin a journey around teamwork and improving their teams with them. Now, what's incredible is that the majority of people that visit us are earned from the strong reputation that we have in the market around our brand and our products. Word of mouth and organic channels, people searching for Jira or Trello or Bitbucket or Atlassian because somebody has said, "You have to check that thing out, it's awesome," account for about 87% of traffic driven to atlassian.com as just an example. We augment and complement that with paid channels. We make sure that we're reaching people who might not have heard about us, but most importantly, augmenting that powerful word of mouth, wherever people might wander.
That helps us reach that high volume that I mentioned. Our goal here is to drive this ever-increasing percentage of people that are coming to learn about us and kick the tires on the product towards the products themselves. Much of the smarts and data-driven techniques I'm going to talk about a little bit later. We are focused on constantly improving the yield from all this interest through experimentation and trying to nudge people towards the products that they're looking at, and then once they're in the product, towards capabilities and product that can help activate them better. I'm pulling the punchline a little bit forward here in the deck, but I think the result of all this is pretty incredible. Unlike most of the business that you'll cover. If you explode a typical day at Atlassian, they all look similar.
They grow over time, but they look similar. This is August 8th, which was a Tuesday, so about a month ago. On this particular day, we conducted $2.8 million in sales from 6,800 unique companies, unique domains that transacted with us commercially in one shape or form. Some of these companies are going to be buying their first product. Could buy 10-user license, a 5-user license, a 100-user license, a 1,000-user license. We don't discriminate. We'll take them all. Some could be existing customers that are buying a second, or a third, or a fourth product. Some could be existing customers that are adding additional users to one product that they began with. Many customers here are going to be renewing either active cloud subscriptions or maintenance for server products. Some are going to be buying add-ons from our marketplace, through the Atlassian Marketplace.
The single biggest transaction on this day was $75,000. I think it just shows the high velocity, high volume of customers that we're serving on a regular basis. Again, in typical Atlassian fashion, the days before and after look very similar to this. Murray's going to talk about the linearity of the business and the shape of that over time. On that same day, 13,000 unique companies started a trial of one or more of our products. Again, this is multidimensional. Some are going to be new customers that are starting for the first time.
Many are going to be existing customers that are maybe trialing another part of the business, another part of that same domain, who might not even know that another part of the business is using Confluence or Trello, might start their own independent version for their own independent team, because they're a part of the business that might not be connected to where Atlassian energy already exists. Some of these are going to convert into paid products, some of them are going to convert into free products. Some of them are going to convert into paid products that don't qualify for our customer definition. It's just a really incredible volume of both commerce and pipeline that we're building. That's phase 1 of just attraction with a little bit of the results and the punchline at the end. Let's talk about converting that demand into trials and customers.
The three funnels that I mentioned earlier generate a pretty remarkable amount of new trial pipeline for the company. On average, over that same six-month period between March and August, 5,000 new Bitbucket trial accounts were created each day. 3,000 new trials of Jira Service Desk, Confluence were created, and 24,000 new sign-ups of Trello on a single day. We augment all of the automated conversion energy and expansion energy that I'll get to in more detail. We augment that with a couple of different teams to better service and support customers alongside their evaluation and growth journey. We think a lot about touch being smart.
We want to be placed at the right time. We use data to make sure if people do need help, that we might be a little bit more proactive with letting them know they can get it. We don't want to get in the customer's way if they don't need us. The teams that do this internally, we call advocates. They are assistance teams in various stages of that life cycle. Externally, we also have grown a very vibrant and important global partner network that helps us further remove friction from customers. They also help us augment our awareness-building efforts. There are five distinct advocate teams. The number of Atlassians on each of these teams is on the slide, so you can see how big they are. I'll describe them really quickly. Loyalty advocates assist with post-sale renewal, and retention work.
Customer advocates assist mostly with billing and customer service issues. Remember, we've got 90,000 customers, as part of our official customer definition. We've got lots more customers that are not part of that. Maybe they're using free products, or they're spending less than is qualified in the definition. There's just a lot of customers that may have a credit card declined or a billing issue. Maybe they have a quote they generated that they can't figure out how to regenerate. There's always going to be little friction points sort of in the billing and purchasing cycle that we want to make sure that we're right there on the spot to help them with. Product advocates assist customers that have questions about our products, really in the pre-sale part of the journey. If they can't find an answer themselves, we let them know that we will help them.
We'll bend over backwards to provide an absolutely killer answer. Then they go back to the part of the funnel that I've already described. I'm going to talk a little bit more about how they work in a moment. Enterprise advocates are dedicated to our largest existing customers and assist with large account growth and scale, and they help these customers better navigate some of our premium enterprise offerings and enterprise services. They also will connect them to partners that have expertise around scale and deployment and customization that they might need. I'm going to talk about them in the expansion section because they are purely focused on large account expansion. Partner advocates assist partners with service-related needs for their business. They might need help with quote generation or co-marketing engagements and things like that.
It's important to note that these teams actually don't grow that much over time, in part because we are constantly focused on improving self-service through all these touchpoints. Creating simpler, more effective ways for customers to do the things that they're asking for help with on their own. We make the discover to buy path, as I mentioned, really easy and friction-free. A customer, if you flip it another way around and they're going from kind of learning to buying, they can sign up painlessly for a trial of the product, and then go through that trial experience where the product is doing a lot of the things that maybe a person would do if I sat right next to them on the computer and tried to guide them through how to get the most out of the product.
Naturally, they can buy online, with a credit card or kind of a variety of other payment methods. 97% of customers that are evaluating what we do just go through this without ever saying, "Hey, I need extra help. I couldn't figure this out. Please help me." For the 3% that opt in, we basically let everyone know, 100% of them, "If at any moment you need additional help, we're there to help you." That's the job of the product advocate. They're there at various phases of this evaluation journey, letting customers know that, if they need extra TLC or extra help, they're there for you. We'll do things like webinars or live demos or live Q&A where people can get questions answered.
There's forms sprinkled throughout, a whole bunch of interaction points that let them basically ask quickly a question, and then that gets routed to the person that has the best answer. During their product evaluation, we'll send them an email that lets them know, "If you run into any snags and you want somebody to help you, simply hit reply and we're right there." In product, we'll pop open chats, and say, "Hey, it looks like you might be struggling with X, Y, and Z. Can I actually help you get through that better?" Just really smart, thoughtful ways of making sure we're complementing all this automation with a human contact point that can help where it's needed. Our global partner program is also a really important complement to all the automation and self-service and smart touch aspects in our own business.
The partner program offers a pretty broad set of complementary things to what we do. There's really 4 types of partners. Solution partners, which I'll talk about in a little more detail, provide sales consulting, technical services around implementation, customizing our products for specific things that our customers might need them. Training partners deliver artificial courseware and certification programs in live classroom settings all over the world, complementing what our own in-house teams do and helping us scale that. The marketplace vendors we've talked about, they build apps and extend our capability and largely improve and increase the stickiness of our own products. Finally, we have official relationships with corporate and government resellers that offer additional purchasing assistance to customers that procure through these groups. Also important to note that the process for resellers, for companies that resell our products, is also heavily automated.
Companies that are reselling buy our products online in the same way a customer does through their partner portals. We have a lot of investment in making sure that that's really seamless and awesome. For the solution partners, there's a series of 3 levels that we use to distinguish and qualify them. All of them sign an official solution partner agreement. Once in the program, there's a combination of investments they need to make in sales enablement and technical professional accreditation in order to qualify to be in these levels. These levels are meaningful because we give referrals through them. Customers use the directory to discover them. On the far right. Oh, it bears stating that the technical accreditation is pretty tough. We have a high bar.
If you're going to have somebody that knows how to do things with our products, you actually have to study and invest in passing the certification and accreditation exam. At the silver level, there's no minimum annual sales requirement, but there are requirements for the technical accreditation and sales enablement. Gold, minimum annual sales of $50K with the technical and sales requirement you see on the screen. For platinum, annual sales of $250K, with a meaningful difference in investment in technical accreditation and sales enablement. We've had the solution partner program for about a decade, and globally have more than 400 partners in each geo. Pretty good reach and I think we're in a lot of markets, where these companies are representing our products and more importantly, representing sort of ways to configure and extend them.
Partners tend to index a little higher in regions that are non-English speaking, maybe where customers or companies in those regions are a little less familiar or comfortable procuring online. They might want to have a conversation with somebody in French or Korean or Japanese. They might want to understand our products in a little bit more detail. We do, by the way, localize our products in 13 languages, 15 languages, and the web and the customer flow is also localized. They also index higher in larger enterprises and organizations where naturally there is a bigger opportunity for the consulting and technical services that is really the bedrock of their businesses. There's more opportunity for growth and expansion. Channel features, importantly in our mix in the overall business, they touch, in one way, shape, or form, about a third of our revenue, where they become the reseller.
The channel both brings in new business to us, but also once they establish a relationship with a customer, may become the procurement vehicle, where they're handling the renewal and the expansion, and other things that the customer might want them to do. They can remove friction points around foreign currency and local billing and contacts, and a whole bunch of other things in addition to the consulting and services. If you think about how to dimensionalize their contribution by geo, it kind of maps where they are, and as I mentioned earlier, where some of those markets might want to engage more directly with somebody locally. The figs in Europe and Japan, China, Korea, places like that. If you walk the expo hall, you'll meet a bunch of them. There's a flavor of 3 I thought would just be interesting.
CGI, the huge Canadian technology consultancy with about 70,000 employees. They've been a partner for about a year and a half. We've had growing practices in all the big SIs, but I think this is a really good example of a successful growth practice. They are also, as most of the SIs are, a pretty large customer of all of our products. Cprime is a solution partner that joined in late 2012. It's a good example, I think, of a prototypical partner of ours. With 150 employees, a little bit more boutique. They specialize specifically in Agile and in the transformation of technology teams and technology development, operating primarily in the U.S. Valiantys is a partner that's been around for about eight years. So kind of a veteran in the program. About 80 employees, headquartered in France, but operate throughout Europe. They have about 80 employees.
They focus on more business-oriented collaboration opportunities around Confluence and certainly ITSM and service collaboration with Jira Service Desk. The net result of this conversion energy is consistent and steady rate of new customer acquisition over time. Recently, getting close to 90,000 customers with an active monthly subscription of close to $10 or greater that Murray will help define for you. That's conversion cycle. Let's move on to the final phase, expansion. There's a couple of different dimensions of expand for us. One of the most important really begins with activating that first user and that first team. We put a lot of energy on that, for good reason.
Naturally, if we don't stick, if we don't engage, if we don't get the first team of 10 or 50 or 100 really loving that first product, we have no opportunity to grow them by either expanding usage or cross-selling. There's a lot of focus that you'll see on making sure that we get better and better and better at activation, at getting you hooked. That is the surest path to the other expansion dimensions. The other ones are maybe obvious. Cross-selling is a big one, where we've got additional products within the portfolio that we want people to be attached to. We do this, by the way, selectively at point of sale. If you go to the web, I think you'll see some smart approaches around bundling and packaging.
Our philosophy is really rooted around helping the customer get started and nail the first thing they came to us to look for. We don't want to overload the cart with five or six or eight things that cognitively they weren't prepared to fully wrap their heads around. We also expand through apps in the marketplace, which is another key point of expansion for us. Part of the expansion effort, for many years, has been to drive usage across an increasing range of teams. This is just one example of that with just the Jira product family. In a survey last month of about 2,000 random users across cloud versions of Jira Software, Jira Service Desk, Jira Core, just over half of the respondents identified as being on a software team. This is in response to the question: What team are you on?
Bear in mind that, again, the users that are identifying there are not all developers. Probably the minority of them are developers. That group includes designers, people within product management. They might associate with, "My job when I come to work, I'm focused on this software project or this kind of innovation exercise." 20% identified as being within IT, probably largely users of Jira Service Desk, maybe a little bit of Jira Core and Jira Software. About a third identified as being on a business team, marketing, HR, finance, that sort of thing. This is just one example. I think when you look at products like Trello and Confluence, probably the dimensions of these skew more towards IT and business functions than the Jira family would.
Much of our expand energy is driven through data that we capture about customers, both as they begin their journey with us and as they use our products. We track a lot of what they do, and we use that to get better and better and better over time at targeting them with things that we think, or in some cases believe, that they're ripe for, and that isn't just a product. In this particular example, where you look at, like Sally Jones, who's a Jira admin, we track what company Sally's in, what industry that is, how many employees they have, her lifetime with us, her actually activity in product, what she does, maybe parts of our web properties that she's visited to learn about different things. We can then use that data to target Sally with particular offers.
In some cases, it could be a cross-sell. In some cases, it could be, "Hey, here's a piece of content that could help you be a better Jira admin." Wherever Sally moves across the diaspora of touchpoints, both in product and across the web, we're smartly nudging her towards things that we think would be best useful. Again, the goal here is not to be obnoxious and not to have this feel like an advertisement. It really should feel like an effective and thoughtful recommendation for what to do next. I'm going to give you a couple of examples because I think it's important to understand. In this example, we're focusing on what is really vital to begin with, active, growing, happy usage of the thing that you first bought from us, in this case, Confluence.
Confluence has a feature that lets people comment and discuss things at the bottom of the page, which is what you see here. Confluence also lets you highlight text on the page and actually have that same discussion in the context of the page itself, where your commenting is actually attached to the thing that you're referring to. That feature is a little bit harder to discover. The one at the bottom is obvious. You just see it. It's in your face.
An example of how we would use the engagement platform is to observe that a user is copying text from the page, pasting it into the bottom comment, then beginning to say, "Hey, what do you guys think about this thing that I've just shared with you?" If we see them do that a couple of times, we might say, "Hey, do you know that there is maybe a better way for you to do this? You can actually do this in line. Let me show you." Now we know we'll do this because we know that engagement of the product activation and use of this feature can lead to better expansion, higher NPS in products, and happier usage, which then leads to an opportunity to potentially sell this Confluence customer the other things.
This is a simple cross-sale or cross-flow example, where the user is a Jira user, I want to let them know about Confluence and start their journey with that particular product. This one's pretty easy. In the product selector at the bottom, we can simply load a Confluence trial. Again, when they click on that, Confluence is there. They didn't have to sign up for anything. They can either learn about it or they can click try, and boom, it's just right in the product. Then the whole Confluence onboarding magic takes over from there. That's kind of a simple cross flow. This one's a bit more sophisticated. Here's another example where I'm trying to sell a Jira user Confluence. In this particular example, we've added a function to Jira called Pages, for this user. You see it right there on the left.
They'll click it, they'll see a series of pages that they create. Some might be specific to Jira, like reports. Some might be a requirements document or a sprint plan. When they click on that, because it's a function that's really relevant to what they're working on, we can let them know, "Hey, this particular page is powered by Confluence, that you're not currently signed up or using for. Do you want to do it?" Then deep link them or take them directly into that particular function, where they see the true value and marriage of those products working together. Final example. Again, ratcheting up the sophistication. In this particular example, same Jira screen that you might not be familiar to, we might notice that, boy, this particular user is writing a lot of text.
I mean, they're writing pages of text in the Jira issue. Or maybe they're copy and pasting links from Google Docs, or maybe they're attaching Word documents. The main thing we can see here is actually, do you know that there are easier ways for you to do what you're trying to do? Again, same thing. Why don't you just click here and create a page? Then boom, you take them right in. Really friction, seamless way to just get people onto the next thing. We use this system to help nurture and nudge thousands of customers each day, and we're continually refining and improving it. There's one other important dimension of expansion that involves growing large enterprises as they scale, both in terms of the number of users and the number of products.
Ultimately, make a decision to potentially standardize on us, so we become the standard operating system for how teams work. Here, there's a couple of different expansion points for large customers. In the server product family, the Data Center editions provide clustering and high availability and performance at scale. You might be running Jira on a single machine supporting 1,000 people, and at that point, we recommend that you really upgrade to the Data Center edition. In cloud, we talked today about Identity Manager, which offers a premium set of features around security and administration and management of cloud instances. Then independent of deployment platform, there's a series of services that are offered for large companies, from our technical account manager program to additional certification to premier support. This is the world of the enterprise advocate.
Again, we make that discover-to-buy path the same, friction-free. You can upgrade your server instance to Data Center. Again, the server instance can let you know that, boy, you should really evaluate this upgrade path. But you might need extra help. We take 1,000 of our largest customers, and we have enterprise advocates dedicated to growing and nurturing them through this journey. They do things like help with invoice consolidation. Maybe a large company wants to understand, but we've got lots of different pockets of purchasing. We actually want to help consolidate those on sort of a single invoice or a single bill. They'll do solution partner referral, and naturally, they'll help introduce or direct people towards the premium upgrade paths. Important to note, again, that most of the customers that buy the Data Center editions do so self-service.
You add it to the shopping cart, you go through your regular renewal, and you say, "Actually, I want to upgrade to the Data Center edition." You can just do that self-service like you can everything else. Expand is working. In the past two years, we've more than doubled the cohort of customers spending more than $50,000 annually with us. In that same period, customers spending about a half a million dollars have grown 5x. From 14 in FY 2015 to 75 in 2017. Here's two examples that highlight kind of this continual growth and expansion. Large global consulting firm, and you can see they may have started small and kind of grew gracefully over a long period of time, and then eventually reached a point where they began to ratchet up their investments in things like Data Center, and Marketplace expansion, and additional products.
Same kind of optical view of a large technology company. Both these companies are spending just over $1.5 million. We've got lots of journey left in both of them in terms of products they don't own and users that aren't using Atlassian. Retention is naturally a really important and vital part of expansion. If we can't keep the customers we have, we don't have an opportunity to sell them other things. Retention is also strong. When we look at that same cohort of customers spending more than $50,000 annually, we've achieved a 98% logo retention rate, keeping them around. I think that demonstrates both the stickiness and the indispensability of our products for these organizations. It's also worth noting that for this cohort of companies spending more than $50,000, 90% of those customers own three or more of our products.
The net effect of all this. A chart that you've probably seen from us in the past, because it's such an awesome chart, is a continuous rate of both new customer acquisition, retention, and expansion over time. This chart shows annual cohorts dating back to our very beginning. Each colored band is a class of customers that bought in that first year. It shows the increasing spend as customers renew, expand usage, buy additional products from both us and apps in the marketplace. There's a couple things worth highlighting quickly. First is that initial fiscal 2002 class that you can't even see because it's a crappy shade of blue and it's really small. They spent $18,000 in 2002. That same cohort in fiscal 2017 spent $1.4 million. Second, maybe no surprise, but the cohorts are growing bigger over time.
We've got more products to expand them with. We add things to the portfolio. We make a bunch of improvements in the expansion techniques. We get smarter and smarter and better and better at the things that I described. Finally, the size of each initial cohort is growing and becoming larger over time. Maybe that's no surprise because we've got a lot more things, both at point of sale and in that initial year of their journey, to expand them with. We get better at focusing on activating them in the very beginning. As many of you know, we are pretty efficient. We do all of what I've just described for 16% of revenue. I think we're both very thoughtful about our dollars and very smart about the investments we make to continue to land, grow, and expand our customers.
I wanted to leave you with just two examples of how all of this, the sales and marketing machinery that I described, how it works from a customer's perspective. Two quick videos.
If you fly in an airplane, Rockwell Collins touched your life and helped keep you safe. If you use the GPS in your cellphone, if you rely on the military to protect your country, Rockwell Collins is involved in that. It's one of the most important things is that we are trusted to provide safe equipment that you can rely on to do what it's intended to do.
We were working in a very siloed environment where we had a lot of single tools that weren't integrated with one another, and all of our teams were forced to log in independently to each of those tools, and they never really talked to one another.
The areas where we're not specifically driving content, the users are picking that up and running with it. Confluence is a great example of that. We had very little prescription in terms of how people in projects were using Confluence, and projects picked that up and ran with it.
One really good example of how the tools are viral in a way, and they attract other people in, is through the dashboard capability. When you share a dashboard with someone and they see that for the first time, it's really kind of mind-blowing. It even takes a little bit to realize that this isn't just a static report. This is actually a live indication of where things are at in the project.
One of the examples of Atlassian tools, I'll call it hopping, is we had Jill come to us from manufacturing. We were delivering this to engineering and Jill got interested and wanted to know if her team in operations could use it. We worked together for a few weeks, got them set up, and they were off and running. They added about 3,000 manufacturing users.
This is not the kind of thing where we're pushing tools out to our users. We're providing these tools, users are clamoring to bring those tools into their projects.
One more quick one from Dice.
DHI is a digital media company. We maintain and run multiple job boards, with dice.com being the flagship lead board. For us, Confluence started as an IT-only tool. Along the way, our customer support department asked if we had any tools that could help them track projects simply and document simply. I sat down with the head of customer support and showed her Confluence.
Brian actually helped me to set up my first space, and then I got my managers involved to look it over and see what should we have in there, what do we need to use it for. Our first big project with it was creating our new hire onboarding program and putting everything in there, all of our documentation, progress reports, everything. Since my department, customer success, implemented Confluence, I demoed it to sales teams and to our customer service team, and they have all started using Confluence for similar uses as well.
Use of Confluence by HR was not a technology project. It wasn't something we asked them to do. They gravitated towards what was easy. HR came to technology to say, "Could you help us? Could we create our HR portal in Confluence?" We went through and we looked at their notes and looked what they're wanting to do and said, "Yes, you can create everything you need." Another big component of our technology roadmap and strategy is around Atlassian integration, given that we work our dev teams on Jira, and we work on HipChat, and we work through Confluence. JSD was a natural evolution. With the Atlassian suite, it's not an executive saying, "I've used this tool before. We're going to implement it here." It's usually from the ground up.
You get a development team, a technology team that comes in and says, "We've gotten this new tool, and it's working really well." Then like wildfire, it expands. That's how the adoption has occurred here. That's how I learned about it when it was first introduced to me, is I had a development manager of mine come up and said, "Hey, it's a very low-cost tool. You really should try it." I said, "Well, geez, for the cost, go for it." What I didn't expect was it to grow and grow and grow across, beyond just that one person's team.
I had never seen or heard of Atlassian software until I joined this company. My reaction was really just pure joy and excitement. Ease of use of Confluence and of Jira and what you could get out of that software was just amazing to me. I felt younger just walking in here, working with the team, how agile the software was and flexible. It really energized me.
Many different enterprise tools are complex, expensive, and convoluted, hard to use. What we found with the Atlassian tool set is you really just control access. We can have different groups talking to each other and just run with it. They create their own bases. They create their own content. They don't need technology to be involved in this. Having stronger teams and being more collaborative with different groups really help us drive new product ideas much quicker to market than we could if we were very siloed like we were in the past.
Cool. From me, thanks so much for letting me hold your attention for an hour. I am pleased to introduce our head of R&D and CTO, Sri.
Thanks, Jay. Hi, everyone. I'm excited to be here. I'm Sri Viswanath, CTO and head of R&D. I came to Atlassian 20 months ago. Before Atlassian, I had worked in various different companies, Sun Microsystems, large enterprise company, VMware, another large enterprise, Ning, which was a small consumer company, Groupon, consumer, e-commerce. Atlassian is a mix of all this. It's super unique. In fact, for all the reasons that Jay mentioned, I was super excited. Having actually seen the flywheel work and having felt younger based on the video, I am 20 times more excited. Today, I'll share why it is. I'm going to cover first what we have been doing for the last 12 months, how we work as an R&D organization from the R&D lens, and just give you a brief overview of what's coming up next. We invest more in R&D than our peers.
37% as a percentage of revenue on R&D. What this means is Atlassian is an R&D first company. It also means that because we can invest more, we can actually get towards our mission much faster. Our mission is to unleash potential of every team, and my job here at Atlassian is to unleash potential of our R&D team. I'm super excited to be here. Let me first cover all the different responsibilities of R&D and the impact that R&D has at Atlassian, which is different or more than other companies that I've been. We don't just build products. We build products that sell itself. We go two steps beyond just building a product. What this means is we need to understand our customers really, really well.
The customers need to love our product the first time they use it and every time they use it, so that when they walk over to the cooler, water cooler, they need to tell their friends and family that, "Wow, I use Atlassian products. You should use it, too." This needs to work in every type of teams, whether it's software team, IT team, marketing, legal, or a combination of people from all these different functions. It needs to work in variety of customer sizes, whether it's a two-person company in Nairobi or a 10,000 person in Atlanta. We have users who started today who are super naive, who are trying to figure out how to log in, to people who have used our products for 10 plus years and admins who are super admins who know more than many employees.
Our products are customized in such a way that the products look super different when in these contexts. The impact of R&D goes beyond just building the products. We offer different deployment models. We have cloud server data center, and we need to make each of them work and give choice to our customers. Because we are a multi-product company, our platform that supports all this needs to scale faster than all of them combined. We have a thriving marketplace, and when we build our products, we also keep building our marketplace. For all the things that Jay mentioned, we need to have our tools for all the go-to-market and all the automation that goes with it so that we don't have to have salespeople. Our customers are in 180-plus countries, so our products need to work in every single of those countries. There's a lot there.
As you can imagine, we've been busy. We've been busy both in adding value for today, shipping features and shipping new releases, and also setting our company for the next 10 years. I can't cover all the things that have happened in the last 12 months, but let me cover the top ones. Before I cover the last 12 months, let's look back before that, three years before that. We have shipped a number of products consistently over that period. Our R&D organization has a strong reputation of shipping, of making things happen. Even in the last 12 months, we have released a number of products and offerings. If you were in the keynote and you've heard the news last time, Jay mentioned this, we released Stride, which is a brand-new communication tool. It's not just a brand-new communication product.
We built it ground up in R&D on a completely brand-new platform. I'll cover that slightly later. We call it the platform Teamwork Platform. We also released Identity Manager, which is adding authentication and security capabilities for our products. SAML, which is a standard way of authentication inside a company, two-factor auth, which we can enforce across the board, password policies that ITs may have. It streamlines IT policies for a company. Atlassian Stack, which is a bundling of server and data center products, which helps the larger customers, server customers, to use all our products. We also acquired two companies in the last 12 months, Statuspage and Trello. Statuspage is the status for all the services that our customers have. As Jay mentioned, it's like the status bar. Trello, insanely popular project management tool.
From the R&D lens, we acquired these companies. We are integrating them in a smart, strategic way. Our platform has microservices. It's how we build our services. We pick the ones that matter the most and only integrate those strategically so that these products when we acquire, they keep growing. We let them grow faster, make them grow faster. Not only did we work on those and a number of features that I can't cover, we also worked on some transformational projects that set ourselves for the next 10 years. The biggest of them all is Project Vertigo. That's an internal code name. It's the first time we are actually talking about this week. People haven't heard it outside. I'll cover this in detail. Before I do, let me cover a few other ones. Suspense for that. Atlassian Design Guidelines.
It's an end-to-end design guidelines from Atlassian. It's the version 3 of this. It makes our product experiences simple and beautiful. You saw some of it in the keynote today. We've been working on it for the last 12 months, we have actually shipped it in every product. It includes iconography, typography, writing style, whole bunch of things on the product, it's not just the product, it's also all the artifacts that touch our customers. What it does is for naive users and for users in non-software and non-IT teams, it makes it a lot easier and approachable for our products, hopefully, it'll increase our satisfaction score. In fact, we A/B test all our products, we know for the releases that we have done when we release products, what the customer satisfaction is.
ADG 3, as you saw in the keynote, we are getting good results on that. Common identity. We have worked on a back-end platform that gives a common Atlassian account for all our products. What it means is, for all the cross-sell that Jay mentioned, if you have an account in Jira, you don't have to create a new account. You already have the same account that works in Confluence. All the cross-product marketing activities work seamlessly because of this common identity. We didn't have that before last year. That should remove a big friction that we had in our products. Mobile apps. Before last year, we only had HipChat, in the last year, we have added Jira app, Confluence app, Stride. It's a brand-new product. We have also acquired Trello.
We have built out those apps using a very platform approach in terms of we built a mobile platform underneath that can help support more apps in the future, that has all the common capabilities that's foundation for these mobile apps. Vertigo. As I said, Vertigo is an internal code name. It revolutionized how we do our cloud platform. Cloud is super important for us. It's critical for us. 75% of net new customers are coming to the cloud. 70% of all new purchasers happen on the cloud. Vertigo sets us up so that we can scale our cloud platform for the next 10 years. Before Vertigo, Jira and Confluence platform were single-tenanted, had a bunch of issues. We couldn't scale as much. If you look at our website, we have this 2,000 user limit for Jira and Confluence because of the older architecture.
We had to run a VM per and a database per customer. As you can imagine, we have tens of thousands of customers, just deploying and undeploying takes a long time, can cause downtimes and issues. Essentially, as a summary, our innovation was slower than we needed. Vertigo solves all that. We moved, as part of Vertigo, we did 2 big things, each of them on its own is really hard. Having been through multiple of these kinds of migrations, these kinds of projects are extremely hard to get it right. We not only changed our platform to be multi-tenanted, which is how all the modern software is built, SaaS software, we also moved it to a third-party service, AWS. Let me tell you what it buys us, what it gives us, what the advantages are.
Number one thing, the reason why we did this is faster dev speed, faster innovation. We can deploy faster for a couple of reasons. One, of course, we can spin up the VMs faster, all the things that you hear. Also, fundamentally, the way we develop software changes internally, right? We don't have to worry about figuring out which database to stand up and get some machines in place and configure all those. We can use what's available on AWS. That fundamentally changes the rate of innovation. Undeploy, for example, rolling back if there's an issue, is in minutes. It used to be days before. We can also spin up newer presence in other regions where AWS is available. That gives a faster experience for our customers.
The new architecture allows us to break through the 2,000 user limit and support larger customers as they grow on the cloud. Lastly, because we did two things where we changed our architecture and moved to AWS, it's also lower cost to serve. We don't have to worry about the plumbing. We can let AWS and third-party services optimize for it. We can focus on adding value to our customers and solving the teamwork problem. It's a massive project. I'm super excited to tell you that we are 94% done. We started migrations earlier this year, and we have migrated 94% of our cloud customers. This is a big deal. Let me just take an anecdote, right? It's like changing both the engines mid-air in your plane, right, with no impact.
We have been seamlessly migrating customers to the back end, so customers, they're getting their Coke and they're drinking Coke happily enjoying. They have no knowledge that the engine's changed. We did one more thing. During this process, we also spun up an Ireland data center, and it's proving to be faster, and we are migrating Europe customers into Ireland, and it's proving to be faster. This plane, we also built a new plane midair and moved those customers over, and they don't have to have hops, and they reach faster. Huge. Huge for us. It's a big testament to our R&D team. Big kudos to everyone involved. It's a large project, but I'm super thrilled to say that we have the, like, we'll finish by end of the year, and it's set up for the next set of growth that we have. All right.
Let me spend a few minutes on how R&D works. Right. Touch upon customers, teamwork, innovation, and ecosystem. The combination of this is different from every place that I have been, and it's unique. The first thing is about customers. Customers come first. We are maniacally focused on customers. As I mentioned earlier, we need to build products that sell itself, which means we really truly have to understand what our customer needs. For that, we employ various different mechanisms. We have strong feedback loops. We have customer satisfaction surveys. We do A/B tests. The other thing that we do internally in the company is dogfooding. We eat our own dog food. We try all our products across the board. Our products lend itself for teamwork, and we are spread over different geos, which I'll come to. The dogfooding helps us make our products really better.
Next is people. Our R&D is set up in all these different locations. We also have some remote people. Sydney is the largest of all. We've grown as a global team. The advantage that we have is that we can tap into the local talent pool, and we can hire the best in each of those locations. Because we are a teamwork company, we deploy all our tools and make them all work and execute on all the complicated projects that I said as a single global team. Next is innovation. We have a number of frameworks in place, systems in place. Overall, starting from the founders, Mike and Scott, who are techies, who wrote the first version of Jira, it's been focused on innovation. Innovation is the lifeblood of the company. ShipIt is one way of expressing that innovation.
If you walk the hallways, if you saw the keynote, we have a ShipIt live happening here. It's amazing to see the energy that comes in our ShipIt that we do every quarter. In fact, Jira Service Desk product came out of ShipIt, and it's growing and thriving today. There are a number of other features who have come up, too. In fact, Jay showed you the graph or the screenshot of inline commenting. Inline commenting came out of multiple ShipIts because commenting inline in a text is a harder engineering problem, because you need to know where to start and end text change. Somebody tried this was a few years ago, tried it in a ShipIt, didn't make it work, tried it in a ShipIt again. It wasn't on the roadmap because that was not high up on there.
Ecosystem. We not only build our products, we also expose APIs for those. When the product starts growing, the ecosystem around that starts growing, too. The use cases that we would never solve, we would never build, would never be on our roadmap, is solved by all these different partners that we have and developers that we have. We have more than 3,000 apps in our ecosystem. From the R&D side, we shipped JSD that came out of ShipIt. A year later, we also exposed the APIs. Now there is an ecosystem building. We launched Stride last week. Today we announced the API for Stride.
That's part of our DNA, to allow our ecosystem to grow along with the products, which helps in various different dimensions of growth. All right. Customers, teamwork, innovation, and ecosystem. The combination of this makes it unique, and the flywheel going on innovation. Let me spend some minutes on long-term initiatives. These are the initiatives that you can expect us to be investing for the next few years. Not specific to the next few months, but overall, these are the tracks that we would invest in. I'll cover these four areas: trust, feature, simply powerful, smarts, and mobile. Starting with trust. It's critical for us to build trust with our customers, and we have done an amazing job on the server and data center over the number of years. Cloud is newer in our portfolio, and cloud in the last four years is growing.
We want to double down on this. Now that the Vertigo platform is done, or almost done, we want to double down on making sure that we keep working on security, reliability, privacy, compliance that's needed, especially as we break the 2,000 limit and get the larger customers. In fact, the Identity Manager is the first release of having that layer, security layer on top of our cloud products, and you can expect us to have many more of these. Next is simply powerful. Our products are really powerful. Jira and Confluence and other products have lots of features. It works well if you're an advanced user. It's much harder if you're a newer user or if you're in a non-tech team, haven't used the products in the software context. We want the products to be powerful. We also want it to be simple.
The combination of simple and powerful is a unique combination that we want to make sure that we keep investing in. The onboarding of our new users, making sure that as you learn more about our products, the progressive disclosure of features, and as you become more sophisticated to enable more features, is one of the tracks that we'll keep working on on all our products. Let me give you one example of how doing this will not just for the software context, but the non-software context works really well Jira. This is the Jira details page. There are a number of different units here, and not all of them is required in every single place. What we have done in designing this track is to make sure that we can modularize Jira and overall as a platform in terms of different components that can exist.
We can take the different components in the page and figure out what's needed for the action that's happening. In this example, it's a board. Somebody's trying to figure out how to change status of the board, maybe add comments, and focus on the action so they can get more work done. This kind of modularization helps us to build simpler pages for specific contexts in other products. Smarts. We hear machine learning and AI everywhere these days, right? It's a buzzword. If you step back, it consists of three different layers of what you need to do. There's the machine learning platform, then there's the algorithms, and then there's data. Every third party provider has a platform these days, and it's getting much better. I consider that as a commodity. The next layer is the algorithms. Algorithms are becoming a lot more available.
There's a lot more research happening. Over the next few years, it'll become much more of a known thing than an unknown thing. What differentiates machine learning is the data layer. Really, if you want to make machine learning work as a company, as a developer, as a product person, it's the data that matters. That's important because for us, as Atlassian, we have that unique data from our customers. Think about this. If you look at the life cycle of the data that we have, we have project management documents of content in Confluence. We have Jira bugs and project management tracking in Jira Software and Jira Core. We have content in Trello. We have source code in Bitbucket, right? We can actually put this together and figure out the things that you can do to improve productivity of the company.
If you start from the higher up as a company overall, you can have statistics on how things are doing in different parts of the company. We can also go all the way down to the platform layer where if you take a single element, we talked about this in the morning keynote, we have built out a brand new platform called the Teamwork Platform. If you take the single element in there, we can make that smarter using the data. We can truly understand you as a person and all the people around you and the relationships and what you use. If you type in @mention, we can figure out smartly what people we need to show and not just alphabetical. Starting from the lowest platform level all the way to the company product and company level, we can make it all smarter.
We are taking a very practical approach at that. We have started building the teamwork platform. You'll see us including it in all our products over the next several years. I guess our products would be smarter inside out and personalized. Lastly, mobile. I mentioned earlier, mobile, we have made significant progress. We had HipChat a year ago. We built Jira, Confluence, Stride, we acquired Trello. You can expect us to invest more in making these better and also solving for use cases that couldn't be solved or can't be solved on the desktop when people are on the go. Combine this with smarts. I think we can make smarter notifications and much better personalized experience on home. There's lots of innovation to be done here. All these things combined gives us better engagement, more active users, and hopefully more dollars. All right. Let me summarize.
R&D powers Atlassian. R&D innovation is the lifeblood of Atlassian. We have the foundation strong on innovation and frameworks in place. Vertigo cloud platform delivers high performance and scalability. We are 94% done, I'm excited that we'll cross the finish line this year. We are maniacally focused on customers to build products that sell itself. We are delivering value today and also setting ourselves up for the future. That's all I had, guys. Thank you.
Okay. Thanks, Sri. Good afternoon. I'm the last presentation before we get to the break, I'll do my best to get through this. I'm going to cover the financials. I'm going to cover a bunch of different items that have come up as questions by many of the investors. Hopefully we're going to be able to answer many of those today. One thing I want to start with is just, again, is our model and how it really works. Now, Jay went through a lot of this, and so did Sri. From sort of a financial perspective, the number 1 thing is you have to have a massive TAM. For the way we price and the way we do things, you have to have a massive TAM. This has obviously been an area that many of you have focused on.
Are you really going to get outside of software and IT? Can you get into large enterprises? We see this as a TAM that's very addressable. We're going after this Fortune 500,000, and we're going after all teams and all organizations. We have a value-driven approach to pricing. Essentially, we have very low pricing. It's highly competitive, and in effect, we're kind of cannibalizing the market. The market is so large, there's no end in sight. When you couple that with the kind of products that we deliver that can sell themselves, products that customers love using, the combination of those two puts us in a highly differentiated position as we tack this large TAM.
One of the things is that I was on the board for four years before I took this role, I didn't fully comprehend the fourth one here, which is You have to defer greed and be patient for revenue in this model for the way for it to work. If we were to go out and hire salespeople and aggressively try to drive a little more revenue growth, what would have happened is we would have increased our prices, which would have made our prices less affordable to the Fortune 500,000, and we'd eventually start narrowing down the number of customers that we could target. We would really need these salespeople to justify the price. We'd have to discount. We'd become less predictable. We'd generate less cash and profits. With this model, you have to be patient.
I remember talking with Jay maybe about 1 year ago, I said, "Jay, I finally get it. You have to defer greed." He says, "Exactly, that's what it is. You just have to be patient." I have to kind of be careful myself when I go over and talk to some of the enterprise advocates, "Hey, I hear there's a large deal coming." It's not to push too hard, just, "How's it going?" Not push too hard. If I do, everybody will start doing things that are unnatural, which means discounting and negotiating things on agreements, and we'll turn into a hockey stick company, and we don't want that. That's just how this model works, and it's very difficult for anyone else to really copy this because you would have to start your company out this way.
You can't have $100 million in revenue and go to this model. It's too late. You would have to take your revenue down because of pricing. You'd have to get rid of your sales team and go become a very patient company for revenue. No investor is going to accept that at that point in your development as a company. In terms of the TAM, we showed this in our IPO. We haven't updated it. It's a large TAM. Software's eating the world. At the time, there was 21 million software developers, and growing. That's obviously a very important place for us. 100 million users in sort of technical teams. Then there's the knowledge worker number, which can be 800 million, 1 billion, 1 billion plus, depending upon where you go to look at that. Our products serve the needs of all of them. They all work on teams.
They all have workflow. They got projects they're involved in. They're creating and sharing content. They're all addressable. We have a long way to go. At the time of the IPO, we had approximately 5 million active users now. A long way to go here. No end in sight. In terms of revenue, you can see over the last 6 years what our revenue has been. I joined the board right after it sort of hit the $100 million marker. I was kind of thinking, "Hey, you need to get salespeople to get to $200 million." In the 6 years I've been with the company, I've heard repeatedly across all the way up here to this $800 million plus is, you're going to have to hire salespeople. You can't expand into the enterprise. When you get to $300 million, you're going to have to hire salespeople.
$500 million. We'll probably get some questions about that again today. Here we are crossing over $800 million on our way to $1 billion with this model. This is my favorite slide here. What you're seeing is the last 8 quarters sales by day. That's how linear our business is. That's sales, that's transactions. Those things that Jay talked about on August 8th, that's the 90 days for the last 8 quarters. It's my favorite. I've worked, I guess, my whole life and thought I'd work for a company that had this. It's kind of like CFO Nirvana in a way, because it's going to be what it's going to be. It comes in every day. Matter of fact, I just want to look at that one more time so you can see that. You might notice there's a little bit of ripples in there.
Those are the weekends. We generate revenue on the weekends, but not at the same level as Monday through Friday. We're global. Sunday here is Monday in Australia. That's how linear our business is. It really helps us in quarter to understand where we're at financially. It helps us if we want to adjust any dials on our expenses in terms of what kind of an outcome we might want to have. It puts us in a position where we're really in a no surprise environment from a top-line perspective. It also is happening when you look at it on an annual basis. This is showing our recurring revenue. This is our maintenance on our server side and then our subscription business, which is cloud and data center. You can see it's increasing over this time period.
As we go into a year, we got a pretty good idea where our revenue is going to be. If I went back and told you in 2015 and 2016 how much we missed our plan by, it'd be shocking. It's so close. We really have a good idea in this model. Now, again, if we were to get greedy and try to squeeze more revenue growth by hiring salespeople, all this will start to fade out. It's a tremendous model and one we want to stay with. This is just showing our server and maintenance business, which is in blue. This is the perpetual license business. It still represents more than half of our overall business. You can see in white, the subscription business, cloud and data center, is growing at a faster clip.
We're not trying to move customers to the white area from the blue. Other companies are going through this cloud transition, and they're really trying to shift customers over. We're not doing that. We're going to let customers choose what they want. We believe that blue area, that perpetual license plus maintenance business, is going to be around for a long time. We'll let customers decide when they want to move over to cloud and to data center. We'll do things clearly in the data center area to encourage them, because we think it's a better solution for large enterprises. This is something that we'll let them make that decision. It also has some implications for gross margin, and I'll cover that later on in my comments. Here again, you can see that we've been profitable for almost since inception. We've generated cash.
This is just the last five years in terms of our non-IFRS operating margins and our free cash flow margin. We had a little bit lower free cash flow margin in 2015, 2016, really driven by CapEx. A lot of that was in those data centers that Sri is talking about that we're moving out of. Our CapEx has really been in two areas. It's been leasehold improvements in facilities, and it's been in building out our data centers. We're moving out of the data center business. We're in the apps business. We'll let somebody else run that. From a capital expenditure standpoint, we are not going to be having the investments in our own servers for our own data centers going forward. You've seen both of these. You can see the outcome of the model.
One is R&D is a massively competitive advantage for us, to have that kind of resources to invest in making our products highly competitive, to bring new products to market, to serve all the deployment models, to have that ecosystem, all those things that Sri talked about. We get a lot of questions from investors about, "How can you guys spend 37%? No one else spends it." We tried to help you today to understand that we do have different areas that we need to invest in that other companies don't, and we need our products to sell themselves to make our model work. In sales and marketing, this highly efficient model has led to this mid-teen sales and marketing as expense as a percentage of revenue. That's where our profits and cash flow come from.
Because that number's lower, that gives us those profits that we have. It also does something else, is that we don't have the complexity of managing that field organization around the world and country managers, and something's happened in Europe, and we got to replace the leader in Europe and all those challenges that come with managing a large field organization. We don't have that in our business. Okay, let's go into pricing. This is an area we get a lot of focus on. I'm going to try to spend some time here to explain our pricing models, give you some ideas of how things compare depending on the deployment model that you choose, and talk about some of our recent actions. We've got these three deployment models. The top is you host. That's a server you've bought from us. You run it behind your firewall.
That's a perpetual license plus maintenance model. There's the data center, which you also run behind your firewall, but it's an annual term license. We recognize the revenue over those 12 months. This is the area we're going to have some changes in our revenue recognition around IFRS 15 or ASC 606 under GAAP, and I'll cover that here in a few minutes. Then cloud is your sort of standard SaaS offering. On our April earnings call, we had hinted that there's going to be this thing that's going to drive us this low single-digit points of revenue growth, and then we talked a bit more about it in July. Now we're actually up and running with this with our customers since the pricing changed on July 31st.
One of the things that we've heard from our customers is they didn't want to buy our cloud offering like it was a perpetual license with these tiers. They wanted to buy on a per-user basis. If they had 25 users and they wanted to go to the 26th user, they didn't want to buy the 50 user tier. They just wanted to buy that one extra user, which is pretty standard in companies sort of in our cloud industry, this deployment model. We did a lot of work around actually how to do this. We've come out with per-user pricing on a monthly basis, just the way they wanted it. Secondly, we heard is they don't want per-user pricing on an annual basis. They wanted these granular tiers. It's kind of PUP-esque, but it does still have some tiers.
As we did all of this, we looked and said, "Are we really capturing the value from the pricing we're charged for the value we're giving?" We made some decisions to increase prices as part of this, and this is what led to the low single-digit points of revenue growth. On the annual side, you get a discount. You got one before, you get one now. There's no real change there. That's one of the things that we would expect to see some customers with the price change say, "I'm going to go from monthly and I'm going to get that discount on annual." You had to factor that in to get to low single-digit points of revenue growth. We had some customers that might have only had 12 users, and they'd be paying $84 for those 12 users in Jira now.
If they went down and became a starter, they'd pay $10. We knew there'd be some of those kinds of things going on. There might be some that would just attrit out. All that went into this low single-digit points of revenue growth. The experience we've had to date at this point is kind of what we expected. There has not been any kind of major customer reaction to this. It's been relatively quiet. It's still early days. Actually, how this all plays out will take a number of quarters. Up to this point, the reaction has been as expected. A 140-page report went into this. This was not a let's figure out how to drive revenue growth on a Friday and then on Monday announce new pricing. We worked on this very carefully because we don't want to screw up that value-driven pricing.
We want to make sure we still capture that and maintain the great value that we have. Here's a way of looking at how does the pricing work across the different deployment models or the ways you could buy from us. This is 199 users. It's kind of a random-ish number, but these numbers will move around depending upon the number of users. Jira Software, if you went cloud monthly under the new pricing, you'd pay $13,000. For the year, for that 199 users. If you went cloud annual, you get your discount, you get two months off, $11,000. Server, $13,200. That would be your license plus maintenance. If you spent $2 with us, $1 would go to license and $1 to maintenance. In that second year, we would get the maintenance again if you chose to renew. $13,200 for the server.
Data Center, $12,000. Really right in that same zone. We don't want to make it create any friction for you to go to Data Center. You'll see in a minute here what happens as you increase the number of users on Data Center and how we generate more share of wallet from the customer. At 500 users, this is the cumulative revenue we would generate if you were on cloud annual versus if you just bought the server plus maintenance. In year one, cloud annual will be lower because you're going to buy the license plus the maintenance in year one. In year two under server, you just got the maintenance. Now we're crossing over. We're generating more revenue on that cloud annual for Jira Software at 500 users, then it continues to expand from there.
This would change again for different products, but this is a good example of just trying to show the trade-offs from a customer perspective. The gross margins are different for us in these, I'll talk about that a little bit later. You'll see here, by year 3, the revenue now is starting to really expand out on the cloud annual side. Today, the majority of our customers in cloud are still on monthly. We would be generating more revenue, the compare would look a little different. This is graphing the difference in the price in your first year on Data Center versus if you went with the traditional server plus maintenance. You can see right there that if you got a few users, you're not paying much for server plus maintenance, getting into the product, getting going, getting that first team started. Data Center is not doing much either. It's pretty low price.
As you get out there now into an increasing number of users, the value of Data Center kicks in. All the things that a larger organization scale really needs in terms of running thousands of users, Data Center meets that need, they're willing to pay for it. When you hear us talk about these deals over $500,000 we're expanding in the enterprise, Data Center is a big driver in that. We are collecting a lot more revenue from them because of the Data Center offering. We do a lot of price changes. They're going on all the time. Some are up, some are down. This is just showing you in the last 12 months, all the different price changes. The most recent one is server on the left.
A couple of weeks ago, we announced we are raising prices on server 10% on new licenses and 10% on maintenance. The license will start on October 1st, maintenance on January 1st. It is already included in the guidance we gave on July 27th. This is not going to add to what we said. This was already embedded in the guidance that we have provided. We did not want to pre-announce it on the earnings call. We knew it was coming up soon. We wanted to make sure the customers heard it first. As you start thinking through all those numbers and how much revenue could this generate, maintenance is where the money really is in terms of the price benefit.
If you are renewing on June 25th, which is about the end of our fiscal year, that is the first time you would have that price increase, and it would go to the balance sheet, and we would get five days of revenue out of 365. We really would not get any revenue benefit from that. The maintenance price increase benefit is really more of a fiscal 2019 than it is a fiscal 2018 benefit for us. There are other changes we are making in the maintenance area as well that actually is going to reduce revenue and make some areas a little more customer friendly. Net-net, it will give us some revenue, but it is already factored into the guidance that we provided on July 27th. In cloud, those are all the different types of price changes that we have made over the last 12 months.
In data center, Atlassian Stack is a discount by bundling a bunch of data center products together to meet the needs for the larger enterprises that want to go in that direction. A few years ago, we had outsized benefit from price increases. We went through that during the IPO process about getting 15 points or 10 points of revenue growth because of the price changes. That was more of a one-time deal. A product like Jira topped out at 500 users, if you had 10,000 in your organization, 9,500 of those users were free. We added tiers up to 10,000 so we could collect some value that we were bringing to them. There was volume discounts. We got a big chunk of revenue when we did that. The startup pricing we are talking about here is more your standard price increase.
It is not a one-time big benefit. Do not look at what we did a few years ago and say, "Wow, they are doing something on server again. Are they going to get 15 points of revenue growth?" That is not the case. Pricing. We are going to continue to be the low price, high value leader. We are not changing that. We will continue to look for ways to optimize pricing and put that rigorous effort into it with focus groups and analytics and all those things to strike that right balance of generating some additional revenue, a little more for the shareholder, giving us some resources to invest in our future, but making sure that we do not lose the position we have competitively as we go after that Fortune 500,000 customers. The other thing is we are going to move to annual increases.
We've gotten this habit over the years of not doing anything, kind of like in cloud for six years, and then we do this price increase. We're moving to annual price increases. We've got salary increases for our employees. Every one of our software vendors has an annual increase. We got rent increases. We need to move to this more standard model that's less disruptive to the customer. We're going to move to annual increases across our business. The amount of those will depend year to year. We'll see where we're at on that. Obviously, if there's heavy recession, that might have an effect on what we do or not do. We're moving to annual increases. A few other topics. Share-based compensation, we get this question a lot. "Wow, you guys have so much share-based compensation expense.
I'm comparing to other companies." Well, first of all, we're under IFRS versus GAAP. Under IFRS, they've determined that you've got to expense it differently than GAAP. This is four-year RSU vest. Under GAAP, you'd recognize the expense evenly across those 16 quarters. Under IFRS, it was decided that it's a better way to do it, is to have 57% of the expenses over those four years loaded in year one. When you're hiring a lot of employees like us, we're just front-loading the share-based expense. Now, if we ever were to slow down our hiring and level off, then it's going to come back and be a real positive. At this point, we're very focused on hiring lots of employees and driving for the long term. We're going to see this front-loading of share-based payment expense.
The point here is, I wouldn't focus so much here to kind of figure out what's going on with share count dilution. We're going to give you that here shortly. This, I think, is a little bit of a head fake based on the way that expensing is done under IFRS. We're IFRS 15. We're not 606. We're going to adopt it on July 1st, 2018, which is the beginning of our fiscal 2019. We're going to go with a full retrospective method. We looked at modified. The decision was full because the internal cost to do these was about the same. We're not going to have much of an impact to our revenue. We went ahead with full retrospective, so it'll be very easy for you guys to compare from a trending standpoint.
The only area we're really going to see any kind of impact is going to be that data center license. That where you have the 12-month license, that term license, and we recognize it ratably over that 12-month period, we're going to have to front-load some of that. It's going to give us some additional revenue in fiscal 2018 and 2019 as our data center business grows. It's not going to be significant. Because we don't have that quota-carrying, bag-carrying Sales force, we really don't have much in the way of commissions, and they're so low right now, we don't think it's material enough to even capitalize and expense them.
Right now, obviously, there's more to do here, our position right now is we're not going to capitalize any commissions, and it's just going to expense as incurred, which is what we've done in the past. Customer count. As a definition, the way we look at this is it's a unique domain with at least one active and paid license and approximately $10 per month that we receive. That's this sort of 90,000 number. We had 89,000 at the end of July that we talk about. That's the definition that was put in place a while back. We tweaked it a little bit for Trello because Trello is $9.99, we wanted to give you visibility in terms of how many customers we acquired. We added them into that, we shared with you when we acquired them. There's nuances here.
This is not a perfect formula to do this because we've taken this conservative approach where you've got to pay us this $10. A ways back, we had server starters. We've talked about they're still coming in every day. They're trying the product. They're kicking the tires. They're doing what they're doing, we've never counted them. They come in. They go. They're one-time $10 server starters. We adopted the same kind of framework on cloud. If you're paying us $10 a month in cloud, we're calling you a starter, we're not counting you as one of our customers. Yet, we've got Trello, which is pretty much at $10 a month, we're counting that as a customer. We're probably going to evolve our customer definition going forward.
We didn't really want to come out and just give you this huge number change in a quarter where you're thinking, like, "Are we trying to obfuscate something?" My expectation will be is that probably in fiscal 2019, we're going to evolve this. Just to give you an idea, there is 97,000 cloud starters out there. When you think of that Fortune 500,000, we've made a lot more progress on going after that Fortune 500,000, that is all expand possibility over time. Some will never move off that number. Others could be a Fortune 500 company that's experimenting today. There's huge numbers of starter customers. We add that in, we're going to be a much bigger company in terms of the number of customers that we've reported. Bitbucket, just to give you another little nuance here.
Bitbucket was free up to five users, you paid $1 6-10. We didn't count you as a customer. We raised prices a few months ago, went from $1 to $2. If you were a six-person user, you weren't a customer because it was $6. Today, it's $12. You're now a customer. In Q1, our customer number is going to go up somewhere between 12,000-15,000 for all the Bitbucket customers that were paying us before, we didn't count as a customer, they're going to become a customer going forward. In addition to that, we'll have our normal amount in the quarter that we generate. We'll give all this to you on the October call, we just want to give you a heads-up it's coming, it's because of the price increase on Bitbucket. Long-term financial considerations.
At the time of the IPO, we had this sort of three, five-year framework with some margin data, et cetera. We're sort of retiring that, we're going to give you some more quantitative considerations as you think about the model going forward. Number one is, assume annual price increases. Whatever our revenue growth would have been, we're going to get some annual price increases, so that should lift our growth, assuming that we do them in each of the years going forward, which is our intent. Again, if there's some kind of big recession, we might reconsider. It's going to be across cloud and server. Two. We've been accelerating depreciation on those servers that we're running in our own data center that Sri and team are working really hard to get out of and move to AWS.
That's going to wind up by the end of this year. As we go into fiscal 2019, we're going to get a little relief from that, but at the same time, we're going to be transitioning more to cloud. That transition to cloud is going to continue, the cloud gross margins are definitely lower than they are on the server side. On the cloud side, we're hosting all those costs, whereas if you bought a server, you would be hosting your own cost behind your own firewall. It's not going to be any kind of big dislocation. It'll just be a slow trend down. We'll do everything we can to try to optimize everything we do in terms of our cloud operations. That mix where customers choose to move from server to cloud will have a modest decline impact to our gross margin over the time.
Think of this as sort of a three to five-year timeframe in terms of what I'm going through. In terms of op margin, there's going to be modest leverage. It's going to be driven by revenue growth. We would expect that our op margin is going to expand over time. Exactly what that is will be dependent upon how much revenue growth we have. We're going to let revenue growth be a big driver of our operating margin expansion. We're not going to save our way to greater glory. The one area that we do have a little bit of noise is that about a quarter of our costs are Aussie dollar denominated, it's been a bit of a volatile currency over the past years.
We hedge a partial amount of that, it goes out over a number of quarters, it's kind of a rolling thing. It's a smoothing function. The Aussie dollar is around AUD 0.80 today. It's been as high as AUD 1.00 a few years ago. It's been as low as AUD 0.70 a year ago. While we'll be hedging that at some level, you're also going to have to keep in mind that if it were to go to AUD 0.90 or at a parity in a six-month period, it's going to be hard for us over time, even in that six-month period, because we don't hedge at all, or if it happened over 18 months, it's just going to smooth it out, that'll have an effect on us.
We'll obviously look for ways to try to offset that, but it'll be something that you'll have to keep in mind in terms of our op margin. We sell in dollars, so we don't have any natural hedge. That may change. There may be at some point, we'd move to selling in local currency as sort of an offset, but the Australian market is not going to be one of our largest revenue markets. There won't be the degree of revenue there that would be a natural hedge against the cost that we have, and Sydney is our largest location. Free cash flow is going to kind of move along with op margin. Looking forward, we would expect to see our Free cash flow margin modestly increase kind of in parallel with op margin. There's two things that could affect that.
Number 1 is that if more of our customers chose to go from monthly cloud to annual cloud, it's going to basically pull forward bookings and Free cash flow. Again, we're not trying to change what they currently do today, but that would be something that could increase our Free cash flow margin for a period of time. It assumes constant facilities investment. This is really where the CapEx is. There's going to be years where we're going to have some leasehold improvements, we're going to fit out facilities, and then we're going to grow into them for a few years. Then we're going to fit out again. When you looked at our Free cash flow margin over the last five years, the lower years were because of CapEx. That's going to have some impact on kind of where we go from here.
There's going to be some years are going to be more and some years are going to be less. The overall trend we would expect is to parallel that op margin and to modestly increase. Share count. Instead of trying to look at share-based expense and trying to figure out what is the dilution, we're targeting 3%. That's an approximate number. There's a lot of things that could go into that could move it around. Number 1 is that when you have a lot of new hires, we're a growth company, that's going to lead to a lot of new hire grants. It tends to be the largest grant employee gets is when they come in the door. If we have acquisitions, that obviously could change this.
This 3% excludes acquisitions, that could have an impact. If our stock was $100 and we're making offers to employees based on value, is what we do. We don't do it on a number of shares, it's on value. We'd grant a lot less shares. If we were $10 a share, we'd have to grant more shares to compete if that's what it took. Our stock price is actually going to have an impact on what share count is. If we perform really well and the stock goes up and all these things, and we're competing on value, there's a way here that we actually could have less dilution. For right now, for modeling purposes, we'd say targeting this sort of approximately 3%, but it can move around back and forth here.
The one thing I will say that having the two largest shareholders at the table on these decisions, they are highly aligned with all of you. I'd almost argue at times maybe too aligned in a sense, in terms of what it takes to compete in Silicon Valley for some of the best talent. There is a lot of sensitivity to this, probably more so than you'd see in some other companies where you didn't have quite the ownership percentage with the founders. Lastly, on taxes, we're going for cash tax. Lots of things that we can do as an organization to try to optimize here. We'll optimize for cash tax first and book tax second. We'll do everything we can to improve earnings per share on that effective tax rate.
If it ever comes at the expense of the cash tax, we're going to do cash tax first. That's the focus there. As we announced on our July call, I'll be transitioning out at the end of the year. I just want to touch briefly on some of the things that have been put in place since we've been public. These are all working today. They've been adopted by the organization. We've got a strategic planning process in place. We look at the three-year model around that, the things that we've got to get done. We have an operating plan, which is the operational actions or the OKRs that the top 25 managers in the company have to get done in the year to achieve that strategic plan.
They got to stand up, and they got to do it with all their peers and say, "This is what I'm going to get done this year." We have a financial plan, which is just the numerical representation of all those actions. It's the basis for our annual guidance. We have a quarterly operating review. Now, we bring those 25 managers back in, and they got to go right back and said, "Here's what I said we were going to do during the course of the year, and here's our status, how we're performing." There's accountability that's there with your peers and the executive team. We take this very seriously. This is up and running. The quarterly rolling forecast, we go out four quarters. We're finishing it up right now for the next four quarters.
It's the basis for our guidance in the next quarter for the remainder of the year. We use it as a way to optimize resources. There may be some areas where we don't think we need quite the funding we thought at the beginning of the year. We're going to shift it somewhere else, or do we take it to margin or whatever we do. This is in place. The executive team signs off on all the commitments across those four quarters. We have a weekly outlook.
Beginning on week 5, every week, those same 25 managers have to say, "Here's where we're going to end on expenses for the quarter." On the revenue side, because we're very data-driven, the finance team does it working closely with Jay's team, is that we have every week where we think we're going to end in revenue and bookings for the quarter. We're running the business by the numbers. This is reviewed by the executive team. You're held accountable. We look at where you were on week 7 on expenses, and we look at where we close and did you know your numbers. It's really important that the managers of the organizations, as we scale, understand where they stand on their numbers. This is all in place.
The last few public companies that I left, where this was in place, the incoming CFO tweaked a few things. It was adopted. I would expect it to continue. It's working well for us today. Just in summary, what you've heard today that Ian mentioned at the beginning is that you heard that our go-to-market approach continues to scale. Here we are crossing $800 million. We're on our way to $1 billion. This model is continuing to work. Our expansion paths continuing to work very well. Jay spent a lot of time on that. We're all about R&D. You can see where that 37% goes. It goes across a wide spectrum. It has to begin with having great products. You saw a little bit about Sri and what we've been able to do. This Project Vertigo is a huge undertaking.
I sit next to Sri at work. Following this all along, it's really impressive what they've been able to do in a few months to do this conversion that so many other companies are challenged with and to get to 94% of our customers is really exciting. That's where the good way and the cloudy. The price is right. We've talked about some pricing today. We're going to continue to look at that as something that is part of our model. We've got a great financial model in terms of predictability, recurring revenue, profits, and free cash flow. We don't see anything that we're going to do that's going to go disrupt that. Great model. We're going to continue forward as is. Okay. Thank you on the financials.
What we're going to do now is we're going to take that 10-minute break that was advertised. Scott and Mike, our two founders, are going to join Jay, Sri, Ian, and me. We're going to take your questions for about, I think, 45 minutes or so. We'll see you back here in about 10 minutes. Thank you.
Okay, great, everybody. We have about 40, 45 minutes or so. It's free flow questions, please feel free to put your hand up, ask a question, and we'll go from there. I think first one down here.
Hey, guys. Thanks very much. I must say, I'm not just sucking up, but this is actually one of the best analyst days I've been to in a long time. I found that almost all the questions I was about to ask you answered one slide later, thanks very much. A lot to meet. Couple of questions. First one for Murray on pricing. You guys have said that you intend to keep pricing low, but you intend to keep raising. Now you're going to be raising prices every year. How do you feel like you're balancing that, in terms of continuing to be that value and price leader, at the same time, making sure you've got the whole elasticity curve right, and that you're not going to alienate customers with those price increases?
Well, that's exactly what we're doing. We approach it just with that. We do not want to change our model. If we raise the prices too much, it narrows down our market. We've got to hire salespeople to justify the price and all that. Clearly, we've got to have great products. We've got the firepower and that kind of investment in R&D that we've got to come out with great products and continue to innovate so that the pricing will continue to make a lot of sense from a value perspective. We'll continue to look at what we think it's appropriate in terms of what we think customers will pay. We'll do all that stuff that you would expect us to do. Customers understand that our costs are going up every year and that a price increase would make sense because everybody else is doing it.
All of our software vendors, every single one of them that I've signed in the time I've been in the company, every one of them charges us an increase because their costs are going up. I think it's going to be more in that kind of modest price increase than it's going to be, we're going to go out and do some kind of 30% increase mode. That would just break the model again. We're going to be just thoughtful about it. We've hired some really sharp people. We've used third parties to help us to really be thoughtful about how we approach pricing so that we do not break our model, but we capture the value that we are delivering to our customers.
One follow-up for Mike and Scott and Jay. I'll keep it short. It's a competitive question. Also, I should have said who I am, Michael Turrin from Raymond James. Sorry. In any case, general perception is that there've been a lot of introductions of new collaboration products by large companies recently. Let's maybe focus specifically on Microsoft and how you think about them long term.
I can take that one. If I think about as we grow as a company, the competitors we compete against increasingly become the larger players. I think there's four areas in which we compete with those larger players, and you'll see that through all the things that we do. One is that we operate open. That's something you saw in the keynote yesterday and the keynote today. That's a philosophy that we have, effectively open by default. If you look at our competitors' products, they're closed by default, and it's actually hard to retrofit that. It's actually really difficult to say, now every document you write is now open to your company. It has to be baked into the products from day one. Two is we're about teams, and they're about individuals.
Again, if you look at all the things, the platform that we're making for team productivity, some of the things we showed you this morning and a lot of the things that are coming over the next year are about things that makes team more productive rather than individuals. Third is we interoperate with a large number of vendors. Whereas most of those big companies say, "Well, you're either in my ecosystem or you're not." We think that's not the right way. I think long term, most companies are going to have a heterogeneous environment. Integrating with large numbers of vendors is the way to do that. Lastly, our products have always been flexible and extensible because we know that teamwork is messy. There's not one way of doing teamwork, and every team works in a slightly different way.
If there was one way of doing teamwork that worked, we'd all be doing it and it would be a solved problem. We make sure our products are flexible enough to handle all the different ways that teams work, and that extends to our ecosystem. You saw we had 3,000 apps in our ecosystem. We have hundreds of vendors. You can walk outside and see them all. They're the four areas where I think we'll compete really strongly against those big players going forward.
Sorry, in the back over there.
Hi. Keith Bachman from Bank of Montreal. Two things. First, Michael Turits, I would like to say that you were sucking up when you complimented the slide deck. Secondly,
Where is Keith from?
For you.
You got to apologize for that.
Murray, for you, could you give a little dimensions on your comments surrounding margin increases? What I mean by that is, what are the puts and takes that you think investors should be thinking about as it relates to your ability to expand operating margins? Secondly, is there any dimensions that you can give around that? Many companies give a long-term model where it's going to be 100 basis points, it's going to be 50, it's going to be 200. I think it's a very different implications for the stock in that your margins have been flat, call it three years on an operating basis. If you could give a little bit more granularity on margins. Thank you.
We fully understand that being more specific on margins, everybody's got an idea about revenue growth and looking at margins, another driver for the stock. We understand that. This company has been built on investing in the long term and getting ourselves hung up on we've got to hit a certain target next year and the year after and letting that start to change our own thinking in terms of thinking long term, we don't want to put ourselves in that position. Now, we understand from an investment community standpoint, you want that. We get all that. This company has been founded on the long term. We know that over time, the margin goes up. We said that.
To what degree will be dependent upon if we see new opportunities to invest and it's going to drive revenue growth downstream, we're going to choose to do that over short-term op margin. The op margin is going to be driven by the pace of revenue growth in a particular year, where in effect, we essentially just can't hire fast enough to keep up with it, or we think that there's not enough stuff for us to invest in, so we just wouldn't do the hiring at the pace that would keep up with the revenue growth. I made the comments on the Aussie dollar because you do need to keep that in mind. While we're going from up to AUD 0.80 now, it's something we can digest, but if it were to go to AUD 1.00, that's going to have an impact on us.
It's something that we'll smooth it out with our multi-quarter hedging program. In terms of if it were to go down to $0.70 where it was just a year ago, it's going to give us some lift in margin. We may just choose to reinvest that. Those are the things to think about. I don't know if Scott or Mike would want to add to that at all.
No, I think you added it. We think long term, we want to deliver returns long term without saying that this year we're going to deliver a specific return, which may or may not line up with the investments that we want to make. I can assure you, Mike and I, as the largest shareholders, are very much aligned with the long-term shareholder value. I think you answered it well.
Thank you.
Hi.
Sorry, yep. Over there, sorry, Gregg. Yep.
Gregg Moskowitz from Cowen. Thanks very much for having us today. Murray, just to start with you, when you talk about annual price increases for cloud and server, is your expectation that barring any significant macro swings, of course, that each of those products would have a price increase each year? Or is it more along the lines of some products one year followed by others the second year to the point where it might be on a staggered basis?
What we're seeing right now is the thinking would be it'd be across all products. That's not to say that it could be one or two we decide not to for various reasons, but we need to get ourselves into an annual increase mode and not this disruptive thing every few years and then we get things spun up a little bit internally about, hey, what's the customer reaction? We haven't done something for a number of years. We just need to get into the program of annual price increase. Our costs are going up, as I said, every year across the different dimensions, and we want to continue to invest in our business, and we're going to be looking to pass on a little bit of that cost to our customers. Again, we're not going to disrupt our model.
As I said earlier, whatever you were thinking our revenue growth is in multi-year out, this would give us a little bit of boost.
All right, great. On the product front, I wanted to ask about Stride. Sri, maybe to start with you, were there any technical issues or limitations that prevented you effectively from layering on some of your new innovations like actions and decisions or focus mode on top of HipChat in the cloud? Then either for Mike, Scott, or Jay, can you talk about the migration path for those HipChat Cloud customers? Given that Stride does carry a price premium versus HipChat Cloud for the Pro version, will there be a financial incentive for them to do so? Thanks.
Yeah. I'll answer on the Stride side, and Mike can answer the rest. As I said, Stride is built upon this new teamwork platform, so we wanted to make sure that the foundations that we are building, we want to build on top of that. That was the primary reason why we didn't want to bolt on to what we had before. We thought as we're changing completely, having all text, audio, video coming together in one product that's seamless, it made a lot more sense to build it ground up, which is why we took that approach.
Yeah, that's a great answer on the engineering side. I've forgotten the question completely.
Migrations.
Sorry. I was just thinking what a great answer the last two guys have given, and I've miffed it. Just on migrations, that's right. From a technical perspective for HipChat Cloud customers, your data is already in Stride, actually, right? You don't notice it, but we're constantly migrating it for you on an ongoing basis. It is a choice-driven upgrade, so effectively, once you log in as an administrator, you can choose to migrate your organization or not. It is a one-way migration, so you can't go backwards, but from a data point of view, it doesn't take any time. You don't need to go down. Your users need to download new clients, go to the App Store, and away you go, right? It's a very seamless migration process. At the same time, we're not pushing customers to do that.
They do it at their own choice, right? It's not like on Friday, everybody has to go.
That's, I guess, the first thing. From the pricing point of view, any existing HipChat Cloud paid customers get a year's worth of the existing pricing. They're grandfathered for the current pricing, basically $2 a user. The Stride free tier does have significantly more than the HipChat free tier did in the cloud. There is some chance that some customers may move from the paid tier down to free if they deem that to be enough. Obviously then, we wouldn't grandfather their pricing in terms of charging them. They would effectively be moving back to the free tier.
Let me just add one thing there, which is the guidance we gave on July 27th, it anticipated Stride's announcement and any impacts to revenue in fiscal 2018. It was already considered in the guidance we gave of $826 million-$834 million on July 27th.
Sanjit?
Yeah, Sanjit Singh, Morgan Stanley. Thank you for taking the questions. I had a couple of product-related questions. First, on Trello. I think for investors, when they see R&D being 37% of revenue, there's maybe an implicit assumption that most of the R&D innovation's going to come organically, and Trello was a rather sizable one. I wanted to see if you could talk a little bit about plans with Trello, and how does sort of one plus one equal three over time? I know Trello's been running sort of independently. Could you talk a little bit about some of the initiatives to sort of cross-sell and integrate the rest of the portfolio with Trello?
The first point, and most important point is Trello in and of itself is a fantastic product. It's a fantastic business. A very fast-growing product of its own. Our number one priority, as we've said publicly a number of times now, is to maintain the momentum of the existing Trello business. Right? We can give them more engineers, we can hire more staff for them, we can help them. Otherwise, they would have to be raising money and those sorts of things, right? There's an ability just in the existing Trello business that we are driving at to continue its momentum and growth. You saw it pass 25 million users. It is a very large product, right?
As Michael said, there's pretty much, with some asterisks, as far as we can determine, a user in every single country of the planet, including now in Antarctica. It is a very, very large product in and of itself in terms of its reach. Now, it's very early in its monetization journey. That's also something we've said we've got a lot of experience at. It's only started monetizing about two years ago, and they've had a couple of shifts in their pricing model. I think that's something we can also help them with. Again, to Murray's caveat, all baked into the forward guidance. The thing you've seen today that's maybe a little bit different is us continuing to integrate Trello across the platform, across the different family of applications. Not just ours, but also third parties.
The Trello embedded stuff, it's now in Confluence, it's also in a lot of other products, right? It's in Microsoft Teams, it's in Quip, appear in Dropbox Paper. Trello being, again, a platform shows, as Scott was saying, our ability to play well with other people. At the same time, it's synced into Jira. We put it into Bitbucket in a really deep and integrated way. You can see it appear in Stride. You also see Stride appear in Trello. That's continuing to help the user bases play well together.
That's helpful. I think the follow-up question I had has to do with what you're looking for in terms of. Oh, I forgot my question.
It's contagious.
Yeah. I'll take a pause here.
Itai.
Yeah. Hi. Itai Kidron from Oppenheimer. Again, thanks for hosting the event, it was very helpful. First question to Murray about the gross margin. I want to tie this into Project Vertigo, which was very impressive. You were able to make this transition with no hiccups in frame performance and uptime while doing this. I guess I'm trying to reconcile the move more to AWS and making that transition to that platform that enables you to scale more users. That's understandable. On the flip side of it, if you're able to scale much more, much faster, much easier, and I understand you're carrying now the hosting cost on your end, but why wouldn't also the increase in scale that you can now deliver in the cloud offset some of that to make for a neutral impact on gross margin rather than a negative one?
The comment that I made on gross margin is long-term in nature. The server gross margin is definitely higher than cloud, meaning either if we were running internally or at AWS or some other third-party cloud provider. It's really going to be the amount of revenue switch over than it is like the cost optimization. Clearly, going to someone like AWS that does it out for a living, they're going to be more efficient than us doing it internally. Yeah, there's going to be something there in terms of savings, but it's the transition from a higher gross margin to a lower gross margin business. Still really good gross margins in cloud, but it's that transition that's going to have the effect on the sort of a slower decline in gross margin as we go out a number of years.
Fiscal 2019 might be a little different just because we're going to get the relief on this accelerated depreciation that we're incurring right now. We won't face that in fiscal 2019. 2019's a little bit different, but beyond that, it's really going to come down to what customers choose, and if they go to the one that's a lower gross margin, meaning cloud, then that could have some of an effect on that. Clearly, when you get to operating margin, it's a different story.
Right.
For gross margin.
Will it be fair to say that the gross margin within cloud itself should improve over time?
It can, but the difference between a cloud gross margin business and a
server business where we're not hosting any of those costs. It's material different. It's really the mix of revenue between the two. It's not the inefficiency or efficiency of operating in the cloud. I would say that, and it shouldn't be coming from the finance guy, but the speed of development and the innovation that we can do on the products is primary, the cost savings second. Performance to the customer and reliability, all that's more important than trying to squeeze that last dollar out right now for us.
Yeah.
We're still going to do it. I sit next to Sri every day at work. "Hey, Sri, when are we going to reduce storage or whatever?" There's always going to be that, but it's really just a mix of the business that's going to have the impact-
Got it
on the gross margin.
Scott and Mike, I wanted to dig in a little bit into Bitbucket a little bit more because there are real clear established competitors in that space. I'm trying to get a better sense of how many of your Bitbucket customers are ones that started with you with Bitbucket versus started with Jira and then added Bitbucket. I'm trying to understand what's the leading pull into that product, how much of that is a developer that truly wants a code repository, I'm going to go and get Bitbucket, or you've managed to upsell me through Jira. I'm trying to get to sort of the natural expansion of this product going forward against established competitors in this space.
Yeah, that is an interesting question. We look at that across Atlassian, too, because many businesses have a very clear, "I land with something, and I expand with everything else." Square, which went public about the same time as us, had the Square Reader, then that's their entry to then sell them finance and cash registers and everything else. At Atlassian, we have many land opportunities, we land with Trello, we land with Bitbucket, we land with Jira, we land with Confluence, we land with all of our products, and we cross-sell significantly. Those two particular ones, I couldn't pull the exact numbers out of my head, but there's not an overwhelming winner where they go one way and then the other one is a tag-along. We actually have a lot of them come from both directions.
It would be a much simpler business to model if we had that land with one thing, because it's a lot easier for everyone. I think it would be a less valuable business because we have a sort of a redundancy, or we have many land opportunities, which means that we're not dependent on a particular entry into a company. We can land many different ways. The destination of an enterprise license agreement or tens of thousands of people using us is the same, but we have many different ways to get there.
Thank you. Great. Thanks.
Sanjit, do you have a second?
I remembered what my question was. That answer last time was just so good, it just shocked me. On the competitive environment, when people talk competition with Atlassian, a lot of it's focused on the messaging space. I guess my question is, in a market where pricing is coming down on sort of per user, right? You guys are attractively priced. Microsoft embeds it in its Office 365 solution. How important strategically is it to win in messaging when the underlying pricing of the market is coming down? Why is this space important? It's not just an account for a significant part of your revenue. I just wanted to get your view on why it's important to win in this space.
I would say it's, as I mentioned at the top of my session, it's part of the way teams work. I think it is vitally important. It plays sort of a role in connecting people with their activity and work product and what they're trying to accomplish. If you actually take a step back from messaging, it is one flavor of communication. Maybe the one word I probably use the most when I described all of our products was communication because Jira ostensibly is a communication product about projects and tasks and activities. Confluence is a communication product about pages and content, Bitbucket about code. There is a real-time messaging and communication connection point that's pretty vital to teams.
Again, from a competitive perspective, remember, and I think this is still true of messaging, there isn't necessarily, even inside an organization, one messaging system that everybody will use for always. In the same way that's not true as a consumer. Most of us have lots of different messaging applications. I know I have three that I use just to communicate with my wife. I think it's important, and the other thing is what you saw in Stride is I think a lot of invention and innovation and not just saying messaging is a solved problem because it's not. Real-time communication is not a solved problem. I think most of you may relate to the problem that we're trying to solve with Stride, where if you've got chatter messaging, you might pivot around a GoToMeeting or Webex or Google Hangouts.
You might pivot back to the thing that's sort of tracking what you all agreed you're going to accomplish. I think there's real innovation that we've introduced with Stride based on the DNA around just teamwork generally.
Nick, over there.
How are you doing? Nick Lawler from Foxhaven. As you've removed the user limitations on cloud, how do you envision or kind of predict the customer base will evolve in its adoption between data center and cloud over time? Does one mode over the other enable different things in terms of what you can do with the product long term?
If anyone knows the answer to that question, please let us know because every person has always tried to predict the exact moment in which all the enterprises are going to switch to the cloud, and it differs by different industry, by different geography. For us, we are agnostic. We get it roughly from an economic perspective, from a cost of support, from all these, we are really agnostic as to where our customers go. We know that in 20 years' time, most of our customers won't want to rent space in a cold room and rack their own servers and upgrade operating systems. They'll want to just pay us to do that for them. The path between here and there, I can't give you an exact time period.
All I know is that we are encouraging customers to move to cloud as we support them and all the things that they need. At the same time, we're just as happy if they stay on the server business. That server business has got a lot of legs in it. It's not something that's going to go to zero anytime soon because a lot of customers still need it, and there may be some customers in a very long time, government agencies, who may never, ever want to put stuff in the cloud. That may have a very long longevity, may have a very long life in the server business as a result.
I can tell you a short story, just because we're at Summit. For about three or four years at Summit, there was a particular large customer that was really pushing Data Center and Premier Support and the TAM program. Like all the enterprise features, they were really pushing. Every year we said, "Well, what're you doing about cloud? What's your long-term strategy? How do we think about how you're going to move so we can learn about all the customers?" Nope. Turns up last year, new CIO come along, 100% cloud. We're all going to cloud. It wasn't an organizational shift that happened with learning, it was new guy came in and went, "We're moving." We were like, "Oh." We started running, moving. He's like, "Okay, we're moving over the next five years." Oh, okay.
The new person came in, was a massive shift for that one organization. When you spread that over the lot, it's going to take a long time.
Yep. Sorry, Steve, down the front here.
Hi, Steve Cameron. I'm trying to get at the Microsoft Office for Teams question because you guys have seemed like you've really nailed the IT software development space. Sounds like Service Desk is doing well, but the upside in the model and the valuation, at least to me, looks like that broader opportunity. I guess qualitatively, how are you doing on that? Then kind of an ask is what can you tell us on an ongoing basis to give us a little more visibility on development of that separate, much larger business?
Which separate, much larger business?
Microsoft Office for Teams, on the business users, not the IT user. If you did your chart, your pie chart, not the IT users and not the developers.
Yeah, business teams?
Yeah.
Yeah. Well.
Jira Core.
No, not just Jira Core. In fact, Jira Core will probably be the minority of adoption outside of business teams. Confluence, Trello, Stride. I think there's a greater propensity for less technical parts of the organization to adopt and expand with those products. Just remember that the point that we echoed, that not all teams are the same. They work differently, and there's a variety of tools that they might need at different points of what they're trying to accomplish, to work better together. Office plays a role in that, largely around individual productivity. You start with something that you need to iterate on just for you. Microsoft has, as they have done for many years, have layered on sharing. SharePoint was one early example of that a couple of decades ago. Teams in the communication messaging space is an example.
In some cases, we will displace the need for that. Doesn't necessarily mean that companies are going to cancel their Office 365 subscriptions. They're still going to have that on your desktop for the things that Excel, PowerPoint, and Word do. You may share that through channels that Microsoft gives you. The team, when they're talking about beginning something that they're creating together that they know this is a specification or requirement or a business plan or a launch plan, they're probably not going to start in Word. They're just going to start in Confluence or a product like it, and then move the whole collaborative life cycle around that. We might display some of those use cases where I've still got Office, but I don't use it for a lot of things that it wasn't really built or equipped to solve.
In many cases, we're going to coexist with it, Office documents might be attached to a Confluence page, or they might be piped into a Stride channel, or they might be part of a Jira project or Jira plan. I think it's just part of serving the really diverse set of ways that people work and giving them more ways to be effective. There's got to be more questions.
You guys must've done a great job earlier today. If there are no questions, well done, team.
Okay. Viswanath, you're going to have one. Yeah. Sorry, Mike, down here.
This is more of a development question, was it Facebook says, "Move fast and break things," and things like that. You have a big R&D staff. How do you kind of balance what Zuckerberg says, don't break the infrastructure, but you can do other stuff. How do you kind of keep that creativity going? You have a big organization. That's a non-trivial job, as they say.
Yeah, no, great question. I think about this all day. My goal is to make sure that the engineers and the execution team moves as fast as possible, but it's not breaking the customer, right? Like I said, we do A/B tests and gradual rollouts, we want to make sure that when they move fast and finish a project from their perspective, it's running in 1% of the customers. We can tune it and make sure that it works well before we ramp up. Every time we have a strong process of how do we deploy and make sure that there's progressive rollouts when we do that. We have gates in place when things go out to customers, but for engineers and for the product team to develop, it's much, much faster inside.
You know there some people think it might, some people think it's real. Yeah. The question is. Some people think it's going to kill us in a few years. Yeah. What are we doing on machine learning? I mentioned the smarts. Smarts is our practical way of implementing machine learning, we have the data. We are working on making sure that we have the Teamwork Platform that's smart and exposes APIs and graphs that's needed to build the products. We embed those features into each of these products, they become smarter. Right? We're taking a much more practical long-term approach to make our products smarter.
Yeah. I would say that if you think about what AI is really doing, it's a series of algorithms. Very, very complicated, very, very smart algorithms and a massive amount of data, right? You put those together, if you Google, you've got a car, right? The good thing for us is a lot of those algorithms are open source, given away. Facebook, Microsoft, Google, all giving away their ML framework, that's great for us. We have a huge amount of data about how teams interoperate. Where are people? Which pages are they using in common? Which spaces are they using in common? If we look at pages and, sorry, pages and issues and messages from HipChat or Stride, we can determine a lot about which people actually work together.
The task for us is to take that dataset and turn it into features that the customer doesn't even need to know are happening, right? If you look at Stride, it's a good example. When you mention somebody, we've learned a lot from all your applications, we make sure the right people within 2,000 people at Atlassian turn up faster. It's a very simple feature. The user just gets a small bit of the light, probably never notices that happened, and that's exactly the way it should be. The more you use, the more applications, the more data we have, the better we can get.
Any other question? Sorry. See you down here. One mic over here, please. That was one back. Sorry.
I'll double-dip. You've got the sort of two revenue lines, which is the license and maintenance, which I think is growing around 20%, and then the cloud business, which is growing, I think, like 60%. I've got a terrible memory, but it's something like that. How much do those cannibalize each other? The other way to put it, because that cloud business is getting to be 45% of revenues, is that actually your natural growth rate, which is 60%? And/or how much are those drawing from different sources and/or do they interact? If you get the question.
Yeah. No, I understand the question, I think. Approximately three-quarters of the new customers every quarter are just going straight to cloud. There's no transition occurring from server to cloud. They're just coming straight into cloud. Where we're seeing transition on those lines would be if you're a server customer today and you go to data centers, you're still kind of running a server behind the firewall, but you moved to an annual subscription, that's getting reported in that subscription line. It's commingled in there with cloud to a certain degree. There is some of the cannibalization, if you will. As you saw, you're probably going to be, data center, a lot of the customer use cases are these much higher prices, as you go through. That's actually kind of a lift for us in a sense.
The server business is not going away. It's still growing. It's still the majority of the revenue. It's not like, again, like 18 months from now, there's going to be this big switch. If the server line goes down to less and less revenue and becomes less relevant and the server keeps powering it, I mean, the subscription line powers ahead, the 60% number you gave, therefore our overall revenue growth goes up that much. There's just nothing that's going to happen in the short term that's going to have those kind of swing. It's going to be something that is going to be very deliberate. Obviously, we're going to battle the law of large numbers and those kinds of things, but there just isn't going to be that transition.
We've said here repeatedly that we're not trying to force our customers to a certain deployment model. Certain parts of the world, they're not going to cloud. Certain government entity, they do not want to go outside their firewall. We have the position of, we grew up in the server. We've got it. You probably wouldn't go develop it today, but we've got it, and we're going to use it as a competitive advantage for ourselves.
I guess that subscription line is growing so fast. I think most people are assuming your revenue growth rate's going to decline. It's 45% of your revenues, and it's still growing pretty darn fast, does it start to dominate the business, and are suddenly we looking at a 40% growth rate company?
Right now we're fiscal 2018, $826-$834, we got price increases in the future. Outside of that, we'll see how it goes. I would say this, if we were growing 60% top line, we'd probably have some operating margin leverage for those looking for that.
[Its a gain].
Just lastly, can you tell us how the CFO search is going for you, Scott?
I can say that. Since we announced, in the quarter four earnings, I can't remember when it was, when the last earnings call. It's going really well. I've met with a whole bunch of candidates. Obviously, really hard to find someone as amazing as Murray. He's done a fantastic job here, but there's some pretty qualified people out there who've also been public company CFOs, who've done amazing careers and are looking for an incredible company like Atlassian. I can't give you any more than that at the moment, pretty excited by it.
Sorry, one down here as well, Nick, I think. Sorry. Right there. Yep.
I guess since you didn't cover it a lot during this presentation, on the Marketplace side, it seems like from your recent reporting, that's been growing really nicely. Can you guys talk about what you hope for from the Marketplace and what kind of extensions of your product you want from the Marketplace versus what you would build yourself?
Sure. Look, we've been really happy with how the Marketplace is growing, obviously, over the last, what has it been, three or four years now. You can wander around the hall out there and just see two things. One is the volume of vendors that we have now, more than 100 just here
1,000 in the marketplace, as we said this morning. The volume of vendors is really good. The other thing is the sophistication of those vendors, right. There's increasing numbers of vendors out there with these big booths. They come with marketing teams now. Their businesses are going really well, which is an awesome feeling. It's great to be able to be helping them to power their businesses. We have always said that there's a lot of opportunity around our products. Our focus is on solving that core collaboration problem at the center, and that leaves a lot of opportunities around that are hard for us to exploit because they're a smaller number of customers or a niche or something like this. Tempo is a classic marketplace vendor that really focuses on time tracking. Consulting companies, people have billable hours, all this sort of thing.
It's a niche, but it's a big niche, right. They built a fantastic business on it, and it's not a core collaborative problem. We can have a really symbiotic relationship with that. We've encouraged our marketplace vendors to look for opportunities like that. There's many out there, pretty excited about where it's going.
I add to that. I know we know that people are customers that buy more, buy add-ons from the marketplace are stickier. They stick around longer. That's primarily for us. We make some money on the marketplace, it's really about making sure our customers are happy, we provide all the things that our customers need. The other one is, from an M&A perspective, if there's ever opportunities we do need to own, it's fantastic. They're already built on our platform. They love our culture. We have, I think, done one or two opportunistic M&A things from our ecosystem historically, and they've always worked out really well for us.
I think we have a question down here.
As I'm sure you all know, Atlassian was featured in the Daniel Pink book, "Drive," where he talks about autonomy, mastery, and purpose, and that's his cultural framework. Is that the cultural framework that you would just say you subscribe to, or is there another way, pithy way that you think about your culture?
I'm not going to say Our culture, we have five distinct values. You can go read on our website or any one of our offices. We bake that into how we hire people. We bake that into our entire compensation philosophy. That runs through everything that we do. I like to think that means that we make better hiring decisions, and we have lower attrition, and we have engaged workforce. We're really strong on that. I don't think mastery, autonomy, purpose, that's sort of more I think that's universal. That's how people want to come to work when they have mastery, good at something, getting better. Autonomy, they own something completely. Purpose, where they're here for Their job has meaning in the world. We subscribe to all those things, but I don't think that differentiates one company from another.
I think our values and our culture would differentiate us from peer and other companies.
[Anyone], maybe. Sorry. Jackie first, it's fine. You're closer. Yeah.
As you all look at your R&D priorities, how do you balance the decision to go deeper into a certain area like Jira, where you did software and service, versus going to another area outside of what you're doing?
Well, I don't know it's an R&D priority. As a company priority, we look at is where do we want to go? Lucky thing is we have a lot of opportunity for us. If you think our mission is to unleash the potential of every team, there's about a billion knowledge workers out there today, and we're only just really scratching the surface. For us, it's a matter of where can we play, where there's a competitive moat long term, where there's real customer value that we provide, and we have a skill set that we can do that well. We take all those things into account. Sometimes we decide, okay, that's an extension to one of our products where we sell into our existing base.
Something like Jira Portfolio would be a good example of that, where it's a product, it sells to a small part of our base, really, we have a moat around that. We're very well-positioned to sell that. There's other products, whether it's Trello or whether it's Stride, that then take us into new markets and new customers and help us acquire net new to us. We take all those things into account. There's not a perfect formula, but we do think sort of competitively where the landscape is before we make any moves.
Maybe two more questions. Ninoi, did you have one there?
Just two quick follow-up questions on the broader opportunity outside of IT. I guess firstly, the examples that you had here on the broader expansion outside of IT was mostly the IT teams using it and then someone else seeing it and then adopting it. Is that kind of the primary way to think about growth in that side, or do you guys actually land outside of IT? The second piece is, are you seeing any specific use cases or departments that are inflecting more earlier compared to the others?
What was the second part of the question?
Within the broader opportunity outside of IT, are you seeing any use cases or departments that are like HR versus legal-
More than others
are doing earlier than the others?
Yeah. The first part of the question is, those two pretty good examples happen to have been server customers, where I think probably IT is a more prominent conduit to the rest of the business. They've already established sort of the product inside of the business, and oftentimes the business is going to come to them and say, "Hey, is there a thing that can help me?" I think in the cloud, the friction path for less technical teams is a lot easier. You don't have to install anything. You can sign up for All of us in the room can sign up for Trello in a couple of minutes, get going with Jira with a team of 10 in a couple of minutes. I think that there's a more predominant land path for business teams.
If you go to atlassian.com/teams, I think as an example, you'll see some of the use cases that we provide as examples to marketing or HR or finance or sales to get going with our products. IT continues to be an important conduit because oftentimes the business is going to go to them and say, "Hey, listen, you guys are the technology team. Can you help me solve this problem that I have?" Or that's part of their job is to identify where they can, but it's also common for business teams just to get going on their own. The second part was?
Oh, departments? No. I think it's like we solve general problems. There's some examples that you've heard us highlight before, the quarterly close process in finance or the marketing campaign planning and launching. I think it sort of runs the gamut. The best way to do this is just walk the halls, grab a customer, and you'll hear 100 different things that actually people are applying work, content collaboration, communication, and messaging products to solve.
Let's take one final one if there's one. Sanjit, last one. Yep.
Maybe just to end up on a strategy question again. I think some of the players in the space are looking at the market in the view that they're trying to win the number of hours per day of the user, right? Essentially trying to become an OS for collaboration. You work in a single product framework and connect to maybe your Sales force or expenses for coupon and sort of own the identity of that user. To what extent do you think whether that's the right strategy for this particular market? You guys have, as you guys mentioned, multiple great landing point products versus a single point product. In terms of this concept of becoming an OS for collaboration, what's sort of your view on that?
Our goal is eventually to have every knowledge worker in every company use an Atlassian product every day. That doesn't mean we want them 12 hours a day in a product, and we want to actually have outcomes rather than just inputs, and especially in the messaging space, there's a lot of hype around how many hours a day do my people sit in their messaging. It feels like we're in a Dilbertesque-type environment where people's bosses are monitoring how many hours are you online, but not actually the outcomes that you're doing. You saw a lot of that this morning with the way we're approaching that specific market is around getting work done because talk is cheap, and it's great when talk turns into action, but if you just talk all day, you're not achieving something.
I do think there is an element of some attention. We want to be across all knowledge workers, but it's not about how many minutes per day they spend. Our tools are to get work done. That's what we really care about.
To your broad question to start around operating systems, though, one thing I do want to say is that we have a history of building very specific point products for specific purposes. Bitbucket, Jira Software, Jira Service Desk, they do one thing and they do it really well. If you think about our model, that ties in really nicely because we land with one thing that's easily definable. If I'm explaining to a service team what Service Desk does, they get that. They can land, they can use it. If they have value, there's more service teams, or there's other products that work well with Service Desk. I know we don't subscribe to the philosophy of kind of a broad-based collaboration product that does all things for all people. I'm just not sure that's likely to be a successful strategy.
We tend to build a lot more point products. Even if you see the way that the Teamwork platform is being built, its capabilities that are surfacing in all the point products, and then various sort of simple overlays to cross-reference bits and pieces rather than a single broad collaborative operating system product, which I'm not sure exists.
Great. Well, thank you. That ends the formal part of today's session. We're going to move outside here just for informal cocktails. The full team here is going to be outside there, so feel free if you have more questions outside. We will be posting slides from today on our investor relations website in a couple of hours. Thank you all, and we'll hopefully chat a bit longer out there. Thanks.