Please welcome Dave Gennarelli, Autodesk Investor Relations.
All right. Good morning, everybody. Welcome again to the Autodesk Gallery and to the Autodesk Investor Day 2018. By now, you should all have an agenda on your desktop. You'll see that we'll take a short break in the middle of content. Please hold your questions till the very end, then we'll do one long Q&A session at that point. Then I welcome you to stick around, have lunch with our speaking executives, then feel free to look around the gallery and spend some time here. With that, we have a lot of ground to cover, so we'll get right underway. Thanks. Ladies and gentlemen, please welcome Autodesk President and CEO, Andrew Anagnost. Morning. Good morning, fans, skeptics, everyone in between. How's it going? Was the breakfast good?
Sorry I wasn't down here chatting. I had a customer call this morning, and that's where you want me to be, so it was a good thing for me to be doing. I'm gonna spend most of my time talking about what Autodesk looks like in five years and the five outcomes we're gonna drive to get there. I think I'd be a little remiss if I didn't spend a little bit of time just kind of talking about, hey, what's happened over the last year, and what's happened since the last time we all got together at the Investor Day. First off, let's talk about what's happened over the last year. Last year, we actually booted up a whole new leadership team.
We rolled out a whole new set of strategic priorities to the company, we used those to increase the executional focus inside the company. That's a lot of change to absorb in a year. I think one of the exciting things to talk about with regards to all of that change is the outcomes we had as we got to the end of the year. By the end of the year, we'd grown the recurring revenue 25% year-over-year, absolutely above our long-term targets. 5% ARPS growth year-over-year, the much-anticipated ARPS inflection, the stock price grew 42% year-over-year. A year with a lot of change, but a year with a lot of results as well.
If we look back at the Investor Day that we last had, there's a few things that actually came true exactly as we said they would, I want to just review them at a very high level. First off, last time we said that the recurring revenue of the product subscription and total subscription line was going to exceed the maintenance line, it did. It's now become the majority portion of the recurring revenue of the company, which is a great outcome, it's continuing momentum. We also talked about ARPS inflecting at the end of the year. ARPS inflected at the end of the year.
The other thing, although we didn't give exact specifics about what the subscription targets were for FY 2018, but it's important to note that the core subscription-based growth, what we've been calling core EBAs, maintenance, and product subscription, is exactly on pace to what we expected when we were standing up here at the last IR Day. Remember, and I'll show it, and I'll reinforce it, the core drives most of what's happening as we go to FY 2020. It's not what's going to drive what happens as we head to FY 2023. Now let's change gears a little bit and talk about Autodesk five years from now. The Autodesk that we're going to be talking about when we're standing up on this stage five years from now is going to be quite a bit different from the Autodesk we're talking about today.
Five years from now, we're not just gonna have a conversation about Autodesk being the leader in desktop design software. We're gonna be having a pretty nuanced conversation about how Autodesk has become the leader in design and make automations. I specifically use the word automations in there, and I'll talk about what that means, and you'll hear a lot more about what that means as we go through the day. What this is gonna result in is that we're gonna be very tightly integrated into our customers' processes. Not only helping them come up with design solutions, but actually helping them figure out how to turn something into a physical asset.
That's going to be a really exciting transformation. It's also going to lead to another important transformation. It's going to drive the change of Autodesk into a customer company. Now, we've been well known as being a desktop software provider with a lot of really strong distribution capabilities. When we're standing up here five years from now, people are going to be looking at us as a company that's really focused on the success of our customers and taking a lot of responsibility for the success of our customers. That's going to be an inevitable outcome of being tightly integrated with their design and make process. Five years from now, every company in our space that matters is going to be a subscription company. We will not be having this conversation about who went first, who took the arrows, who actually got out in front.
Everybody that matters is going to be a subscription company. Because we went out there first, as we head into the discussion that we'll be having five years from now, Autodesk would have been able to apply machine learning in combination with the data that our customers agreed to share with us to really create push-button automations. Push-button automations that allow them to automate the process of taking a design into a set of instructions that can either be manufactured or built. That's going to be a very significant change, because even though every company we compete with is going to be a subscription company, Autodesk is still going to be ahead on the business side because we would have become what I'd like to call a consumption company. What that means is we're not just subscribing and selling access to solutions. We're actually selling people outcomes.
Push the button and turn your design into a set of instructions that can instantly be streamed to an automated factory. Push a button and turn your design into a set of constructions that can be used on the construction site. Not only will people be paying access through a subscription, they'll be paying for outcomes through consumption. That is going to be a really significant change, and it's going to be something exciting that we can talk about in five years. It's also going to mean that we're going to empower our customers to make a lot more products, make them better, with higher quality and reliability, and with less negative impact on the world we live in. I want to spend just a few minutes on this slide.
You'll hear a little bit more, I want to tell you why I think this is so important. First off, this concept of more. More is inevitable. Population's growing, the middle class is growing, developed world, there's going to be a huge increase in urbanization in the developing world. People are going to want more. More buildings, more roads, more infrastructure repairs, more products. They're going to want those products to be very customized to what they want and how they want to use them. The problem with more is it's really difficult to deliver more with the current capacity. In the built world, there is a huge capacity constraint problem. There is not enough people. When you think about automation taking away jobs, let me be very clear here.
In the built world, there is not enough people to meet the capacity of what we're trying to do. There's not enough money, there is just not enough natural resources to do more. If you provide the right kind of automations, you're going to let people be able to do that more, do it better, do it with fewer costs, lower cost per project, per product, less energy, and less use of resources. We believe that this is going to be our new competitive differentiator in the space. We're working really hard on it. You'll hear a lot more about more, better, with less negative impact on the world as we move through the next five years. That brings me to the five outcomes. What's happened for Autodesk over the next five years that's brought us to being a different kind of company?
Here's the things I want to walk you through. Throughout the day today, you'll get a little bit more information about each one. I want to touch on each one just a little bit, one at a time. Let's start with, we would have completed the subscription transition in five years. It specifically says, completed the subscription transition and not the business model transition or the business model change. The reason we use that language is because this isn't just about creating value for ourselves. It's about creating value for our customers and ourselves. Because the customers need to come along with us on this transition. For them to come along with us, they need to believe we've created a valuable outcome for them.
When you look five years from now, people will still be using these traditional desktop products, AutoCAD, Revit, Inventor. They'll be very different. How the products are delivered, what they're connected with will be different. The customers will see a value increase that they're really not expecting right now in terms of capability, functionality, and value. That's an important focus of what we're trying to do is ensure that these customers believe that the subscription transition has been good for them as well as good for us. When we look at the outcome for ourselves in terms of value creation, when we last talked at IR Day, we rolled out 24 320. And by the way, the $1.4 billion in free cash flow, which on a constant share basis translates to $6 in free cash flow per share.
It won't surprise you at this point because we've been setting the stage as we've been having discussions on the earnings call that we're actually modifying this paradigm a little bit. I'm going to show you how we're modifying it. It's moving to a formula of 25 6 18. Actually, what we're doing is we're talking about a better outcome today, higher growth in recurring revenue at higher ARPS and at a lower subscription CAGR. You're going to get a lot more detail on this as we walk through the rest of the day. This is a pretty important change. It's a great change. It actually says we're going to deliver more with less, not to play off that earlier phrase. What we're not backing away from is the $1.4 billion in free cash flow.
There's a bunch of things that we're changing but reaffirming too, and it's an evolution based on the performance we've had to date, and I think it's a really positive outcome. Let's look at how we're going to get there, and I'm going to bring back some slides I brought up last time. Some of you probably remember this slide where I broke out the ARPS and the ARR and subscription growth numbers between FY 2016 and FY 2020. What you see, and what I want you to remember, is that the core, in terms of getting to FY 2020, is the big driver, and the core being the sum of product EBAs and maintenance.
Yes, there's contributions up there from the new businesses related to the cloud. He put his glasses on as soon as this came up. You can see that the core is the big driver, and you can also see how we modified the subscriptions. You're going to learn a lot more about how we've consolidated the cloud portfolio to a tighter set of subscriptions later on in the day. You'll get a lot of the logic around how this changed. You can also see how the ARPS have trended up relative to previous statements. If we look out from the gap between FY 2020 and FY 2023, this is something I didn't show last time.
This gives you a snapshot on how we continue to drive growth with the new businesses and with the core. One of the things that might surprise some of you is the core continues to deliver value out to fiscal year 2023. The cloud and the new businesses, and the cloud's really just a proxy for the new businesses, becomes a much more important component of the growth that we deliver in FY 2023. You can see it's of the scale of half a billion dollars. I mean, the scale is kind of obfuscate how big some of those numbers are in those little boxes. You can see the CAGRs from ARR from 2020 to 2023 and the subscriptions, and you can see where the ARPS start to trend up.
That should give you a feel for how we drive to some of the financial outcomes we'll be talking about today. Again, you're going to get a lot more detail. The next thing I want to talk about in terms of the five outcomes is digitizing the company. Five years from now, Autodesk will be a fully digitized company. What does that mean and why is it so important? If you want to be a design make provider in the cloud, you have to enable a high degree of self-service and visibility to the customer that allows them to understand what they're doing with their Autodesk products, how they're using them, and how they can manage their relationship with Autodesk all on their own.
At the same time, you have to be able to provide instantaneously real-time insights to everybody inside of Autodesk about what are the customers doing, how are they using the products, and how can we help them be more successful. When we talk about Autodesk being digitized in five years, that's what we talk about. A flexible, scalable environment that not only delivers a great customer experience, but delivers a great ability to provide self-service to our customers and self-service internally so that we can understand our customers a lot better. We'll be able to do that five years from now. We'll be able to do it a lot sooner than that. The outcome I'm particularly excited about is this third one. This is this idea of moving the building information model across the entire construction process from the start to the finish.
What does this mean? What this really means is just like what's already happened in manufacturing today, is that the building information model becomes the record of everything that's happening. Not just the design decisions, but the decisions that go into how is this particular assembly of building components disaggregated into parts that can be manufactured and then delivered to a construction site and assembled in a particular sequence, which, by the way, starts to sound a lot like a manufacturing process, which we'll talk about a little bit later. The key point here is that the building information model is the single source of truth across the entire process. Again, this is another one of these inevitable outcomes because this is exactly what happened in manufacturing, and it's exactly what's going to happen in the AEC world. Stay tuned for that.
When we're standing up here five years from now, it would have happened. In the manufacturing space, we got an even more exciting opportunity in terms of impacting what our customers do with their process. That's this idea of automating the idea, the concept of going from design to manufacturability. Always knowing that what you design is manufacturable, but by a particular technique, and then being able to make those decisions instantaneously. If you look out five years, and again, we'll get more detail as we go through the day, you'll see that we've automated the process of creating geometry based on what the customer is trying to do and the manufacturing facility they're trying to use. Instead of somebody trying to create the geometry and shove it into something, the geometry is created to adapt to what they're already using.
This is an important fundamental change. We've been talking about it for years. We're way ahead in executing on this. Five years from now, we would have made huge progress, and we'll be talking about a lot of the exciting automations we've been driving into the customer base. Which brings me to the last one of the five outcomes. This is this idea that manufacturing and construction converge. This is another one of these inevitable things. When I talk about them converging, I'm not talking about them converging to the same thing that manufacturing is today. No, both of these industries are going to converge to something different. Construction is highly flexible. You could do anything you want on the construction site. Everything in those drawings, everything in those building information models today is open to interpretation. Manufacturing, on the other hand, highly inflexible.
Factories optimize to build one thing really well, pump it out in mass quantities at high quality. What's going to happen in the future is construction's going to absorb some of the goodness of manufacturing and become more inflexible in certain processes, and manufacturing is going to absorb some of the goodness of construction and become more flexible, so that a factory that used to build one thing really well can rotate to build many things well at high quality. This fundamental change of these two very seemingly different industries starting to look a lot alike in the way they build things is something only Autodesk can drive. Only Autodesk, all right? I want you to understand why only Autodesk can do this.
Not only are we the only company that understands the building information model and the data and the leader in it, but we're also the only company that simultaneously understands manufacturing processes, manufacturing data flow. This is going to happen. Construction sites are going to have robots. They're going to have machines. They're going to have the equivalent of manufacturing engineers, and factories are going to be more flexible. We're going to help the two get there together. That's going to be one of the big exciting changes over the next five years, and I'm looking forward to talking about it. We're going to talk a lot about how we get there today. These are the five outcomes.
We've tried to set up kind of a narrative today so you can get at least a good chunk of information about each one of them. How we completed the subscription transition, which is not going to take 5 years, that's a lot shorter. How we digitized the company, driven the BIM model through the entire process, automated design for manufacturability, and brought construction and manufacturing together. All over a 5-year period. All ending with Autodesk becoming the leader in design-make automations. To help you understand the flow, here's what we're going to talk about today. Got a new management team. You're going to get to meet all of them. The first one you're going to get to meet is Lisa Campbell. She's our Senior Vice President of Business Strategy and Marketing.
She's going to talk about the construction and manufacturing opportunities, the TAMs, what we're doing today to address them, how we're going to address it moving forward. Scott Reese, who you haven't met yet, is our Senior Vice President of Manufacturing, Construction and Production Products. He's driving the cloud platform and all the cloud products that are focused on construction and manufacturing, as well as video production, which we won't talk much about today, but we'll talk about in the future. His job is to help you understand what the next generation of construction and manufacturing solutions are going to look like. Steve Blum's going to come up. You've met Steve before, so that'll be a familiar face.
He's going to talk about how we're building machinery to deliver the ARR growth, not only up to FY 2020, but into this new world of being a design-make company in the future. Scott's going to come up and give you a view of how the financials play out over FY 2020 and beyond. I'll come up and offer some closing remarks. With that, I would like to introduce Lisa Campbell.
Hi, good morning. Today, I'm going to talk to you about the next two big growth opportunities for Autodesk, construction and manufacturing. Construction really hasn't changed much in the last 50 years. Still very complex, traditional processes, not much productivity or technology tools. What we're seeing with population growth and with urbanization is that the number of construction projects are going up. That means they need technology, and that's going to create that first big growth opportunity for us. With manufacturing, we're seeing technology that is completely transforming the way you design products, engineer products, and manufacture products. The old ways of manufacturing products with mass productization, where you have the big scale, that's not going to work anymore because what people want is they want to have products that are made to order, and that is where you need new manufacturing processes.
What I'm going to do is I'm going to first dive into our construction opportunity. The United Nations has estimated that we are going to have 7 billion people living in cities by the year 2050. Let me just give you a little context about that. Today, we have 3.5 billion. Imagine all of us living in the cities that we have today. It's kind of unimaginable. The other thing that that means is the number of buildings that we have has to go way up. In fact, estimates are that we have to build 1,000 more buildings per day than we're currently building. Imagine doing that with the low productivity that we currently have in construction. What's the problem? Again, if you look at this chart, it's pretty staggering. This is construction.
Look at how much construction productivity lags against the total economy, look at how it lags against manufacturing. Manufacturing is three times more productive than construction, we all know why. They've been investing in IT technology for several years. Construction has not been doing that, but they need to do that. The reason for this productivity lag, as you can see, you look at this chart and you say, "There's complexity." That's because that's the current construction process. It's fragmented, there's no standardization. Coupled with that, it's all manual. Look at the Post-it notes. This is how we're planning complex construction projects. There's been no technology applied. I thought it would be good to actually show you a real construction project.
What you're looking at, if you can't guess, this is actually in Hong Kong, this is a metro project in Hong Kong. What's the first thing you think of when you see that? You think complex, you think messy, you think they need some technology. Imagine if leading-edge construction firms said, "How about if I use some advanced manufacturing techniques on my construction site?" Imagine what that could look like. Imagine what this kind of control management would do to the construction process. They have 30% waste on construction sites today, imagine what this kind of manufacturing technology would do to that kind of waste. The other big challenge. We said we have population growth, we have urbanization, we have to build 1,000 more buildings a day. Well, those buildings have to be built sustainably.
Right now, the estimates are that every building takes up, with all the buildings, rather, in the world, it's about 40% of our global energy supply and about 25%-35% of CO2 emissions. By the way, it's not just new buildings. What about all the existing buildings that we have? Those have to be refreshed. Oh, by the way, we have to build the infrastructure around those buildings. You take all of this into account, you might say, "Why is Autodesk so excited about this construction opportunity, and why now?" I'll tell you why now. Because change is happening. Three weeks ago, the U.S. put out a jobs report that we have added 313,000 jobs. One in five of those jobs is in construction. By the way, those are high-quality jobs. These are jobs where they need technology.
If you think that number is impressive, this is from the McKinsey Global Institute. With the current investment trends, what they are projecting is that we will add 53 million new building construction jobs and 34 million new infrastructure construction jobs. That's at current investment levels. If we actually built what we needed to build in order to hit these population targets in urbanization, you're talking about over 200 million new construction jobs for both building and infrastructure. All of these high-quality jobs that require technology. What's the other big change that we're seeing besides all of these new jobs being created in construction? We're starting to see technology infrastructure be adopted on the construction site. I'm talking about laptops, tablets, smartphones. Daily usage is going up. What does that mean? Because we've all seen this before.
When you see technology infrastructure be adopted, we saw it with CAD, and we saw it with BIM. Software applications follow. It's inevitable for construction. This is going to happen in construction as well. What I thought I would share with you is what does this mean in terms of a new expansion TAM for Autodesk. We are talking about an incremental $10 billion in new construction TAM. Last year, we talked to you about the $12 billion in design TAM that we already successfully pursue in AEC. In addition to that, we are going to be pursuing this $10 billion new TAM in construction, and we believe this will be the next billion-dollar business for Autodesk. By the way, we're not the only ones seeing these trends happen in the marketplace.
There's been over $1.6 billion in venture capital funds that are going into new construction software startups. These are startups that are seeing this exact same opportunity and the fact that technology must come to construction site and to digitize construction processes. What is Autodesk doing to secure this opportunity right now? We're building solutions in three areas. Oh, by the way, we're connecting all of these solutions. Solutions for the office because that's where pre-construction planning happens, that's where costing happens, and that's where procurement happens. Solutions for the trailer because that's where tasks are assigned, that's where safety protocols are defined, and that's where performance is analyzed. In the field because that's where construction actually happens. That's where they do progress inspections, where we track and log issues, and where performance is actually measured.
We are connecting all three of these, the office, the trailer, and the field. I thought what I would do is share a few examples of customers that are already benefiting from our construction solutions today. JE Dunn, for those of you who don't know JE Dunn, they're a top building contractor. This project is the Lenexa Civic Center in Kansas City, Missouri. Here's what JE Dunn did. They used BIM 360 in conjunction with two internally developed applications. What they were able to do is they took the BIM model, and they were able to provide real-time data, pricing data on every element of that BIM model. That went to the pre-construction team. Why is that a big deal? Because they were actually able to make decisions quickly and efficiently. Structure type, skin type, square footage. That process is usually highly fragmented.
It's always interrupted. There's always delays because people are trying to find data to make these decisions. They were able to streamline that entire process using BIM 360 with some of their internal applications. Look at what the result was. $11 million shaved off the construction estimate and three months shaved off of the schedule estimate. Let's just pause here for a second. When's the last time any of you saw a construction project that was coming in under budget and ahead of schedule? Okay, not many people are nodding. Here's a second example. Balfour Beatty. Balfour Beatty is in the U.K., and they're a general contractor. What they are doing is they are working on a renovation project for the London Olympic Stadium. They are using BIM 360 to literally digitize their construction processes on the site.
Here's what they've been able to accomplish with BIM 360. You do progress inspections, right? When somebody finishes a room, you have to go do a progress inspection. That went from three hours to one hour. Then you have to log issues. Issues went from 10 minutes to one minute. Some of you might say, why is that a big deal? Because in a project of this size, there are hundreds, if not thousands, of issues like this that are logged. Going from 10 minutes to one minute is a big saving. They also have to do weekly quality checks. They have a lot of staff on site. That went from six hours to 30 minutes. Balfour Beatty is actually using BIM 360 on all 70 of their projects right now. What does that add up to?
That's 14,000 hours that they believe they have saved because of BIM 360. I think that's an amazing return on investment. By the way, it's not just JE Dunn and Balfour Beatty that are benefiting from BIM 360. Take a look at this chart. This is BIM 360 subscriptions over the last three years. We've had 10x the growth. This is showing you how successful BIM 360 is when it's being used on construction sites. You get the same kind of benefits that JE Dunn got and that Balfour Beatty got. What else are we doing? We are trying to simplify the portfolio. Andrew talked to you about the fact that we're going to get to a tighter set of applications.
We've done a lot of acquisitions, and we've built a lot of tools, and you can see that here for our BIM 360 offering. Our plan is to go from eight subscriptions to four subscriptions. You're going to have BIM 360 Docs for collaboration. You're going to have BIM 360 Design, BIM 360 Plan, and BIM 360 Build. They are all going to be tightly integrated with the building information model. The building information model is going to be the currency of construction in the future. What else does this mean? This means that you're going to see the number of subscriptions go down, but you're going to see ARPS go up over time for these cloud offerings. Why is Autodesk positioned to win in construction in the long term?
Well, to reiterate, we are the only ones that are building this cloud platform that is the single source of truth for design, plan, build, and operate, all anchored to the building information model. The majority of the world's BIM data is created and edited in our tools, and everybody knows Revit. Revit is a leading BIM design tool. We're also building a network of third-party application developers that are feeding data to make this building information model more robust. What is that doing? This is allowing Autodesk to deliver analytics and insights to construction firms. Everybody knows that construction firms want to mitigate their risk, and I want to protect my construction margins. Imagine being able to do predictive analysis on your construction project in the future.
Imagine if you knew that you had built in the past on a specific location at this certain time of year, and you knew you experienced 48% delays on that project. What would that do for your scheduling? Imagine if you were able to project price increases on materials. What would that do with your costing estimates? How about if you had an aggressive schedule and you know that one of your subcontractors is always late. What's that going to do to the schedule that you're proposing? We're able to do this because look at the amount of data that we have in this platform. This data is growing every single day. This is the kind of value that we are providing to construction customers. By the way, that's not the only unique advantage that Autodesk has.
We have 93% of the top Engineering News-Record global contractors using our BIM design tools. What does that mean? It's a natural evolution for these companies to want to work with us on our new construction platform. You heard Andrew talk about industrialized construction. Here's the way I look at that. Think of a construction site as an open-air factory, and you can do prefabrication off-site. You bring components to the site, and you assemble them. You bring new advanced manufacturing techniques like robotics or 3D printing to the site. Imagine what that's going to do for the construction site of the future. As Andrew said, Autodesk is an expert in manufacturing processes and technology. Autodesk is the leader in BIM.
We are the only ones creating that end-to-end cloud platform for the entire life cycle anchored to the building information model, and we are experts in manufacturing to help the construction organizations move to industrialized construction in the future. Only Autodesk has all three of those things. Our possible process to give you a feel for the CNC machine. What are we talking about? A cloud platform that is automating design and manufacturing processes with native generative technology. What does that do? You get thousands and thousands of unique products made to order to meet the demands of the new population. This is what's creating that new opportunity for Autodesk. That's an $11 billion manufacturing TAM.
Last year, we talked to you about the $16 billion TAM in product design and engineering, and now we have this additional TAM of $11 billion in manufacturing that we'll be pursuing. By the way, we're not the only ones who are seeing these trends in the marketplace. Over $32 billion of new funds were announced for hardware startups in 2016. Since 2016. By the way, these startups recognize the need for the scale and for the scope. I'd like to introduce you to one of those companies that is a startup that's actually using our technology. Briggs Automotive Company is based in the U.K., and they're a new hardware startup for automotive. They're bringing a street-legal race car to market. 30 people in their company. They're using new generative design and manufacturing solutions to do a few things.
Number one, they cut the time to market in half of what it takes to get a conventional car to market. In addition, look what they were able to do. With generative design and manufacturing, they were able to lightweight many of the components of a car. With the same engine capacity, these cars go much faster. This has been unheard of. In addition to hardware startups, we also have big manufacturing companies that use our technology. Greenheck Fans is based in Wisconsin, and they make industrial fans. Using our solution, this is what they were able to accomplish. This is pretty remarkable. They process 1 million pounds of sheet metal products per month, and every one of those products is unique. Every one is unique. Think about that. That is scale and scope, the future of manufacturing processes.
We're not only winning with these commercial companies, we're actually winning the hearts and minds of the next generation of designers and engineers, students. That's creating this flywheel of momentum. I want to show you how this works. First of all, we will get a big, large enterprise like, say, GKN Aerospace, and they're going to fund a project with students at the University of Warwick. The project that they funded was they said, "Let's see if we can build a drone that can help find lost hikers." What happens? The students build the drone with Fusion 360, and they fall in love with Fusion. Then what happens? They graduate, and they go to work, for instance, at a job shop. When they go to that job shop, they bring their love of Fusion to that company.
In this case, Ben Bingham has brought Fusion 360 to more than two employers in his career so far. Then what happens? A job shop like Saunders has this huge community because they work with big vendors like Tormach, and he has this social media following of probably over 800,000. Then what happens? Go to the next illustration, please. They're wanting to build up the suspense. Okay, back up just one, please. What happens is that you then get job shops that are engineering service providers. What do they do? They use generative design and manufacturing to create things like this manifold. By the way, this manifold used to be 15 parts. Now it's one part. It's 80% lighter and 20% cheaper. Then what happens? You get an engineering service provider that provides to new hardware startups like Modbot.
Modbot is a startup here in San Francisco, and they are revolutionizing the sixth axis robot market. What happens? You get big companies like Siemens who buy this new technology because they want to automate and do new things on the manufacturing floor. Big companies like that are the ones that fund student products. What we get is this flywheel of momentum, building the community of the future for manufacturing. By the way, we know this is happening because if you take a look at the Fusion monthly active users, look at this acceleration. Our students like to go on winter break. They like to go on summer break. These dips are the fact that they're actually not using Fusion on the ski slopes. What you will see is that we have dedicated customers.
See, we have students and we have commercial customers and startups that are dedicated to using Fusion 360. The other thing I want to point out to you is we are accelerating and compressing the timeline to add every new 50,000 MAUs for Fusion 360. We went from seven months to five months to three months. What else are we doing? We are trying to streamline our Fusion 360 offering for customers to make it easy for them to buy and easy for them to implement. Again, we're going to go from nine subscriptions down to three subscriptions. What you're going to see is we'll have Fusion 360, Fusion Lifecycle, Fusion Production, again, all of those built on a cloud platform that is automating design and manufacturing processes with native generative algorithms. Again, what are you going to see?
You're going to see less subscriptions, but over time, you're going to see ARPS go up. Why is Autodesk going to win in manufacturing in the long term? Because again, let me point out, we are the only ones building this cloud-based platform to automate design and manufacturing processes with native generative technology. What we are doing is we are feeding product models into this platform, and it's coming from our tools as well as other tools in the marketplace. We have a network of third-party developers that are feeding product model data into this platform. What does that allow us to do? That allows us to provide analytics and insights to our manufacturing customers. Manufacturing customers can use this to do predictive analysis.
They can add new aftermarket services for their customers, or they take the performance data of their product from out in the field, and they understand, how does that change my design process or my manufacturing process? Autodesk has a seven-year head start on all of our competition with generative technology. We started on this seven years ago. Autodesk is the only one that is building that cloud platform for design and make processes with generative technology. We are winning the hearts and minds of the next generation of designers and engineers, students. They are the ones that are creating that flywheel of opportunity and community, building the future community for manufacturing. Only Autodesk has all three of these things.
I want to thank you for your time, I'm going to introduce our next speaker, which is Scott Reese, he's going to go into more detail on the new solutions for construction and manufacturing. With that, I'll turn it over to Scott.
Thank you, Lisa. Good morning. Thank you all for coming. My teams are responsible for building a lot of the products that you see here at Autodesk. You've already heard a lot about the opportunities that we see in both construction and in manufacturing. What I wanted to do is to spend a little bit of time giving you some perspective as to what we're doing from a product perspective to realize those opportunities. If you think back over our 35-year history, where we've spent a lot of our time is in helping our customers design the world around us. If you get to know our customers, and many of you know them well, that's not where their work stops. They have to bring those designs to become a physical reality.
Our opportunity is to converge this notion of designing something with the process of actually making it. That's actually how we see this big opportunity ahead of us. Let's talk about the big opportunity in construction, the $10 billion growth opportunity that Lisa talked about. This is an industry riddled with inefficiencies. It's messy. There's a lot of heroics. It's also an industry being met with unprecedented levels of urban growth. There's a lot of opportunity for growth for every one of our customers, they have to work differently. Before we get too far ahead, let's remind ourselves of the process. What does it take to bring a building into the world? All starts out with an owner. They have an idea for a project of what it is that they're trying to accomplish. They sit down with a design team, typically an architecture firm.
They flesh out the next level of details. From there, they go into what's called the pre-construction phase, the planning phase, where they're planning for the cost, getting an idea of the rough schedule. Once they have some level of agreement there, they move into the physical construction of the building out on the job site. Once the building is erected, the owner takes over. They occupy, they operate the building throughout its useful life. When we talk about construction, we use it as a broad, sweeping term. We're talking about everything in the middle. There's a lot packed in there, we're going to peel that apart a bit and talk to you about what we're doing to realize those opportunities. The thing that's different about Autodesk, we can do more than just address this piece in the middle.
All of you know us as the BIM leader. It's indisputable that we are the leader in building information modeling. Our customers are creating an enormous amount of BIM data today. We see the opportunity to help those very customers leverage that data to drive the entire construction process, something only Autodesk can do. We'll talk about that as well. Let's start off by going out to the job site and talking about what we're doing there. Our strategy here is all about digitizing the job site. It's a little weird for me to stand up here in the year 2018 and talk about digitizing anything. I mean, just look around the room. Everything in our lives is digitized. Remember, the construction industry has had some of the lowest IT investment for decades. There's a long way to go. That's changing.
The appetite is growing, and for good reason. There's credible data out there that shows that the construction industry stands to save over $1 trillion over the coming 10 years by simply digitizing their operations. The business case for digitizing the construction site is crystal clear, and that's exactly what we're focused on with BIM 360. BIM 360 is where we tackle a lot of the here-and-now type problems that happen out on any job site that you go to. A lot of these are the tedious, mundane, day-to-day type problems that our customers deal with. These are the very problems that are critically important to manage if we're going to get our arms wrapped around efficiency and drive greater productivity gains. What BIM 360 does is it helps them take the bigger construction project.
They break it down into smaller and smaller tasks where they can manage the contractors on the job. They can manage the build schedule. They can track quality issues across the entire project, not just a piece of it. Ultimately, they can know their cost to complete that project at any step along the way. It gets even more interesting. That information is digital, imagine what can happen when we marry that digital information up with machine learning, like the engine that we have in our BIM 360 Insight product. That's exactly what we're doing. We're piloting this product actively with several customers today, including 16 of the companies on the popular ENR 100 list.
By working with these companies, the system has already processed over 225 million individual data points, and 30 million of those alone were problems, real problems that happened out on a job site that caused some kind of disruption. By working with these customers and processing all of this information, what are we doing? We're training our machine learning algorithms to understand every step in the construction process, understand which steps lead to problems, understand the steps that remedy those problems, now fast-forward and think about what we're going to be able to do. With BIM 360, we're going to be able to predict and ultimately prevent a lot of those problems from ever occurring. Talk about productivity. That's the type of productivity gain that it's going to take to help our customers realize the opportunities they have in construction.
Something that's unique about Autodesk in construction, we're not a point tool solution. We could do so much more. There's no one that has the potential to accomplish the breadth of problems that Autodesk does. In fact, we talked about how our customers are creating an enormous amount of BIM data already today, and we're focused on helping them leverage that BIM data to drive the planning process for construction. If you look at BIM 360 Design, what we're doing is we're connecting that BIM data to the cloud. By doing that, we're naturally driving the planning process. There's no guessing. They're using the actual BIM data. BIM 360 Design starts to bridge workflows, but not just for the architects and the designers that we've historically worked with.
We're bringing in the general contractors, the engineers, the trades workers on the job, all into one common data environment, giving everyone access and control of the critical project data, something they've never had before. Our customers are already today standardizing on our authoring tools. With BIM 360 Design, they now have the common data environment they need to deliver those projects successfully. In this past year, we've added hundreds of new capabilities into BIM 360. It's a big investment for us. It's getting better and better. We believe connecting the BIM data to the construction process is only the beginning. If you think about our breadth and our reach into our customer base and into these industries, we believe we have an even bigger opportunity to drive the entire end-to-end AEC process.
In fact, we have work already in flight that's going to show up in BIM 360 later this year that will help the designers, the construction professionals, and get this, the manufacturing professionals coordinate across all three disciplines. Why is it important to bring the manufacturing professionals into this BIM process, into the construction process? It's important because we believe that is an essential element of enabling what we see coming next, which is the convergence of construction and manufacturing. If you think of those processes, this is what we talked about earlier. This is the industrialization of construction. It makes sense. You visit any construction site that you see, and you look at the types of problems, the things that these customers are struggling with. They're the same types of problems that manufacturing companies have perfected over 20 years.
The job site of the future is going to become increasingly automated. It's going to start to resemble more and more of what you see today in a factory. This isn't some crazy idea that we just sat around here and dreamt up. Our customers are thinking this way today. They're starting to drive these processes, and we have work underway to help enable these workflows. Let's take a look at a solution that we're working on that takes the building design that our customers are already creating and breaks the building model down into panels. It looks for unique panels and understands all of the elements that have to go onto that panel. It feeds that information into Autodesk Inventor to create fabrication models and then feeds it into Fusion Production to manage the fabrication process.
These panels are manufactured in a factory off-site with a high degree of efficiency, high degree of quality, and a high degree of predictability. These are adjectives historically not associated with anything to do with construction. Autodesk is converging construction and manufacturing processes, the industrialization of construction. It's clear the construction industry is going through this radical transformation. It's undeniable. Autodesk is best positioned to continue making that happen. We talked about a few different elements. We talked about what we're doing to help our customers with their here-and-now type problems out on the job site with BIM 360, digitizing the construction site. We're doing it differently. We're not a point tool solution. We're taking the BIM data that they're creating already today and putting that at the center of the experience.
We're not tackling part of the problem, we're tackling the problem that they have, the end-to-end AEC workflow. We are helping them realize the industrialization of construction by converging construction and manufacturing processes, something only Autodesk can do. There's no company on the planet who has our depth of product portfolio and expertise in both manufacturing and AEC. Only Autodesk can make this happen. It's a big opportunity, a big area of investment for us. Like Andrew said, we have an equally exciting opportunity in this $11 billion growth opportunity in manufacturing. If you think of manufacturing, it's being disrupted as well, but in slightly different ways. The means of production that manufacturing companies have used for decades, they're all being disrupted, challenged, replaced.
You and I as consumers, our expectations, our requirements of manufactured goods, they've all shifted, the old ways of working for these manufacturing companies are simply inadequate for them to remain competitive. Autodesk's vision for the future of making things will help these manufacturing companies thrive in this new, highly iterative manufacturing economy. We've built a solution to help them converge the process of designing something with the process of making it. If we think back over the past 3 decades, one thing you've seen from us is we've acquired many companies that have given us capabilities that span everything from conceptual design all the way through advanced manufacturing. These acquisitions have given us a deep, deep technology moat that's helped us build a highly differentiated solution to help these manufacturing companies work differently and thrive in the future. That solution is called Fusion 360.
As you watch this, what you're going to see is world-class conceptual design tools coming together with 3D parametric modeling, including all of the simulation capabilities, all of the deep manufacturing capabilities that are required to manufacture these products, all in one integrated environment on top of a single data model. The first of its kind, only from Autodesk. This is the convergence of designing and manufacturing. Like Lisa talked about, we're seeing rapid adoption. In fact, we're coming up on having a quarter of a million users who use Fusion 360 to do their work. When you peel that number back a little bit and look at the demographics inside of there, I find a couple of interesting things. One, nearly half of those users are coming from competitive systems that are simply not keeping up. They're finding a better way to work with Fusion 360.
They're seeing the change, we're responding. The other demographic that's interesting is how we're winning the hearts and minds of the next generation of engineers, the students. Which brings me back to the story that Lisa started off with. These students at Warwick University in the U.K., they had an idea that they were passionate about. Building a drone that can help in difficult rescue situations, deliver critical supplies before the emergency responders are able to reach them. They knew that they had a challenging process. None of the students had the same schedule. They all needed to be able to work at different times of the day and night. They were distributed all over the place. They're not going to be working from one location. They had a variety of skill sets. They needed one system that works for all of them.
They started out on this project trying to use traditional tools. In this case, they tried to use SOLIDWORKS. It didn't work the way that they think. They found a better way of working. They found Fusion 360, and they delivered amazing work on this project. These kids are going to change the world, and they're going to do it with Autodesk and with Fusion 360. Everything I just showed you, everything in that video, everything I talked about with the Warwick students, that is Fusion today. It gets even more exciting when you think about where we're taking Fusion next. Now that we have all of those capabilities built into one integrated environment, all working on one common data model, we're working hard to integrate those with our generative algorithms to drive the entire end-to-end process in an automated way.
What you're seeing in this video is that I'm not using the system to capture an idea that's already in my head. With this next generation of Fusion 360, I'm framing the problem. I'm telling it where I need these points to be connected. I'm laying out some preferences. Maybe I have some material preferences, or I have some weight or safety requirements. Maybe I have a favorite manufacturing technique. I'm not telling the computer the solution. I'm telling it the problem. Then I'm letting the computer compute and come back and suggest to me the actual best answer, not my best guess at what the answer might be. When I talk about this being the future of making things, we're not talking about years into the future. Everything you're seeing in this video is actively in development as we speak.
This is the convergence of designing and manufacturing. This is the future of making things, and it's only Autodesk who's delivering it. In the mechanical space, a lot of our users are using Autodesk Inventor. We have a huge install base of highly successful customers, and we've built a bridge for them, a bridge called AnyCAD that helps establish an associative relationship between their work and the data that they create in Autodesk Inventor in the future of making things with Fusion 360. Our customers can have the best of both worlds and never miss a beat. The best thing, AnyCAD doesn't just work with Inventor, it works with all 3D parametric modelers. We've built a bridge for the entire manufacturing industry to find their way to the future of making things only from Autodesk. We've talked about the disruption in manufacturing.
Equally disruptive as what we're seeing in construction, but in a slightly different way. We have the vision for the future of making things. We have a huge technology leadership position, a huge technology moat that gives us a unique opportunity to deliver on it. We're marrying all of that technology up with real generative algorithms to automate the entire end-to-end process, and we are converging the process of designing something with the process of manufacturing it. Only from Autodesk. Equally important to what we're building to help these customers realize their potential is how we're going about building these things. Autodesk Forge is our common cloud platform that brings us 3 distinct competitive advantages for years to come. First, this is how we scale our R&D investment.
Autodesk Forge is how we accelerate the development and delivery of these important solutions for construction and manufacturing with BIM 360 and Fusion 360. The value of the Forge platform goes well beyond our internal uses. The 2nd distinct advantage, it helps us get closer to our customers. Our customers have all kinds of gaps in their workflows, in particular between their business systems and their engineering systems. This is a company named Royal BAM Group. They're using the Forge platform to fill gaps in those workflows. This is a custom configurator they use the Forge APIs to build that gives their designers the ability to select parts to include in their design that can be manufactured in a factory off-site with all those efficiencies and quality gains that we talked about, and then delivered onto the job site for final assembly.
The Royal BAM Group is realizing the promise of industrialized construction. They're realizing it with Autodesk. They're realizing it with Autodesk Forge. The BAM Group's not alone. Last year, 25 of our named accounts, in addition to the applications that they bought from us, they included the Forge platform because they know the promise that it brings of bridging gaps in their internal business and engineering systems. It's a trend that we're going to see for years to come. In fact, last year, we did an 8-figure EBA deal where the customer came back and said, "The reason we bought from you, the deciding factor to go with Autodesk was Autodesk Forge.
A big differentiator for us." The 3rd distinct advantage that our cloud platform brings to us is something not necessarily new, but it brings it in a slightly different way, which is the ability to build a robust third-party ecosystem. If you think about the breadth and depth of our vision for the future of making things, both in manufacturing and construction, there's a lot of room for our partners to play a role, and that's exactly what we're doing. Let's look at GTP Services. GTP's expertise is all about piping and HVAC systems. It's not software development. GTP was able to use the Forge platform to bring to market an application that bridges design and fabrication for piping and HVAC.
They did it in less time than they were ever able to in the past, and they did it at lower cost, driving down their internal development costs. In the manufacturing space, computational fluid dynamics, some of the most sophisticated simulation capabilities of any. A company called CCTech used the Forge platform to make this sophisticated technology available to the masses, something historically available only to a few. They used the Forge platform and our application framework to deliver this technology via web and mobile devices, making it available to the masses, as well as driving down their internal development costs. GTP and CCTech, they're not alone. We have over 8,000 developers actively building their applications on top of the Forge platform today. You're going to see so many of these come out.
The Future of Making Things platform is going to be the most robust solution across manufacturing and construction, and our partners will play a big role. We've covered a lot of ground. If you think about what we're doing in construction, we're helping our customers realize the opportunity they have today, tackle the problems they have in the here and now with BIM 360, but we're doing it in a different way, a way that only Autodesk can. We're putting the BIM data that they're already creating at the center of that experience. In manufacturing, we're delivering on the Future of Making Things, a highly differentiated vision that helps these customers respond to the disruption they're going through. We're converging the process of designing something and manufacturing it, and we're doing it all in a very different way.
We've built Forge, the cloud platform that is accelerating efforts all across the board. As you can see, when it comes to construction and manufacturing, we're poised for growth. We're poised for growth today by addressing the problems that our customers have in the here and now, but we're also poised for growth for years to come. Thanks for coming.
All right. Hope you enjoyed the first half there. We're going to take about a 20-minute break. We'll come back at 10 o'clock, and then we'll get on with Steve Blum and Scott Herren in the Q&A session. Thanks.
Ladies and gentlemen, please welcome Autodesk Senior Vice President, Worldwide Field Operations, Steve Blum.
Well, thank you. The front row. I love the front row. Welcome back, everybody. Welcome back from your break. Find your seats. Make yourselves comfortable. Today, I'm going to step you through all the things we've been doing to drive ARR growth. As we've been talking with you, ARR is the single most important metric, and we've been doing everything to align ourselves to be prepared to drive the ARR growth we need for many years to come. Since the last time we got together, I wanted to give you an update on how we think about our customers and our customer segmentation, and specifically about our go-to-market approaches from both sales as well as customer success. For those of you who were here before, you'll notice there's a new box here. I've always talked about our named accounts, talked about the territory.
I've talked about hub sales, renaming hub sales digital sales. I'm going to go into some details about why that is and what's composed there. I've added a new segment called the mid-market. The mid-market is where we're taking best practices from our named accounts, and we're applying them to the next level of large accounts. We believe, by focusing on them with account-based selling and marketing, we can drive growth. If you think about the way I've got this structured now, named accounts and digital sales is where we're taking a very direct approach. We're engaging directly with those end customers. We're even transacting business directly with those end customers within named accounts and digital sales. Think about the mid-market and territory as where we're working through our partners to drive scale.
Even though we'll have direct sellers focusing on accounts in the mid-market, as an example, transactions will still be indirectly processed through our partners. We've also changed the way we have aligned our sales force, both from an inside perspective as well as our outside sellers. We've now moved to a position where most of our outside or field-based sellers are focused on account-based selling practices, either with the named accounts or the mid-market. By the way, the mid-market is many more accounts in this space than named, but again, we're focusing on those accounts. Those field reps are not assigned geographies. They're assigned specific accounts, and they spend all of their time working hand-in-hand with a partner to drive expansion and growth within those accounts.
The territory and digital sales are geographically aligned. The way we go about working with partners in the territory is through inside sales. Inside sales was something that didn't even exist 4 or 5 years ago, and now it's actually one of the biggest parts of our go-to-market. Think about those ways that we go to market. I'm going to show you now how we're going to drive growth from existing customers, how we're going to continue to drive growth through new customer acquisition. Of course, in a subscription and consumption model, it's critical to ARR growth to focus on adoption and renewals. I'll give you an overall update on our partner strategy. I know how we work with our partners is always an area of interest, and I'll give you some updates as to what's going on there.
Let's start with existing accounts. What I want to do first is give you an update on how we're working with our named accounts and how we're continuing to progress in our engagement with named accounts. This is an area I'm really pleased with as far as how we've changed the relationship with our customers, focusing on developing strategic relationships. We're leading the way with Enterprise Business Agreements. I wanted to give you a reminder of when I talk about EBAs or Enterprise Business Agreements, what it includes. All right? Because many folks think it just includes a licensing model. It's a consumption-based model. We make our entire portfolio available to a customer who's in an EBA, so they can access anything they want, anytime they want, anywhere around the world for any user that they have within their enterprise, and it's through a tokenized mechanism.
EBAs are much more than that. We focus a lot of our own direct resources on ensuring that we are helping customers achieve their most important business outcomes, and we do that by assigning customer success managers to these accounts. We have enterprise priority support. We have strategic consulting that's applied to help them actually drive implementations overall. When you combine these things together, that's what we mean when we say EBAs. All right, here are some of the results we're continuing to drive through our EBA focus with named accounts. When we move customers from maintenance or our old MFlex model, that was the old EBA we had from years ago, to a Token Flex model, the new EBA, on average, we're getting about 2.5 times the number of users.
We're driving enterprise expansion and scale and growth within these accounts by moving them to an Enterprise Business Agreement. Over the last 3 years, we've had 55% compound annual growth in our EBA ARR. That's outpacing the growth of other parts of our business. As of the end of FY 2018, the fiscal year we just ended, approximately 45% of our named accounts have moved to EBAs. If you remember last time we were together, we were about 25%. We're making good progress and we have a lot more to go do. There's plenty of opportunity for expansion and growth in this part of the business. As I mentioned, strategic relationships and changing the engagement process with our customers is really the ultimate goal. The EBA, from a licensing perspective, is one of the ways of accomplishing that.
I want to give you an example of how we change the way we work with customers in this space. I want to give you an example about AECOM. You may or may not be familiar with who AECOM is. They're a global engineering design firm. They're in the Fortune 500. Both Lisa and Scott talked about the ENR. AECOM is the number one design firm in the ENR. They're a very, very large global firm. We've been working with them for quite some time. In fact, they renewed their Enterprise Business Agreement with us back in FY 2018. The relationship has progressed to the point now where we're actually joining together with their pursuit team to help them win new business. We actually work together with them to put together numerous proposals that they actually were able to win.
An example of that is with the Naval Facilities Engineering Command, NAVFAC, where AECOM won an $85 million project for environmental remediation. We actually helped them put that proposal together. We helped them be positioned to win that contract. That's the kind of strategic relationship we focus on establishing with our named accounts. We also help our customers implement their projects and to gain the efficiencies that Lisa and Scott were talking about. I want to give you two examples. This may be one that some of you from New York are familiar with, the Barclays Center, really beautiful arena in New York City that AECOM was the construction and designer of. We helped by applying adoption specialists. We helped them implement BIM 360 Plan and BIM 360 Build to be able to reduce their overall time by 30%. That's for the designers and for their contractors.
They also were able to reduce file sharing time by 70%. In less than eight months, they were able to install over $350 million worth of project work by leveraging the tools. We help them gain efficiencies, we help them reduce costs, we help them deliver on their project timelines by focusing on the delivery and the outcomes of that project. Another really interesting example is the Olderfleet Building. This is a renovation project in Melbourne, Australia. They were using Team around the world. They needed help with global collaboration. Again, we applied adoption specialists that helped them implement BIM 360 Design. As a result of that, they were able to remove or improve 20% the production time of getting this project done.
The project required lots of design reviews. They were able to reduce the preparation time for those design reviews by eight hours per review. Again, we help them achieve their overall business outcomes. Again, these are the types of ways we're engaging with our named accounts to ensure that we're building strategic relationships. We've had this success continuing to grow over the last several years. As a result, we are expanding this account-based selling approach into the mid-market. That was the new segment that I highlighted. We're actually turning all of those field sellers into account-based sellers. They'll work together with the partners, like I said. A sales rep in the mid-market will have an assigned set of accounts. He or she will put together account plans for those accounts.
We'll work jointly with the partner sales reps to make sure that we're on the same page, that we're executing on a common strategy together. What's critical here, though, because there's many more accounts in this part of our space, is we're leveraging account-based marketing best practices as well in order to drive more opportunities to bring more opportunities into our pipeline. We're leveraging customized content that's been created by Lisa's team that can then be targeted specifically to particular accounts in particular industries or even industry subsegments, and even it can be segmented down to the persona. We're getting the right messages to the right people that matter to them, that focus on the disruptions that they're managing in their business.
When they respond, we can retarget them, we can bring them to more customized content that will help them ultimately see how we are bringing unique value to help them solve their problems. As I said, this will end up helping us drive many more opportunities into our pipeline. The way this works really comes in three different flavors. We have customized content that's available. A sales rep can do a self-service process. If I'm a sales rep on a particular mid-market account and I know who the key people are in that account, I can now start sending the right messages to the different folks, the different personas within the account to engage with them, and then to reengage with them and retarget them overall to help them understand how we can help them with their business issues.
We also can take a programmatic approach, this is where we actually will look at all of the accounts we have in the mid-market. We'll pick the accounts that all have a similar industry segmentation, we start sending them messages specific to their business needs. Again, it's persona specific. Someone that's in the design space will get a different message than someone who's in the make space. This, of course, then allows us to retarget, to re-engage with them, ultimately it builds more opportunities in the pipeline. As we've been doing in named accounts, as we'll always do in account-based selling, we can take that high-touch approach. This is where we actually engage in large events with a particular customer. Maybe we put together a joint customer success planning process focused on a particular project they're going after.
These are the ways we're going to be driving new demand within the mid-market, leveraging account-based selling. Now we are also driving growth within existing accounts by focusing on collections, as well as moving our maintenance customers to subscription. We're having great results. We had great results in FY 2018. In fact, we grew our collection subscriptions by over 60% through FY 2018. Since we introduced our maintenance to subscription program, the program to move our maintenance customers to subscription, over 30% of the maintenance renewals that came due actually moved from maintenance to subscription, which was a really good result the first year out. What's very interesting about that is of the seats that moved from maintenance to subscription that were eligible to be upgraded or upsold from an individual product on maintenance to a collection, approximately one-third of those seats moved to collections.
Andrew and Scott talked about this on the earnings call. How you should read this is, by moving these customers from maintenance to subscription and having them take a larger percentage of collections, we're driving growth in ARR, our number one goal. We're driving ARPS up, we actually end up in many times with fewer subscriptions. They gave an example about having half the number of subscriptions when a customer moved from maintenance to subscription, 10% growth in ARR. These are the types of things we're seeing to drive incremental growth in ARR by moving our maintenance customers to subscription. Let me talk about our continued focus on new customer acquisition. The best places for new customer acquisition come either from our digital sales approach, from focusing on our legacy customers.
Legacy customers, as a reminder, are folks that purchase perpetual licenses but don't have those perpetual licenses on an active maintenance contract. They may never have attached maintenance, or they may have attached maintenance at one point in time and then did not renew. That's our legacy base. Then we have non-paying users. These are the folks that use our software but never paid anything for the use of software. Think of this as license compliance. I'm going to go in each one of these sections, let me start first with digital sales. Last time we were together, I talked about our hub sales approach. Digital sales is an expansion of hub sales. It's leveraging our digital processes. Andrew talked about how we're digitizing the company. We've built e-stores that are focused around the world. Sorry, I thought that had moved over.
We have e-stores around the world now. We've built an online Autodesk Knowledge Network, millions of items that are digitally available to customers to access when they need it. We have chat capabilities, we have virtual agents overall. We have webinars. We have trial downloads. All of these things have been digitized to drive scale overall. We marry them up then with our inside sales approach. This is where we do have our inside hubs where we co-locate inside sellers, technical specialists, customer support folks, marketing and nurturing folks, and even inside customer success people. The combination of these two things is digital sales. In FY 2018, we had great results through our e-store. We exited the year with over $140 million of ARR from our e-store. We had 65% growth in ARR from the e-store. We had 50% growth in subscriptions from the e-store.
Here's something that you may find really interesting. We've talked about this in the past. 50% of all the new subscriptions that we get on our e-store are from net new customers. This is how we're acquiring new customers and bringing them into the Autodesk ecosystem. Let me move on to the legacy base. These are the folks that purchase perpetual licenses. They're not on maintenance anymore. We gave you some statistics about the size of this base. I'm going to give you updates on that. We talked about it two different ways. If you think about the last five years back of releases, the last five releases, the current base is 2.6 million licenses. Of that base, there's 1.2 million active users. These are folks that are using the software, even though it's not on maintenance.
Five years back, 1.2 million active users of the software. If we go beyond five years back, we have 5 million more licenses out there, perpetual licenses not on maintenance. We estimate that the active users of that base are about 800,000. If you add up the active base total from five releases back, plus the estimated amount from more than five releases back, we believe there are about 2 million active users of licenses that are not on maintenance. This is the target base that we're going after. If you recall last year, Andrew actually showed you these same numbers, and we had about 2.2 million in the base. We are actually converting legacy customers to become subscribers. How we do that and why they decide to go is basically, first, we offer promotions.
We've been offering promotions over the last several years in this space. We're going to continue to do that. We have a promotion actively running here in Q1 focused on that legacy base. The value or the discounts associated with those things have been going down, and Scott's going to actually talk more about that approach when he comes up next. Why will a customer actually select to move from an old maintenance or old perpetual license to subscription? They want to get access to the latest features. We continue to add more features, more capabilities into our offerings. We're helping them drive better collaboration through the use of the tools. There's a better customer experience in subscription compared to the older perpetual licenses.
They're given more flexibility in controlling their environment and managing their users, managing their purchases, gaining insights on how the software is being used. Product subscription also includes our basic support, which is not available to maintenance or to perpetual licenses, not on maintenance. These are the reasons that these folks are moving. There's also one more really important reason. Every year that clicks by, every release that falls one more release behind is becoming more out of date. These folks are doing work with other companies, and they have ecosystems, and they have to share data, and they're falling further and further behind the companies they're working with to share data. Sometimes they have to actually move to the latest versions just to be able to collaborate with the folks in their supply chain or their ecosystem. All right.
Let's talk about that third group. This is the group of non-paying users. These are the folks that haven't paid us at all for the use of the licenses. Again, this is where license compliance fits in overall. The size of this base is 12 million users around the world that are using our software, not paying us anything for that use. 4 million of those users are in mature markets. When we take a look at that 4 million base of users, we have detection capabilities that give us an opportunity to identify just under 25% of them, meaning we know what companies they are, what people they are, where they're located overall. Which means that actually a little more than three-quarters of that 4 million, we don't have enough information to actually have outreach directly to those customers.
What do we do between the known users and the unknown users? We have numerous ways of approaching this. First and foremost, we're continuing to build out more detection capabilities so we can actually identify more of that base of unknown users and turning them into known users. When we know who they are, we then can help them run audits. We can do audits for them. They can run self-audits. We provide them tools for software asset management. If you're sitting there saying, "Why would they do that?" Many of these customers may not actually be aware that they're non-compliant. They may actually think they purchased even though they haven't. They want tools so that they can actually evaluate how they can become compliant, or at least evaluate if they're non-compliant overall.
When we know who those folks are, we can have outreaches to them from our inside sales teams. We can go through our telesales teams and make offers to them to help them become compliant. Something that I'm very excited about that's new, that has just rolled out, though, and it's going to give us access to the larger base, the full 4 million, is in-product messaging. We've now developed a capability that within the license, it can determine if it's an actual subscription that has been purchased or not. If it recognizes that it's not, it provides the user with a warning that it's not genuine Autodesk software. It gives the option to purchase or subscribe to that software, either by connecting to our e-store or contacting a partner or contacting us.
This will give us more range and scope into the larger base of users overall. We've just gotten started with this. The capability is available in AutoCAD in the U.S., but we do have plans to expand it into more of our volume-based products, as well as expanding it into other regions around the world. More to come on this, but something I'm very excited about. Let's talk about how we focus on driving ARR growth through adoption and driving, of course, renewals as the renewal base becomes a huge component of our ARR. We put together and now have formed a formalized customer success organization, and this team is focused on the entire customer success lifecycle.
When we land one of those new customers that I was just talking about, we need to make sure that they adopt, that they start using that software effectively right after the purchase. That gives us the opportunity to expand our position. They may want to buy more subscriptions, or maybe they bought a subscription for an individual product and they want to move to a higher value collection to get their job done. If we're doing these things properly, that renewal should actually become more of an automatic task, an administrative task, because the customer is effective in using the software and solving their business problems. We've been doing this, we've been taking this approach with our named accounts. It's part of the reason why we're having great results with named accounts. We have dedicated resources focused on those named accounts.
As I mentioned, we have customer success managers. We have dedicated enterprise priority support folks. We have strategic consulting folks that work directly with those accounts to ensure that they're driving the outcomes in their projects. We put adoption specialists in that go help them. Now, for all other accounts, the focus from our CSO organization is to ensure that we develop repeatable plays that we can scale. We want to scale them through our own inside sales teams, through our partners. We're going to be focusing on putting together repeatable best practices and enablement capabilities, packaged offerings that our partners can take to market to ensure they're focusing on driving customer success through adoption that ultimately leads to renewals. We continue to enhance our early warning system.
This is a system that actually evaluates which customers are most at risk of non-renewing, so we can start taking action as early in this lifecycle as possible to ensure they do renew at the very end. All of this comes together through driving digital efforts. Part of the digitizing the company approach that Andrew talked about is focusing on means of leveraging digital capabilities to truly drive scale that will have huge impacts on customer success and customer experience that will drive great adoption and ultimately the renewals that we want. I want to give you an example or two of what I'm talking about. If you actually go out on our website, you can get to our Autodesk Knowledge Network. This is actually the home screen of the Autodesk Knowledge Network.
There's several areas I just want to highlight for you and show you how we truly drive scale through digitization. If a customer has a question about either how to use a product, maybe they're actually even looking to find out how do I actually make a part. They can come to our site, and they can click on the support and learning, and there they will have access to millions of digitized white papers, best practices, and other things that we've solved for customers in the past. The we is either Autodesk, it's our partners, or it's even other customers. We create the digital content and make it available to those customers. To give you an idea of scale, we are currently getting about 7 million sessions focused in this part of our Autodesk Knowledge Network per month. 7 million per month. All right.
Sometimes folks want to actually reach out and ask for specific help. They want to leverage technology, but they actually want to connect with people to ask some questions and find out how they're doing things. We have the community section over here, and the community actually takes you to forums. Forums that provide customers the ability to help customers. They help each other solve problems. We also participate on the forums. Our partners participate on the forums. We're getting 4 million sessions per month through the forums. Combined between those two things, 11 million sessions per month are helping customers drive a great experience and ultimately focus on their success through digital capabilities. That's truly scale. Now, the part that gets very interesting is this part in the middle. It says customer service.
Customer service is where we take customers now to our Autodesk Virtual Agent, which is based upon AI and machine learning. We've talked about AVA. In fact, that's the image of AVA that people will see if they start chatting with AVA. Through learning, we can train AVA to resolve numerous repeatable questions or issues that come up. The more we get them, it's easier to train AVA to be able to answer those questions. We've gotten to the point right now in our training process where AVA's solving problems for 1,000 customers per day. What's really beneficial, though, is the resolution time as a result of using AVA has dropped the time to resolution so much that most of these issues are being resolved in less than 5 minutes.
Some of you probably haven't gone out and tried this, I'm going to give you an example of how we are using it in one particular example. What you can't see down below here, but customers contact us because they need prior release serial numbers. They need to get access to some prior releases. They contact us for that. The scale, this is a time scale, starts in September 2016 all the way to January of this calendar year. This is the cases along the side. For those that can't see, the cases go up to about 2,000 per month in some situations. We only had human agents for a long period of time as we started to train AVA. You can see as we've trained AVA now handles most of these cases overall.
The resolution time, the mean resolution time has gone from many hours, sometimes over 12 hours, to less than 5 minutes. Think about the customer experience, having to wait 12 hours, but now actually getting these issues resolved in 5 minutes or less. Think about the scale that we're driving. Think about the cost benefits to us that we no longer need people to handle many of these actions and activities. A really good outcome for our customers, and a really good outcome for Autodesk. Again, part of digitizing the company. All right. One other area I wanted to focus on in the area of adoption, this is going to tie into our partner strategy. In FY 2019, this new fiscal year, we made a fairly significant change in the partner framework.
In the past, in all the years I've been here, we've paid partners on landing new customers and renewing customers. For the very first time, we're now paying them not just on landing a customer or renewing the customer, but actually focusing on adoption activities. In FY 2019, partners can earn money by ensuring that when they sell subscriptions to their customers, that those subscriptions do 2 things. They activate the license within the first 30 days, and they run 1 adoption play. We've established adoption plays based upon the offering that's most appropriate to ensure that the customer is going to have the best experience and ultimately renew. We also know if customers don't activate in the first 30 days, they become more of a renewal risk. That's why we're focusing on those areas today.
We're going to be on a journey in this area with our partners overall. I expect that over time, we're going to move to more meaningful metrics focused on customer outcomes, true adoption metrics, perhaps even consumption down the road. More to come, but it's the very first phase. One more thing I do want to highlight. We believe that it takes the most amount of effort to land a new customer. We think it takes the second most amount of effort to drive adoption and effective use. If we do that, it takes the least amount of effort to renew a customer. You're going to see us changing the way we pay our partners to reward them on the activities that have the most resource requirement, the most time requirement.
Expect that over time, we'll pay the most on new, second most on adoption, and the least amount on renew. When Scott comes up, he's going to give you some more details about that, and especially what the impact is to ARR on the renew. All right. A few final things on the partner strategy overall. First, for those of you who may want to recheck on, okay, why do we even have partners? I want to give you a few metrics. First of all, we get scale. We get coverage. We get capacity. We have people representing us each and every day in over 170 countries around the world. For each quota-carrying sales rep that I have on the Autodesk payroll, we have 15 quota-carrying sales reps through our partner ecosystem.
That's the kind of scale and coverage that we need to drive the growth that we know is available to us. I wanted to give you an update on a number of partners, because we've been seeing partner consolidation, we've been driving consolidation through this business model transition overall. Right now, we're about 1,350 partners worldwide. If you remember, last time we were together, it was about 1,600. This is tracking to our expectations. Our best partners are in a better position now than ever before, and they're investing in the business going forward. One other big change we're making is we're moving more of the partner dollars from the front end, which is more of an entitlement on the sale, to the back end, the back end's tied to performance-based measures like the adoption incentive I just talked about.
In fact, the single largest back end incentive is tied to growth in annual contract value. They're giving growth targets that align with our growth targets. We use ACV because it's the best proxy of ARR. They're now going to need to ensure they're driving growth in ACV, annual contract value, which aligns with our ARR expectations, in order to earn that same amount of money. All right, one last piece. I know that it's always interesting to see the mix between direct business and indirect business, so I wanted to give you an update. The size of the pie represents the size of our business as it grows over time. FY 2018 is the year we just finished. I'm giving you an estimate for FY 2020, and then an estimate of where our terminal or future state will be.
You can see that the pie gets much bigger, as you should all be modeling out and as we are. You can also see you don't have to worry about measuring this. This is 50/50. All right. You don't have to take out your protractor and worry about that overall. I know some of you are going to try to measure that piece, and go right ahead. Here's the key point, and here's what we just told our partners out at One Team Conference. The amount of business, indirect and direct, so the direct business and the indirect business, they both grow. They're both much larger in FY 2020 and in the future state than they were in FY 2018. I'm telling our partners, "Don't focus on the percentages. Focus on the total dollars." They're getting bigger and there's fewer partners.
That means fewer partners are dividing up a larger piece of indirect business overall, and this is why our partners continue to invest and continue to grow, at least the ones that are aligned with where we're heading as a company. Let me summarize by saying our account-based approach to selling and marketing is going to be one of the best stimulants and drivers of growth for us within our existing accounts. Customer success is going to be critical here in the world of subscription and consumption. If we do our job properly, we focus on customer business outcomes and drive adoption, those renewals overall should become automatic, and that positions us to expand, and that's how we drive the growth like we're getting with our EBAs with named accounts. Customer acquisition will continue.
We're continuing to bring new customers into the ecosystem, and our digital sales approach should be the best way of doing that. It's leading the way. Our partner relationships are very strong. They continue to get stronger, and our partners are aligned with our strategy overall. When I net it all out, we've built an engine to drive ARR growth, not just through the end of FY 2020, but for many years and beyond. Thank you for your time. It is my pleasure now to introduce to all of you, Scott Herren.
Nice job. All right. Thanks, Steve. Steve touched on a lot of the themes that I'm going to touch on as we go through today and talk about the path, first of all, of how we get to fiscal 2020, and then what things look like beyond fiscal 2020. Let me jump in. First thing I just want to touch on is the outlook. There's no change in the outlook. I actually had a question at one of the breaks, did we change There's no change in guidance versus what we gave back at the beginning of the month in early March. The one thing I'll just point out on this slide is, if you remember, we gave guidance both ways. We gave it in 605, the old rev rec standard, and 606.
605 gave you a chance to compare it to how consensus had been laid in, right? Consensus hadn't yet reflected the 606. We do expect you to build your models based on 606. Since we're using modified retro, we'll actually report it both ways all four quarters of this year, but I'd like you all to adjust your models to 606. As we talk about the path to fiscal 2020, there's three key things you need to understand on how we get to fiscal 2020. First is ARPS and what's driving that longer term. The second is the growth in the core business. Andrew showed you the money slide, right? The slide that says how we'll get there. You saw how big the core was as a percentage of that ARR.
If you understand the core business, you understand how we get to fiscal 2020, so I'll talk about the volume drivers. The third is the $1.4 billion of free cash flow. I'll give you a slide that actually builds that out so that you can have some insight into it. Second part of the presentation, I'll focus on from fiscal 2021 through fiscal 2023. As a part of that, we clearly will be investing to go after some of the exciting new market opportunities that we heard from Lisa earlier. In construction, in manufacturing. That is a part of the growth story between those two years. The core business continues to grow as well. Many of the things that I'll talk about that'll grow the core business between here and fiscal 2020 will also grow beyond fiscal 2020.
All right, Andrew gave you this slide earlier, 25, six, and 18. Uptick in our expected compound annual growth rate, that's not just the to-go, right? That's from start to finish, from fiscal 2016 end to fiscal 2020 end. A better result on ARR, really driven by a better result in ARPS. Of course, the $1.4 billion, as you know, we're not allowed to quote a non-GAAP liquidity metric, like free cash flow on a per share basis. If you were doing the math, here's the way the math would work. Our basic shares outstanding at the end of Q4 was just short of 220 million, 218 and change. As we flip back to a profit, the diluted shares will come back into the share count.
If you divided that number by a share count that's between 220 and 225, that would get you to the $6 of free cash flow per share. The money slide, how we'll get there. What I've done is taken Andrew's slide and actually broken it out by year. The first three years you already had 2016 through 2018. You have a good sense of 2020. I just wanted to give you a full picture of how this looks. There's a couple of takeaways. Again, you see how big the core business is as a part of the ARR, which is the most important metric on here. I'll talk about ARPS, I'll do a whole section that dives down into ARPS.
The one thing that I think is also important to note is that obviously, the CAGR on subs is a little bit below what it had been historically, and we've talked about all the reasons for that on our Q3 and our Q4 earnings call. I think the thing to remember is the number of users is not going down. In other words, as we continue to add to the cloud base, we're adding users as we do that. We may not be adding them as fast as we had initially thought, but we're adding cloud users. When we see the consolidation that we talked about as people go from maintenance to subscription and they buy up to the collection, the example I gave on the call, a customer had 42 maintenance subs.
They converted to 20 collections at the time they moved over, 19 collections and one AutoCAD license. The number of subs went down from 42 to 20. The number of users stayed at, it was 20 before, it was 20 after. The ARR actually went up. More than 10% in that case. Just bear in mind, there's not user leakage happening as the subs growth comes down slightly below where we initially expected. Okay, I talked about the three things that I think you need to understand. Let's start with ARPS. There's four key drivers to our ARPS going forward that we've already seen have an effect on our ARPS, and that's why the ARPS was higher. It'll continue to be in effect through fiscal 2020, and frankly, even beyond fiscal 2020.
It's important to walk through this because I think there's a misperception in some people that the only reason ARPS is going up is because we have price increases. The bear case that goes behind that is, well, you can't continue to take prices up forever, or you'll annualize that price increase and there's no more growth in ARPS. Actually, price increases are not as important as the four factors that we've got on here. These are listed in order of importance, by the way. Growing the renewal base is the first. Increasing our direct sales mix. Steve just showed you the pie charts. Which, by the way, was a conceptual representation. If you notice the things, if you're going to start to measure those pies, just remember that's conceptual.
The product mix shift to collections, which we talked about a lot on the Q4 earnings call as a driver, and lower discount on promotions. Let's take those one at a time. What this represents is the base of renewal subs, so both maintenance and product subs, and you see how that base grows out through time. The important thing is that as that grows, Steve just talked about this as well, what we pay our partners for a renewal is about 10 points less than what we pay them for a new sale. As the renewal base grows, the yield to Autodesk grows at the same time. That's a big driver of ARPS for us. Second is the direct sales mix. I think what's important within this, we've talked about this going to kind of a 50/50 blend.
You can see where we are today and where we think we'll be in fiscal 2020. What's important is a lot of that growth, and again, don't measure this because this is conceptual, a lot of the growth in that direct piece is being driven by the digital sales that Steve just talked about. That's the eStore, and that's our inside sales teams. The reason that has a big effect on ARPS is most of those sell at very close to SRP. Again, the yield to us is significantly higher when we sell through our inside sales team or we sell through the eStore. That's a lot of what's driving the growth as our direct sales becomes a bigger piece of our business. That's the second key driver that we've seen already and that we'll continue to see in ARPS.
We've talked about the mix shift going to collections, and I think we've done a really nice job of driving collections as a % of total sales, driving it up. We've priced it to be attractive. We've marketed it. We've put a lot of wood behind that arrow to drive people to collections, and it's having great success. As it does, of course, on the new product side, that brings in a higher ARPS sale than a standalone product would. The other interesting point is, as people move from maintenance to subs, back to the example we talked about, we see in the fourth quarter, an increasing rate of those who are eligible buying up, trading in a single user license and buying up to a collection. That, of course, drives ARPS up as well. You see where we are on the total mix of collections.
There's more room there. There's more upside in that. We're not yet quite at the level that we ultimately got to on suites as a % of sales. Finally, lower discounts. We've talked about this several times. This is one of the ways we go after that 2 million legacy users. It was 2.2, it's down to 2. Price is only one lever. Steve talked about a lot of the reasons that they'll come over is it's higher value. If they're on a perpetual license and it's not tied to maintenance, it's aging. The people that they're working with in their ecosystem, as they move up, it gets harder and harder for them to trade files. There's a lot of reasons people will move. Price is only one of them. You've seen us run this play.
When we hit the end of sale of upgrades, we ran this too, with a steadily declining discount offered inside the promos. The first time we ran this was the first quarter after we ended sale of perpetual license on standalone products. We withdrew it in Q4 of 2016. We ran this at a 70% discount. We ran it again in Q3, targeting that same base at a 50% discount. We ran it twice last year at a 30% discount. It was actually bifurcated in Q3 between collections versus standalone, and it's running right now at a 25% discount. As we do that, of course, the yield to us on the new sales goes up, right? Because we're not discounting as heavily. The other interesting point that drives ARPS up, if you go back here, these were a pretty big %.
They were outsized percentage of the total product subs base in that first quarter after we stopped selling perpetual licenses. As we build up that base of product subs, these get dampened out in the effect that they have on the overall ARPS. Lower discounting on promotions as we go after, not just as we go after legacy, but lower discounting overall will drive ARPS up as well. Here's the trend on ARPS, and the green line is what you've seen. We've reported this. That's the total ARPS, and you sort of see it bounced around pretty flat from Q3 of fiscal 2017 until Q4, the quarter we just announced earlier this month, where we saw a nice uptick in ARPS.
The interesting thing is, behind the scenes, what you couldn't see is the core business ARPS actually inflected up a year earlier for all the reasons that I just talked about, right? We've seen a nice uptick in overall ARPS. Even more impressive, we've seen a bigger uptick in the core ARPS because of the combination of these effects, and they don't stop in fiscal 2020. I'm just showing you right now through fiscal 2020, this doesn't suddenly hit an asymptote and stop growing. We'll continue to yield benefit and ARPS growth beyond fiscal 2020 for these same reasons. All right, that's the P. Let's talk about the Q. If you think of P times Q. Let's talk about what's driving core subs volume. Again, four things to understand. The organic growth in these core markets, and I'll peel that back for you.
You may be surprised, it might be a bit bigger than you think. Then three that Steve talked about. This drives volume into our core business, right? The growth of enterprise, the conversion of legacy customers, and the conversion of piracy. Design markets. Lisa showed you both of these numbers earlier, the $12 billion for architecture and engineering, the $16 billion in manufacturing. It's a $28 billion TAM just in design that we're chasing today. That CAD market grows at a compound annual growth rate of about 8%, right? This is a Cambashi and Oxford Economics estimate. Nice growth. Nice growth in the core business. Obviously, we do quite well in that space. That's the first reason that we see the core business continuing to build. I'm not going to spend a lot of time on this. Steve just showed you this exact same slide.
The important thing is we're less than halfway penetrated through the named accounts that we've targeted to sell EBAs to. As we continue to sell to them, we see about a two and a half times increase in the usage, right? That drives volume into the core. A lot of room left to convert these legacy customers. We'll continue to go at them through just price discounts, but that's only one lever. You can see that's not going to tick back up, right? That trains the exact wrong behavior if waiting gets you a better price. All right, don't expect to see those discounts as we go after the legacy customer tick up.
What you should expect to see, though, is as those license age more and more and get harder and more difficult for them to use and that we continue to build value into the product sub, you'll see that $2 million come down. Finally, piracy. I thought Steve did a nice job walking through how we're going after piracy.
The key tenet to me. Our CIO was on my team, we're a buyer of software, and we get companies that come to us and say, "Hey, we want to do a compliance audit on you." There's a way to do that that is where I say, "Yeah, if I'm using your software, I want to pay for it." There's a way to do that that is very much a stick in the eye, and it's very alienating between the customer and the vendor. We're not going to do the stick in the eye.
We are going to get paid for the software that people are using, where we have the opportunity to in this $4 million in mature markets, but we're going to do it in a way that's respectful and that demonstrates that we are actually a customer company, as Andrew said. Lots more to do in that space. The good news is there's still an enormous opportunity. This isn't a static number. It's not like that's it, no one else is going to pirate the software tomorrow. That number will continue to grow. There's a lot of opportunity there in piracy conversion. Overall, what that nets is 14% growth in the core business. Maintenance plus product subs plus EBA. We've seen that growth in subs from 2016 to 2017 to 2018. We see that continuing at that 14% growth rate out through fiscal 2020.
Again, this is consistent with what we talked about on the Q4 earnings call. That's P, what's happening on ARPS, what are the core trends that are driving that? It's Q, what's happening in the core subs. Let's now talk about the effect that that has and talk about free cash flow build. It's one of the most popular slides that I showed when we had this event about a year ago, I wanted to give you an updated view of that. This shows you the buildup of how we get to the $1.4 billion. Visually, I don't care how far back in the room you are, you can tell these two bars matter, right? If you're way back there, you might not be able to read what those are. The first one is net income.
Net income is a huge driver. The second is the change in deferred. If you understand those two, there's non-cash charges in there's some CapEx. If you understand those two, you understand how cash flow builds. Let me peel those back for you. I'll first talk about net income. Simplistically, net income starts with revenue and you subtract out spend. I've just talked about P times Q on where revenue is headed. On the spend side, I think we've done a nice job. I think we've done a nice job demonstrating that we're disciplined on our spend. All the way out, if you just do the compound annual growth rate from where we ended fiscal 2016 out through our expectations for fiscal 2020, it's less than a 1% compound annual growth rate over those four years.
To do that, we've been very targeted because we do have investments that we have to make. Andrew's talked about his priorities, digitize the company and the opportunity in construction and manufacturing. We need to invest in those spaces. If we're going to keep spend flat, we have to be very focused in fewer areas. We've made some divestments. We've actually done a lot. By the way, with those divestments, we reallocate the resources that we're working on to be able to invest in the things that we need to invest in. This mindset also drove the rebalancing that we announced with our Q3 results that I think unsettled a lot of people, but it was for this reason. We need to make investments. We're going to find areas that are not as strategic and reduce our investment there and reallocate those resources.
That was not a cost savings focus. That was about maintaining our spend at this level, but investing where we need to invest. I think we've done a nice job on spend management and demonstrated the discipline that many of you, I know, wanted to see. The second piece that drives free cash flow is deferred revenue, the change in deferred revenue year-over-year. Deferred revenue is driven by billings. Billings minus reported revenue gives you deferred revenue. To understand deferred revenue, you really need to understand billings. Let me talk about the drivers of billings from 2019 to 2020. Some of these we've touched on. The organic growth of the core business I've already touched on. As that business continues to grow at this 14% volume rate and ARPS grows at the same time, that obviously grows billings.
There's a bleed back of unbilled deferred, and I'll show you what that looks like. The renewal base continues to grow. Talked about that earlier when I was talking about ARPS and what's driven our annualized revenue per subscription up. That obviously has an effect on billings as well. Multi-year billings have been depressed, and we see that returning to really the level of around fiscal 2018, the year we just closed. I'll show you what that looks like too. Talked about core billings. I'm not going to go back to that, but that obviously is a key driver of the billings growth from 2019 to 2020. Here's how the bleed back of unbilled deferred looks like. When we closed fiscal 2018, one of the statistics that I gave you on the call is we had unbilled deferred revenue of $326 million.
What that is we sold EBAs during fiscal 2018. We billed them for year one. The customer committed three years. We billed them for year one. That was in our fiscal 2018 results. The second two years, even though they're committed, we haven't billed yet. That's what builds up the $326 million of unbilled deferred. Three-year commitment, so half of the 326 will show up as billings in fiscal 2019, the year that we're in. The other half, the last year of that three-year commitment, gets billed in fiscal 2020. Same thing happens this year. We'll sell a lot more EBAs. We'll bill one year. We'll put two years in the bank. Unbilled deferred will grow meaningfully this year. Even as $160 million comes out, unbilled deferred will grow meaningfully again this year, but you see the stacking effect that that has.
It normalizes in fiscal 2020 because by fiscal 2020, you're taking, in effect, two years out of unbilled deferred and putting two years in at the same time. This drives a lot of the year-over-year growth in billings from 2019 to 2020. The renewal base continues to expand on the billing side as well. What this shows you is the growth in billings, what % of that growth is driven by renewals versus driven by new sales versus driven by our cloud products. You can see about 60% of our billings growth is driven by renewals. We've already talked about renewals come to us at a higher yield than a net new sale does. The increasing mix shift to collections actually has an effect in both. If it's maintenance to sub, that mix shift shows up here.
If it's new, it shows up there. We will see a return to normalcy in multi-year. Let me just talk about that for a minute. What this shows you is the % of maintenance and product subs that were sold as multi-year. This is a unit slide. What % of the units sold were sold as multi-year? They spiked a little bit back in fiscal 2016 as we started the end of sale of perpetual licenses and people wanted to make sure they could lock in their maintenance agreement, we've steadily driven that down. A lot of the reason we drove it down is as we launched maintenance to subscription program, remember, we gave them insight, visibility to three years of maintenance price increases.
What we didn't want is someone saying, "Oh, if I renew now, I can get ahead of those three years of price increases." We eliminated the ability for anyone to buy multi-year maintenance. Maintenance was the biggest part of this base, certainly until we got to the end of this year. As multi-year maintenance went to zero, the % of total sales that were multi-year came down pretty markedly. The interesting thing that happened during that time frame is, of course, we're selling product subscriptions at the same time. As we sell product subs, and you look at the % of those, there's no price incentive on that today. As we sell product subs, and you look at the % of those that are multi-year, it's in that range that it's always been historically in, that 20%-25% range of product subs.
As product subs now become, it's bigger than maintenance. Product subs are bigger than the maintenance base as of the end of fiscal 2018. As that product sub base becomes the biggest part of the overall base, that will pull the overall average and multi-year billings up. Here's what it'll look like. The blue bars are total deferred revenue. What you see is that you can see what the year-on-year growth looks like. Importantly, what the green line is how much of that deferred revenue base is long-term, right? That's, of course, where you'd see year two and year three of a multi-year agreement. We'll hit the bottom on that curve this year in fiscal 2019, and it'll tick back up in fiscal 2020 to about the level as it was in the year we just closed, in fiscal 2018.
Still below where multi-year was in both fiscal 2016 and fiscal 2017. That's driven by the product subs, which are already at about that level. Just becoming a bigger and bigger part of the base. That will drive billings growth this year and next into fiscal 2020. That's how we'll drive the $1.4 billion of free cash flow or the $6 of free cash flow per share if you wanna do the arithmetic, which I'm probably not allowed to do. All right. Two other things I want to, or a couple other things I want to touch on quickly before we talk about fiscal 2021 through 2023. First is CapEx allocation, there really hasn't been a change in our overall tenor or our policy on CapEx allocation. It's first and foremost to support the business. We've done a lot of shareholder returns.
If you look at the Cash we've spent, certainly over the last two years, it has been almost exclusively spent on shareholder return. We spent $1.3 billion buying back stock in the last two years and reduced the share count, the average, the outstanding share count by about 6 million shares, a little more than 3%. The interesting thing inside there, I've been asked several times, how are you doing that? What's the methodology you're using to buy back the stock? As the share price rose, we shifted the balance from systematic by X amount every day to, okay, we'll have a small amount of systematic buying, we'll have a large amount of opportunistic buying based on price grids and some Monte Carlo modeling and a couple of momentum indicators.
As we've shifted to more opportunistic buying, of course, the markets have helped being somewhat volatile, we've actually saved almost $5 in the average price that we've bought back stock relative to the VWAP over that timeframe. If you do the math, that says we've saved about $80 million on the buybacks we've executed over the last two years. I think we've done a nice job on CapEx allocation and on CapEx return. Another question that's come up several times since the end of the year is tax reform. "Hey, Scott," the question kind of goes like this, "When you rolled out the $6 of free cash flow per share, that was before tax reform, that must be a nice tailwind to you as you come into fiscal 2020.
What does the $6 become?" The answer is it really has not much effect on the $6, and I'll explain why. First of all, let's be clear, the tax reform is good for Autodesk, right? It will take down our blended rate out through time, and it'll make it much easier for us to move cash freely across borders, which was not the case historically. As we've come through this transition, we're not paying a lot of U.S. cash tax. The way this saves you money is if you have U.S. cash taxes, those U.S. cash taxes go down because the rate is lower. We're not spending a lot out here. We don't have a big outflow because we haven't been U.S. cash, sorry, U.S. tax profitable. Right.
Once we do flip to U.S. tax profitable, we've got about a couple of hundred million, a little bit more, of tax attributes built up on the balance sheet. The first thing we'll do as we come back to a point where we actually owe taxes in the U.S. is we'll consume our tax attributes. I don't expect the reduced rate from tax reform to drive an impact to us in cash, certainly through fiscal 2020. Our current modeling shows it's somewhere in fiscal 2021 or later that we actually consume all the rest of our tax attributes, and then the tax rate would become a benefit. All right. I just want to hit that because I understand the logic behind the question, but it doesn't really, turns out, it doesn't really have an effect on us.
The famous snake slide that we've all seen, 28% op margin now in fiscal 2020, which is below the last time you saw this slide. The last time you saw it said 33% for fiscal 2020. Let me just walk you through what's changed. Annualized recurring revenue, you've seen. It's growing faster than we thought at that time. I think that's a positive. As we've gone through and really focused where we need to focus, and I mentioned that we've done several divestments. Some of those divestments had non-recurring. It hasn't affected our ARR, but non-recurring revenue attached to it. Individually, none of those are big, but when you sum it all up over the last two years, those divestments added up to about $200 million of revenue that previously had been in our fiscal 2020 revenue line that aren't anymore.
This isn't a change in spend. You saw the spend slide. Spend is still where it was, but the non-recurring piece of our revenue stream in fiscal 2020 is about $200 million lower than what it had been last time we looked at this slide. That's what's driving the change in the overall op margin. It also changes the functional P&L. I show this because one of the other questions that I get a lot is, "It's great that you guys are staying flat in spending, but are you under-investing in the changes that you need to make? Are you under-investing in innovation?" Well, you saw Scott Reese up here earlier talking about the things we're doing from a product standpoint. I think it's clear we're not under-investing in innovation. Are you under-investing in the sales transition that you've got to make and the evolution in sales?
You just heard from Steve. That's not an issue either. I wanted to show you this so you get a sense of course, the change here is not about spend, the change here is about the denominator, right? That you can get a sense of how we spend money. That's what gets us then to the goals that we've laid out for fiscal 2020, to the 25% compound annual growth rate of ARR and the $1.4 billion of free cash flow. All right. That's the path to fiscal 2020. I'm sure you'll have questions. We'll have time for Q&A after Andrew wraps up. Let me shift forward now. We're in fiscal 2020. Let's talk about how we go from fiscal 2020 to fiscal 2023. Again, Andrew showed you this slide earlier. There's a couple of things that I think are interesting.
If you look at the growth in ARR, of course, the compound annual growth rate drops from 25% from 2016 to 2020 to 18% from 2020 to 2023. That's just a function of the law of big numbers. As the number gets bigger, the compound annual growth rate goes down. We're still adding more than $2 billion of ARR over those three years to the top line. Again, still heavily dependent on what's happening in the core business in the blue bars. However, the green bars are a lot more prominent. The green bars are more prominent on both the sub side, where it's easy to see, but also on the ARR side. Fair amount of growth in the new opportunities. We call it cloud. The reason we called it cloud is you needed visibility to, okay, let's do the old model.
Remember, we had perpetual license and maintenance. Let's do the new model, which was subscription. Then we had cloud as a third. We're not selling cloud. No one comes to us and says, "Hey, can you sell me some cloud?" They come to us and say, "I've got a construction issue. Can you help me get more efficient on the job site? Can you help me through construction? Can you help me link my design to the make on the back end?" At some point, cloud's not going to be the right definition for this, but to stay consistent because a lot of the models are built this way, I want to show you what cloud looks like out through time. You see the ARPS growth there. That's a compound annual growth rate on ARPS, by the way.
It's a little bit below 5% through that timeframe. How do we get there? The first, of course, is the core business, and all of the factors that we talked about in the core business will be maintained out through fiscal 2023. Now we will invest and expand into construction and manufacturing. I wanted to get this number into the presentation somewhere. That combination of the core business growth and the new opportunities in construction and manufacturing will yield about $2.4 billion in free cash flow in fiscal 2023. For those of you doing the math at home, divide that by 220 to 225, that's $11 of free cash flow per share. That's still the goal. As we think about the focus and what's the most important to us going forward, it's not just free cash flow by itself. It's not just a single metric.
The sum of free cash flow and revenue growth, and most of you know the rule of 40, that's really the way we think about that. The $11 is unchanged, but our focus will be not just on the $11, but also on revenue growth out through fiscal 2023. All right. How do we maintain the foundation in the core business? ARPS continues to grow. The growth rate comes down for the same reasons we've talked about, but it's also the same factors that are driving the growth in ARPS between now and fiscal 2020. Those same factors continue. The base of renewals stack up at a higher net to us. Direct sales continues, and particularly the digital sales piece in that, which sells at pretty close to SRP, and the mix of collections. New subs grow the organic business.
We see it growing on a subs level at about 14% between here and fiscal 2020. It's going to continue to grow in that double-digit rate out through fiscal 2023. EBAs, there's more gas in the tank on EBAs. We're less than halfway through the accounts that we want to convert and continuing to go after legacy and piracy. At the same time, the new opportunities in construction and manufacturing are pretty compelling. Compelling not only from a market opportunity. These are the TAMs that Lisa showed you earlier, the TAM for the new opportunity in construction, the TAM for the new opportunity in manufacturing, but also because we're pretty uniquely positioned to go after that, having a foot in each camp. As manufacturing and construction come together, that puts us in a pretty strong position. What matters most?
There's been a huge focus both internally and for many of you on the free cash flow metric. I just want to run through a couple of quick slides. What this plots is free cash flow margin on the x-axis and revenue multiple on the y-axis. These are real public companies. You probably can't today read all these down here, but you'll get the slides, and you'll be able to read those. These are real data points. You see it's pretty much a scattergram. You see the R squared of 0.15. It's not a great predictor of market value. Let's look at the other metric. Let's look at revenue growth now on the x-axis and revenue multiple on the y-axis. Again, you have pretty much of a scattergram and a pretty low R squared value in terms of the predictability. How does that correlate?
The interesting thing, and this is obviously what underpins the rule of 40, when you add those two together, you get a much tighter grouping. We've annotated on here some of the companies so that you can see them. You get a much tighter grouping and a much higher correlation to really driving shareholder value. As we think of the longer term, the $11 is still in play, but focusing on 11 by itself is not sufficient to drive value. We need to focus on both revenue growth and on the free cash flow margin. The interesting point is you see where we are out there, and that's looking at fiscal 2020 with the goals we've laid out for fiscal 2020. It's below the line. Key takeaways. Then I'll hand off to Andrew. Start with fiscal 2020. The goals are on track.
The things that we've talked about are on track. ARPS is growing for the reasons that we talked about. We actually are growing as a result, growing ARR a percentage point faster, compound annual growth rate of one percentage point faster than we previously thought. Core is driving that, the core volume is staying strong. The free cash flow is really just the sum of net income. You've seen us demonstrate now great spending discipline. The top line plus our spending discipline is what drives net income and then the growth in billings. Beyond that, the trends that get us to fiscal 2020 don't just stop. It's not like we hit an asymptote and go level at that. The core business continues to grow out through time, and we've got some pretty exciting new market opportunities to go after.
That's both the here to fiscal 2020 and fiscal 2020 and beyond. With that, what I'd like to do is, first of all, thank you all, and then bring up our CEO, Andrew Anagnost.
Thank you, sir. I hope you were all paying attention because Scott Herren had the defining quote of the entire day. We are not going to be a stick in the eye. If everyone was paying attention to when he was talking about that, we're not going to be a stick in the eye. One of the things we tried to accomplish over the day was talk about the five-year view of the business and the five outcomes that are going to get us there. I'm going to summarize some of the things you've been hearing and some of the points that were made and try to bring home some of the key takeaways for you. Let's go back to the five outcomes. Completing the subscription transition, right up there on the front and center.
What were some of the things that I want you to take away from this and remember? The first thing is completing the subscription transition also means delivering value for all of these people, our customers, the ones that actually make all of this possible. You're going to see us doing a lot of work to ensure that our customers see the same level of value creation from what we've done moving to subscription and the whole technology portfolio you got a little bit of an insight into today. They're going to see this build out the strength of the products they use every day. You also heard we're staying on track and affirming our path to the $1.4 billion in free cash flow.
You know we're basically heading in exactly the same direction with actually better results on ARR growth and better results on ARPS growth, which brought us to this slide. What's changed? I want to make sure I summarize for you exactly what's changed. What hasn't changed is the number of users we're adding to our ecosystem. It's going up, and it's going up a lot. What has changed is that the number of cloud subscriptions we're delivering in that ecosystem is going down because we're consolidating that portfolio. We're growing our cloud subscriptions dramatically. You can see the numbers over that year. An 18% CAGR over that timeframe in terms of the total subscription growth, but a much, much larger CAGR for the cloud subscriptions. Same number of users, significant user growth in our base, fewer cloud products being sold to the same users.
Dramatic growth in our cloud business, but fewer users. You see the ARPS appreciation, you see the results, higher ARR results, lower subs CAGR, stronger ARPS. A better outcome as we're getting closer and closer to the end game for FY 2020. Now, in terms of digitizing the company, we talked about a few things you're already seeing right now that create the proof points about what it means to drive digitization. One of the things that Steve talked about was how we're employing new types of agents and machine learning algorithms and AI algorithms into our customer interactions. Low-value activities like requesting a license or reactivating some software are being done significantly quicker, and you can see how the time goes down to basically nothing on this scale, but without any human intervention involved.
That's allowing us to drive not only efficiency and scale, but a dramatically improved customer experience at the same time. They get what they need faster. We get more efficiency and scale so that the people who used to be doing these very low-value-add activities are actually engaged in higher value-add activities with our customers. That will continue to increase over time as we digitize more and more of back into the company. The other big thing you're seeing around the digitization efforts is what's happening with the digital sales. We're building out a whole system that not only allows us to transact digitally, which is nice, and you saw from Scott the impact that's going to have on our realization over a multi-year period.
We're building out a system that allows us to understand how to help those customers that we're bringing in digitally, help them buy more, help them get more out of the solution, and ultimately renew more effectively. Those are some of the things that are already showing up in terms of the digitization efforts of the company. We talked about my big favorite one, driving the BIM model through the entire process. I hope you got a real understanding of the fact that we're moving from being a big leader in this $12 billion TAM to also being a significant leader in the $10 billion TAM that's ultimately going to be arriving out there in the construction space. We highlighted a few things, both from the market and the technology perspective, that we're driving our advantage and our long-term ability to do this.
First off, we're the leader in BIM. Autodesk is the only leader in BIM. We're putting BIM at the center of the entire platform that's being driven into the construction market right now. We've built a cloud construction platform, and we're targeting an end-to-end workflow. We're not building out a point solutions. We're building ecosystems with third-party partners, with capabilities that we have, with capabilities we're developing on our own to build out the end-to-end workflow. More importantly, we are the only company that has both AEC and manufacturing expertise, and that is going to help us do something that our construction customers are very eager to do, and that's industrialize the construction process. That's why only Autodesk is going to be able to lead in this particular effort.
Now, the other piece I talked about was this whole notion of automating the process of designing something for manufacturing. Again, we talked about the TAM. We are vigorous competitors in this $16 billion product design and engineering TAM. We're starting to become a more significant player in this $11 billion TAM in the manufacturing processes because of this disruption in the way our customers are working. We are actually going to help these customers get to the new world of robotic hybrid manufacturing, where 3D printing to subtractive printing work together in highly automated factories. That's something that we're highly committed to, and it's what we're doing with the generative work that we've invested so much in over the last few years. If you look at some of the advantages that Autodesk is bringing to this, we've got a seven-year head start.
We've been paying attention to this market well ahead of our competitors and actually building a design environment that targets the new way people are going to work. As a result, we've established some pretty clear technology leadership, and we intend to keep that technology leadership. The platform is focused at converging design and make together, bringing these two processes together, and it's built on these generative algorithms that don't just modify geometry that somebody created but actually automatically create geometry based on specifications of how you're going to build it and what you're trying to accomplish. Finally, we've invested a lot in winning the hearts and minds of the next generation. Fusion, in particular, is getting rapid adoption with students. If you look over a five- and 10-year period, when you win the students, you win the market long term.
That's what we've been focused on, and it's working fabulously. Which is going to allow us to exploit their passion for bringing design and make together inside of companies all over the industries we serve. A lot of things that only Autodesk can do, and we intend to keep doing them. When we look again out the five-year timeframe and the five outcomes, this idea of bringing construction and manufacturing together, of turning a construction site into an outdoor factory and employing the techniques that have been learned and hardened and perfected in manufacturing into the construction environment, but also introducing some of that flexibility back into the manufacturing world. This is something only Autodesk can do. We're the only player that understands these two customers to the depth and degree that needs to be understood in order to drive this transformation.
You combine that with our knowledge of the Building Information Model and our leadership there, our knowledge of the product model, and you've got a real combination to transform the way the customers work in this space. That's how we're going to become the leader in design and make automation over the next five years, which is going to have real tangible results for our business. Again, if we look out here, when we get out to FY 2023, you're looking at almost half a billion dollars in those new businesses. Like I said, it's hard to tell because you're dealing with a scale that actually goes up to $6 billion. We're delivering immense value, and that is just on the organic path of building out the solutions we are already developing today.
Again, because we're consolidating the portfolio, the CAGR in subscriptions is at 12%, but not because we're selling to fewer users. We're just selling fewer types of products to those same users. Again, at higher ARPS. We know how to turn all of this technology, all of this focus on the five-year outcome into long-term value for the company. Now it's time to break into the Q&A session, and I appreciate all of you sticking with us and paying attention to everything we had to say. Team, come on up. Thank you. Oh, yeah, no. Thank you. We've already got the first question. We haven't even got everybody up on stage. He beat you, Jay. Oh, you did? Okay. Did he? Does anybody see? I wasn't paying attention.
The replay. I have to go to replay.
All right. Well, unfortunately, ceremony-wise, Jay always gets the first one. All right. I'm going straight to you. Jay.
Thank you. One for Andrew, one for my friend, Mr. Blum. Andrew, you spoke of certain eventualities today. Eventualities in terms of how AEC and manufacturing-
Yep
would evolve. Eventualities in terms of your own business, particularly that you would be a consumption model, and you were even more emphatic about that perhaps today than you were at AU when you spoke about it. I appreciate what Scott Reese showed today in terms of certain technologies. It seems for all of those eventualities to occur, and you spoke of only Autodesk this and that, which is conceptually very Adobe-like. There seem to me some very basic underlying architectural requirements that you have to execute. This is the famous Quantum question, I suppose. Just talk about getting there in terms of tying all this together-
Yep
Because the onlyness can only occur, I think, if you have the right architecture and platform.
That's your first question?
That's the first question.
Okay.
Let me just see. According to the 10-K, 31% of your business last year came from Tech Data, and they've always been a big relationship, but it's growing. By the time we get to 2022, 2023, if they maintain that percentage or even grow it, that's a more than billion-dollar relationship with one partner. Talk about the infrastructural or margin or efficiency implications of something that immense in terms of the relationship.
I'll start with question one, and then I'm going to let Scott elaborate on question one. You're right. Everything Scott showed you is based on an architectural framework that serves up building information model data that is specifically tailored to types of workflows and types of consumers. That underlying infrastructure is what we've called Project Quantum in the past. It's basically the way we deliver up information to these individual views. You remember that video that Scott showed where we're actually-
Feeding out the pre-fabrication information and then stuffing it into Inventor so that it can actually be pushed into a manufacturing hub. It's Quantum that's pulling that information out of the building information model and moving it over to Inventor. Something that our customers have been asking us to do in a seamless way for a very long time. The ability to move the data into the cloud, to deconstruct it and feed it up in particular ways is allowing us to get there. You're right. There has to be a certain technological underpinning to make that happen. Scott, do you want to add anything?
There's not a lot to add. It's an interesting question. If you think about the challenges that we've had across all of our industries, across anybody who participates, a lot of it is rooted in the way information is passed back and forth in the form of a proprietary file that has to be deconstructed, and then read in. You always lose information, you lose intelligence every time you pass one of those files back and forth. Quantum, as you are seeing, that is the underlying way that we're going to resolve that and allow data to flow between everything that Autodesk does, but also with anyone who participates in the market, competitors included. Yeah.
All right. Wait, we're going to answer the second part, and then you have priority on the next question. Steve.
Okay. First of all, we really do enjoy the relationship we have with Tech Data. I'm not really wanting to talk about a specific partner at any event or things like that overall. What I do want to comment, so Tech Data is a distribution partner of ours. Our relationship with our distribution partners has evolved over time. Where they added value in the past and where they're adding value today is actually different. It continues to change, and it continues to evolve, especially as we build out more of our own digital capabilities.
We've been very open with the folks over at Tech Data because we have a very strong relationship with them, saying, "Continue to innovate, continue to find ways to add more value as we continue to digitize and automate ourselves so that as we continue to build out the digital capabilities and the transactional volume and scale that we're focusing on with our priority of digitizing the company.
I love the way Steve says this. It's so constructive. I'll tell you, every year when I meet with Tech Data, we sit down because every year I meet with them, I say, "All right, why should I work with you for another year?
Yep.
Every year, they create a reason why we should work with them for another year.
Yeah.
Is it going to be worth $1 billion in the future? Completely unclear. Every year, I sit down and say, "Tell me why we should be working together for the next two years." They bring a new reason in terms of value creation that they've managed to do. They know as a distribution, given the world where we're going, they have to find new ways to add value. As long as they continue to add value, we'll keep working with them, and they've done a pretty good job of that.
Yes, they have.
All right. There you go.
Hey, Mike Nemeroff from Credit Suisse. Thanks for hosting us today. Two questions, one to you, Andrew. At the Autodesk University a couple of months ago, there was a customer panel, and the use of BIM from those customers was relatively low. I know it's-
Yep
regulated outside the United States. A lot of what you talk about in the future drives BIM throughout the platform. I just want to understand, why do you see this massive shift-
Yep
in BIM over the next five years, How confident are you and what gives you the confidence that that's actually going to happen? That's where a lot of the strategy goes.
Yep.
For Scott, on the operating margin guide, just want to understand the reduction. Is it all from divestitures, or is there some increased spending? I just want to clarify where that reduction is coming from specifically. Thanks.
All right. Let's start with your first question. To frame it, there's actually two things that are going to drive this. Nothing's done in five years, but it's made significant progress. The first thing is that the AEC space has been progressively following the exact same path as the manufacturing space was following for about 20 years now, just slower. All right? It's a more complex ecosystem. What inevitably happens when people are trying to increase reliability, predictability, and diminish the amount of waste in a process, they start queuing off the model to control the process. When I say that's inevitable, it's inevitable because it's following the exact same pattern of what's happened in manufacturing. The inevitability is no prediction of when it's all complete timeframe-wise or what's going to happen.
This queuing off the model to manage predictability, to manage waste, it's going to happen. It's already happening. Our smartest customers, and they come from certain regions, are in the construction. They're already doing this, and they're coming to saying, "Hey, how can I do this with a BIM model? How can I do this like what they did with automotive design? How can I do this?" We're already having pretty significant discussions with some really big design build firms about how they can look more like an automotive company, where they have a design department and a department that looks at manufacturing engineering and another department that then looks at kind of the plant engineering and how you actually build it. It's inevitable. They're already thinking this way. The other aspect of this is how you continue to change the ecosystem over time.
Just like in the evolution of manufacturing, the large OEMs ultimately drove a pretty significant change in the contractual, the relationship environment, and all the things that happened in the ecosystem of manufacturing. The owners and operators of buildings are going to start driving significant changes with regards to how the ecosystem now gets. We participate in that. We encourage owner and operators to specify BIM for all their products. They get it. They get what the benefit is for them. They get why they want to do it, and they also get that if the process is fully BIM-enabled. If it's BIM from start to finish, they actually get back an as-built asset that helps them manage and operate the building over time, or even something that they can use to retain and enhance the value of the asset over time.
those are the two levers that we're going to be working over time to drive the sequel, that we're actively involved in both of them.
to your question, Michael, I know you haven't had a chance to measure the bars yet on the slide. I'm sure that'll happen. Spin is almost exactly where it was historically. I think the other important point is, it is about revenue reduction and that falling through to the bottom line. It's non-recurring revenue. You see ARR is actually growing faster than it had before. We've done a series of a lot of our consumer products, Creative Market. This is something that we spun out. Digital Arts is something we spun out, Continua. It's a lot of sort of cats and dogs that we had accumulated over time that we spun out that had small amounts of non-recurring revenue. You see the flip side of it, though. You see where we already are in recurring revenue.
The end of Q4, we're at 93% recurring. That now goes up into the high 90% range out through fiscal 2020. Yes, it is non-recurring revenue that's fallen out, and that's what's changed the margin, not any change on the spend side. You see the flip side of it in an even higher % of recurring revenue.
I think what I think is interesting about this whole discussion, we're constrained to fiscal years to have this conversation. When you really look at this, what's really going on is if you just moved out like a month or a quarter further, it all shoots right back up because of the accumulation of recurring revenue we're seeing. We're talking about fine gradations of the flow of money here and how it accumulates. In the big macro sense, it's all heading in the right direction.
Hi, it's Gregg Moskowitz from Cowen. First question, I guess, is for either Andrew or Scott. You've talked before about how fiscal 2020 is not steady state, but there have been some who have been really concerned that there was going to be a meaningful drop-off post that period. You've also today guided to a really healthy 18% ARR CAGR from 2020 to 2023. Scott, you talked about four main drivers of ARPS, and frankly, none of them included pricing. My question is it fair to say because of that when you look at that 18% CAGR from 2020 to 2023, that you're not currently assuming any meaningful price action?
I'm sorry. Do you want to do it?
I was supposed to tee up the question.
All right, go ahead.
What I was going to do is say, well, I was going to make one comment and hand it to Scott. I want to be very clear. We are not dependent on any kind of significant price action to hit our numbers. In fact, that's not the way we want to work with our customer base over time. I just want to lay that out as a critical point. Scott, there's not a whole lot else to add other than-
I think you will see us get on a more disciplined, kind of a regular rhythm of small annual CPI-type annual price increases. No, there's no significant uptick or step-up. I think that's why on the one slide I talked about that as a misperception around ARPS, because I know that had been circulating a bit that, hey, once you annualize that, there's no more increase, and it's really, price obviously does have an effect. Even as we go through Intuos, it has an effect on ARPS. Those four were listed in order of importance before that I talked through.
Okay, great. As a follow-up for you, Andrew. You provided us with cloud ARPS and core ARPS in fiscal 2020 and 2023. You're talking to a group of analysts. You had to figure some of us might try to reverse engineer that. Based on my math, it looks like you had cloud going from 14% of the subs mix in fiscal 2020 to about 21% or so in fiscal 2023. First off, is that correct? Secondly, how do you envision the ramp occurring over that time period when you look at BIM 360 relative to Fusion 360?
Yeah. All right. You're probably roughly right. I don't have the exact mix memorized off those charts. I should go measure them myself. You're probably roughly within the right ballpark. When you get to more higher fidelity with your rulers, you'll get it better because the charts are accurate. When I look at the ramp, a lot of the ramp is going to be driven by BIM 360 initially, because right now the whole construction opportunity, it's market ready. It's technology ready. The market's ready. They want technology to digitize the construction site. They're hungry for it. They're gobbling it up. You're going to see that ramping up faster. The technology that goes into Fusion and is related to manufacturing is more technologically complex.
People are still trying to wrestle with the, "Wow, so I have to change my process in the future to adapt to new ways of working." It's going to ramp up slower than the construction opportunity. The great news about that is as you move past FY 2023, it provides additional fuel even beyond that threshold, because we're only looking five years out. Do you want to say anything?
No, that's good. Thanks, Greg.
Hey, guys, this is Rich Hilliker from Wells Fargo for Phil Winslow. Unfortunately, Phil got stuck in N.Y., so I'm here in his place. Just two quick ones. If we take the 2 million active non-subscribers and we add the paying subs, we get to 5.7 million, and that's about 10% higher than the last analyst day in late 2016. I know you touched on cloud subscriptions helping that, but I'm wondering if you can give us a little bit more color on what's driving that increase in the active user base. My second question is around the 5% CAGR in pricing after 2020. It seems a little bit conservative, so I'm just kind of wondering, given that end-to-end promotional pricing will end in 2021, can you guys comment a little bit on that?
Who wants to take the first question?
Yeah, I'm not sure I followed all the math in your first question. You said the active base was up
Yeah. If we take the 2 million active nonsubscribers and we add the paying subs, we get to 5.7 million.
Right? I was just comparing that to the last Analyst Day from late 2016. I was just saying that that's about 10% higher.
I know that you've touched on cloud subscriptions quite a bit. I was just wondering if you can just give us a little bit more color on what's been contributing to the active user base.
Oh, got it. There is growth in the core business every year, 14% of a big number turns out to be a lot of growth, right? We added net 600,000, 611,000 subs during the quarter. Don't equate each sub with a new user.
Yeah.
If that's the math you're trying to do in your head. Even as subs go up or down, that doesn't necessarily mean users scale at the same rate, right? Yes, it's growing as we continue to sell more licenses into the market. It's really the core that grew the active user base because what I actually said was that the amount I was calculating of active users from that base actually is down compared to where it was the year before, but it was the growth in the core users. Right, that actually is accounting for that overall.
Just on the second part, about the 5% CAGR after 2020 down here.
Yep.
It just seemed a little conservative given that M2S promotional ends in 2021. Just wondering if you can comment a little bit on that.
No, there's nothing hidden deep in the bowels of the spreadsheet that I'm trying to protect you from. As we ripple that out through time, we're not expecting. Just be clear on M2S. When someone gets to the end of their grandfather, their 3-year period, their grandfather from the time they convert, keep the same price for 3 years, and they don't pay up front, they pay annually. At that point, we've said they just go back to the terminal price. They don't go all the way up to the SRP at that point. There's not this mountain that we're going to get and that our customers are going to face as that 3-year lock-in period comes back. It's just doing the math out through time.
Okay.
Hi, this is Monika from KeyBanc. Thanks for hosting us. Your revenue expectations for fiscal 2020 is about $200 million lower than your previous expectations, but your free cash flow number is unchanged. There must be some positive surprise from your previous expectations. Could you walk through what are the?
Scott, why don't you take it? Yeah. Monika, it's the things that I tried to walk through on the few times queue. Remember, ARR is actually up from the previous guide. Nonrecurring revenue has come down as we've divested some of the former consumer product groups and small products that we had that actually were driving revenue. Any one individually was not significant, but when you sum them all up, it's about $200 million of nonrecurring revenue that's going. What's driving the overall billings growth, which is driving the deferred revenue, is the trends that we talked about around ARPS and around the growth in the core business. It's actually all the same trends that we just talked about.
As a follow-up, for renewals, you talked about channel getting almost 10% lower discount. Maybe, do you think you could get some kind of backlash from your channel, just given that 70% of your revenue is still from channel?
Yes. Steve, why don't you take it? Yeah. Well, it depends on what you determine backlash to be. Do partners like having less dollars paid to them on a particular activity? No, they don't like that at all. They actually can make more on selling new in other places. Ultimately, we're trying to align, as I mentioned earlier, where do they have to put their effort? Where do they have to invest their resources? How do we align the way we compensate them for that, to ensure that where their biggest investments occur, they have the opportunity to make the most. If they're doing their job properly, the renewal should not require a lot of effort. It should not require a lot of investment of time and resource. That's the methodology we're focusing them on.
If you ask a partner, are they happy that we're paying them less for renewals this year than the year before, I assure you will get a, "We're not happy about it at all." They understand what's going on, how the economics should align with where we want them to invest and focus. Yeah. I think the last part of what Steve said is really important. Steve and his team, and working with my team, have done a great job modeling out the economics of a channel partner and having them understand there's a message under here. We want you to drive renewals, but actually, you're going to make more money if you allocate resources to selling new and to driving adoption.
If they do that effectively, they're actually our biggest. We just came out of our sales kickoff meeting earlier this month, and our biggest resellers actually understand that quite well. Right. The other piece that goes along with that is drive more of your profits as a reseller out of the services business and not out of just straight product markup. If all you're doing is product markup, it's going to be a tough business long term. The interesting part is by focusing on the services part of the business, which is a way for them to actually earn incremental dollars, the renewal even becomes more likely. Exactly because the service business actually drives the effective use that drives the renewal. By doing that, they bring more income in, they get more gross profit, and they have to spend less time on the renewal.
We're just aligning with all that. Thank you. Thanks, Monika.
Thanks. Saket Kalia from Barclays. Two questions for me. First, maybe strategically for you, Andrew, or you, Lisa, on construction. How do you segment the competitive landscape in that market just as you attack it even more aggressively? That's the first question. Secondly, maybe for you, Scott. You gave us, and thank you for the specifics on ARR for the model into FY 2023. Can you give us some broad brushes on what margin and maybe billings looks like in that model as well?
Lisa, why don't you take the competitive question?
The way we look at the construction market is you would actually say, we look at competitors across. There's design, there's pre-construction. You saw Scott walk you through that whole thing. There's the construction site, there's operate and manage. When we do analysis of the competitive landscape, there's players in each one of those areas, but we don't have anybody that's doing what we're doing, which is across the entire life cycle. That's how we monitor the competitive landscape. Who's in design, who's in pre-construction, who's doing the site automation, who's in operate and manage. All of our charts, they show that, we understand what's happening there.
So far today, we have not seen one that's doing what we said we're doing, which is everything based on that BIM model for the entire life cycle, and it's the one source of truth.
On the margin question and billings growth out from 2021 to 2023, I don't want to get pinned down to a number on that. Margins obviously expand, right, as we work our way out through time. Billings growth continues to grow strongly for all the same reasons that I laid out here. We've given you a lot of the breadcrumbs to be able to back into what that looks like by telling you what ARR looks like and what the free cash flow is that that yields. You can probably model it pretty closely. I think the way we're thinking about it longer term is shifting away from just the free cash flow metric as an individual hit this and nothing else matters number, to that's not enough, right? We've got to focus on both the free cash flow margin and revenue growth on the top line.
I'm not saying free cash flow margin is not important. It's very important, and we'll continue to drive that. You can drive it with zero revenue growth, or you can drive it with high revenue growth, and we're on the second path.
You want us to optimize the sum of those two numbers, not any one of them individually.
That's right.
Excellent. This is Keith Weiss from Morgan Stanley, back here in the bleacher seats.
One strategic question.
You got to get here early, Keith, if you want to see us, Ron. Jay was here at 7:05 A.M. this morning.
One more strategic question, one more tactical. On the strategic side, again, relating to construction. Really good job of laying out a really big opportunity, and a compelling value proposition in automating what's going on in the construction market. Frankly, that's existed for quite some time. We've heard about the 25%, 30% wastage in construction for a while.
The under-investment in technology has persisted for a long time. The question is, why now? What do you guys see in the near term that's actually going to inflect that? Is it going to get those construction companies to start spending on technology? From your perspective in getting the right products into the field, you mentioned the VC investment going on in this industry. How much of it comes from internal investment? How much of it is going to be M&A, and you guys getting back into the market for M&A and sort of picking up those technologies? That was the strategic one. The tactical one is on cloud, and specifically cloud ARPS. If I'm not mistaken, the FY 2020 targets for cloud ARPS remain relatively the same.
I thought a lot of what we talked about in the last quarter, in the last two quarters was fewer subscriptions, higher value subscriptions. Why doesn't that cloud ARPS number move higher?
All right, let me take the first one, I'll give Scott the second one. All right, on the first one. Let's talk about what's changed, right? In some ways, nothing's changed. In some ways, almost everything's changed, right? The nothing that's changed is that the rate at which some of these companies are taking up technology hasn't really changed that much. It's actually been pretty rapid already, at least in terms of construction site automation and bringing mobility to the construction site and everything else. We're just piling on top of that with better data, more tightly connected to the building information model. That's kind of stayed the same, and we're going to ride that wave. What's changed is that everybody's starting to wake up and say, "Hey, it seems like there's something going on in construction.
It looks like it's starting to be real. Maybe I should start to participate in that as well." What you see is in the environment, you actually see the beginnings of consolidation activities with other portfolios going on. That actually has the effect of accelerating the seeding of more and more technology into different spaces, because as companies consolidate smaller companies in their portfolio, they start seeding that technology into other parts of their processes. The climate we're competing in terms of people starting to consolidate, the smaller companies starting to suddenly show growth that they weren't showing, that's absolutely changed. This kind of early market part that we've been chasing in terms of site execution, that's just humming along at where it was.
You combine those two things together, now is the time to start really going hard on that opportunity because you can already see that over a 5-year period, it's going to start to take off, right? Nothing changes dramatically in some of these technology areas, but because some of the problems are not technologically complex, you're going to see that amplification start to happen pretty rapidly. That's changed. When you look at the strategic question, look, historically, we've been a very acquisitive company. It's actually the last year and a half or so that has been an exception, not the norm. When you look at the size and scope of this opportunity, and when you look at what's happening in the market already and the number of venture-funded companies that are out there, you can certainly expect us to be acquisitive again moving forward.
Keith, on the cloud. Sorry.
No, go ahead.
On the cloud ARPS question, certainly some of the things that we've talked, stopping the seeding programs helps. Some of the packaging that Lisa walked through, kind of packaging at higher levels that are greater value helps push cloud ARPS up. What's pushing it and to kind of keep it at equilibrium where it was is we've had great success selling user packs, and especially within BIM. The importance of user packs, obviously, if you buy 10, you get one price per user. If you buy 100, you get a lower price per user. If you buy 500, it's less still.
If you think about the way BIM gets adopted, it's probably going to get adopted on kind of a bowling pin strategy, where you get the big company at the top, and initially they're going to buy the users that they want, and they're going to mandate that you use BIM 360 out on the job site. User packs have been quite popular, and as they are, that puts downward pressure on the cloud ARPS. Those are the two offsetting factors.
Excellent. It's very helpful. Thanks, guys.
Yeah. Hi, guys. Alex Dow from Deutsche Bank. Just a couple from me. The first one, still quite early days with subscription churn rates, but what are you seeing there? You talked historically about maintenance churn being in the sort of mid-teens range, probably best in class for mid-market software. Could be more towards the 10% range, and obviously, you're trying to go up more into the enterprise as well. What trends have you seen over the last few quarters around the churn rates and subscriptions, and where do you think that can evolve to in the coming years? That's the first one. The second one, Andrew, you talked in the vision about machine learning and automation increasingly in construction and manufacturing.
I think anyone who talks about machine learning talks also about needing lots of data, and I think they also talk about needing to get that data from a cloud platform. Now, in the FY 2023 vision, cloud is still a minority of the subscriptions, quite a significant minority, and it's mainly talked about as a net new opportunity. How are you going to migrate the existing customer base to cloud platforms in manufacturing AEC to drive that vision of machine learning and also-
Let me answer the second question first. You're absolutely wrong. All right. The amount of data we have on our cloud platform right now relative to our customers' construction and design products is massive. Right. We've had cloud infrastructure out there for over seven years, and customers have been moving data into that environment for quite a while. All right. In addition to some of the field data we have with BIM 360 Field, BIM 360 Glue, we already have massive amounts of data. Some of those things you saw from Scott with regards to BIM 360 Insight and some of the capability we were seeing there, that's all based on some of the information that we've been accumulating. We're already well ahead in terms of accumulating data. You're right, there's two things that are required for machine learning.
You have to understand the damn space you're trying to drive machine learning into, because you can't teach a machine to do things you don't understand, believe it or not. You got to have lots of data. We're already well on our way to getting down that path. Don't confuse the size of the revenue opportunity with the footprint in terms of user and data that we've already got. Just applications like BIM 360 Design, previously Collaboration for Revit, huge amounts of information in the cloud about how our customers are using and sharing information. BIM 360 Glue, clash detection, tons of workflow information about what works and what doesn't. Okay. Scott, do you have anything you want to add about the-
That's absolutely right. It's petabytes of data.
You probably got some great stat that I don't have in my head.
Yeah. It's a few petabytes of data, actually, that are in the cloud. It's because if you look at our footprint in the design space, a lot of those users have been using our cloud to store and collaborate around their data for years. When we talk about BIM 360 and that's not really represented there because it's more associated with our design products. Yeah, we have several petabytes of data in our cloud that we're able to learn from and train these algorithms with.
I'm going to turn the renewal rate question over to Scott.
Sure. You said two different things, Alex, and I want to make sure that I answer your question. On the maintenance side, the churn rates, we actually quote it the other way, right? We talk about what the renewal rate is. Our renewal rates have stayed steady. Even in an environment where we've announced maintenance to subscription plan, raised the price 5% on maintenance in the first year, given visibility that maintenance price mid-year this year goes up 10%, mid-year next year goes up 20%. We've seen a huge uptick of people converting. As they convert, you get the example like the one we talked about on the earnings call. Right? That example, those 20 look like maintenance churn. Right? Even in that environment, we're still showing pretty steady renewal rates.
Those renewal rates, just to be clear, are not dollar renewal rates. They're unit renewal rates, right? Of course, we have the highest non-renew on our lowest value product. If we did that on a dollar basis, you'd see those renewal rates be a little bit higher.
I was talking more about the subscription renewal rates that you were seeing at this point. Basically, are the renewal rates going up? I was talking more about the subscription renewal rates, i.e., are they going up in a subscription world where if you don't renew, you can't use the product.
Yeah
which is what you would expect.
Yeah. Of course, they are. Yeah. They are going up. They're in line with our expectations, by the way. We expected them to go up, but they are going up, and it's in line with what we thought.
You're getting a lot of questions about the construction side, which I think what caught me by surprise was the size of the opportunity. 2020, I think you put a $10 billion TAM. When you use the word construction, one, it's not a homogeneous word, and it's not a monolithic sector. I'm wondering a few things at a granular level. What vertical? What's the activity? What's your entry point into construction, and who's the buyer? Is it the GC?
Yeah.
Is it someone else? Where is it in the food chain? Vertical, activity, and who's the buyer? Because that's a big opportunity.
Yep
for something you're talking about two years out.
Let's talk, it's kind of a combination of things from both what Scott and Lisa said. All right? First, in terms of the vertical, it's commercial buildings right now, what we like to call vertical construction. Right? This is where the hotbed of activity is. It's not residential. Residential is adopting to digital, but in a different kind of domain. It's commercial buildings that's the key vertical that we're engaged in right now. You're not seeing as broad adoption in what we call horizontal construction or infrastructure yet, but that's the next wave. Commercial buildings first. Okay. What was the other aspect of your question? You asked for the vertical.
Yeah, Tim.
Who's the buyer?
Oh, yeah.
What's the activity that gets you in?
To understand that, you got to go back to what Scott Reese showed, where he showed the breakdown of the various modes. What you're seeing is two anchor activities that are going on right now. One is in field automation. The buyer there is actually the GC and the subs. The subcontractors and the GC are buying these tools for field automation, quality check, field issue management, and the things associated with that. The buyer over there is there. When you look more towards the other side, where we're talking about pre-construction, clash detection, layout, and plan, the buyer's on the design side, but that's still into the construction TAM. That's where you see people starting to ramp up on BIM 360 Design, where they're trying to share monolithic model data across multiple players.
You see a little bit more of the purchasing on the design side. What you're going to see as time goes on is it's going to bleed deeper into the contractor. It comes from the contractor from field execution in towards the center, and then you're going to see it bleed from the design side over deeper into project management and project planning. Right now, you got the design side, the owner, and the architect/design build firm, and then you've got the GC and the subs over here on the field execution side. Make sense?
It does. If you're pitching me, if I'm the GC and you're in my office wanting me to take you on, what do you tell me? What's the three-sentence pitch on, "This is what I can do for you right now?
I can deliver you two-tenths of an increase in your margin %. Two-tenths. Right there. How would you like that? Would you like another 2.2 in your margin increase? Most of them will say yeah, since usually the margin on a project. What's the average margin on a project?
3%-5%.
Yeah. They usually don't believe that you could even do that at first.
Right.
It opens the door.
Well, you also remove the risk of that 3%-5% margin because that's what they're trying to do. They're always trying to do the risk mitigation, and how do I protect that margin, right? Because of the predictability problems. I'd add one other thing, too, to what Andrew said. From a go-to-market perspective, you can actually go this way and this way. You get your software into all the project sites. I'm talking about where the cranes are. You get the project folks to start using it, and then what happens is you get the attention of the GC in the home office because they're saying, "Wait a minute, you're being used on 50 of my sites?
Right.
I want to talk to you about something bigger." You come in from the top as well because they've been building BIM design models with our tools for years. They're saying, "How do I get more value out of that into the downstream processes?
Right.
Hi, Ken Talanian from Evercore ISI. It's a bigger picture question. If we were to hit a recession between now and fiscal 2020, fiscal 2023, what kind of impact could we see to your growth metrics? What consideration have you given to your cloud forecast relative to that possibility?
What? I didn't--
What consideration have you given in your cloud forecast relative to the possibility of a recession?
One question.
Yeah. I'll give you the answer to the first part. We have a lot more resiliency to a downturn than we've had in the past. Obviously, when you're selling perpetual licenses. Go back and look at the impact the great financial crisis had on Autodesk, and our perpetual license revenue fell 40% in 2009. What's happening behind the scenes is someone says, "I've got fewer employees. I don't need to buy new licenses from Autodesk this year. In fact, I'm going to put those licenses on the shelf. As I hire people back, I still don't have to buy new from Autodesk until I've consumed what I already had." In a subscription world, obviously, that doesn't happen. The first thing is, if you took a 10% cut in your workforce, you still have to pay for the 90%. Right? That annual subscription keeps going.
Even in the financial crisis, we saw our recurring revenue, maintenance revenue, actually grew through that. We are a lot more resilient to the next downturn. It's not that we'd be immune, of course. As the workforce shrinks, the number of subscriptions you need shrink. We'd be a lot more resilient. I think there's a lot of scar tissue from the way our revenue contracted in the last economic downturn that I think is still out there. I read some of the reports where people say it's a very cyclical stock. It's like, it's not that cycles won't affect us, but it'll have a lot lower effect than it had historically. In terms of, we've run different scenarios on that. You could probably run them as well based on your own model.
What's built into the model is no assumption of a recession. I didn't try to predict a recession will start at this point and have this, that, and end at that point. We expected the current economic environment to be where it is.
Okay. Thank you.
Zane Chrane with Bernstein Research. Really great presentation today. Appreciate all the helpful detail. Question on your thinking behind consolidating some of the cloud products. It looks like you're going from a lot of products in manufacturing and construction down to a few. Could you just elaborate more on the thinking that was driving that? On top of that, it looks like there are maybe a little bit of negative surprise at the very low end of the cloud solutions this last year, particularly with Team, but the higher end is doing well. What gives you confidence that the new strategy is the right way to go to market, and you'll have optimal uptake of the new offerings? Thank you.
Let me answer the end part of that, then I'm going to turn it over to Lisa to talk about the repackaging. At the low end, Team, that was a problem we created on our own. What we were doing is we were seeding the market with Team as a way to attract people to use some of the higher-end offerings in the BIM 360 portfolio. It turns out that's just a bad way to do things, all right? It's better actually to go in and start selling the value of BIM 360. That seeding strategy just wasn't a good way to build up interest in the business at the low end. We just stopped doing it.
People are still buying BIM 360 Team when you need it, but because it's such a low-end application, it makes a lot more sense to consolidate it into something else, which I'll leave to Lisa in terms of what we're consolidating into what and why.
Right. Let me start off with BIM 360. You saw that chart, right? We have all these different acquisitions, a lot of that consolidation actually was driven from conversations with our customers. What we're looking at is what are the top business problems and workflows that they're trying to solve. Instead of saying, "Well, here's 13 things that you could pick from," we actually started consolidating around those workflows. The BIM 360 went from eight different offerings, that could have been even more offerings, we said, "Listen, people either want to collaborate on the site, they're focused on BIM 360 Design, they're trying to either do design, they're trying to do planning," right? With planning, what am I doing? It's all the clash detection that's going on, right? Then when I'm out in the field, that's the construction site.
That's where we really came up with those offerings. Quite frankly, a lot of it's driven from what we've learned from our customers, the way that they want to buy and the way that they want to implement. That's why we actually feel very confident about what this pricing and packaging looks like. I would say the same thing with Fusion 360. We've done a lot of acquisitions. On the break, somebody asked me, I thought it was a great question, "Well, what about simulation and" One of the things that we have learned in this new way of generative design and manufacturing, this stuff is all embedded in design. It's not a separate step anymore. What we're doing is we're embedding all of that capability, life cycle management, simulation, that's all a part of the design and manufacturing process.
Instead of making those separate offerings, which in my mind is the old way, that is not about the new way. You actually have to integrate all of that technology because that's how you design and manufacture in the future. We're really packaging based on what has to happen in the future.
Yep. There's one more thing I'll add, too. When you're out there competing in the market and you're competing against various startup companies, you start to look at their strategies and their portfolios, and you find, we're going in with eight SKUs and they're going in with two. It's like, okay, well, I guess we don't need eight SKUs. It just makes life a lot easier to have the conversation with the customer.
That was what I was going to add. It's a lot easier to go to market with a fewer number of things than a lot of things overall. When you're selling both direct and through partners, we add a lot of complexity into the business when we offer many different little things that all are composition of a solution. Actually, from a go in the market perspective, it's much simpler for us, for our partners, and for our customers to be able to deal with these consolidated approaches.
Right.
It takes just as much effort to sell a Team seat to a customer for the first time as it does to sell a more robust solution that goes a lot deeper into their process.
Yeah.
It's just a better ROI.
I think it makes adoption and renewal easier, too, because I have more value now in this offering, which makes it stickier. It's in more of my processes, therefore, I'm going to use it and I'm going to renew on it.
Right.
Great. Thanks for hosting. Hey, guys. It's Gal Munda from Berenberg. I have one question. I'm trying to bridge the sub-number between FY 2020 and FY 2023, going from 4.9 to 7. That 2.1 million extra subs, we know that you told us around 800 of that is going to be cloud, but can you conceptually talk about the rest of it, 1.3 million? How much have you built in active non-subscribers still coming in at that stage, and how much of that is pirates potentially coming and potentially the growth of the legacy business coming up?
Yeah. Gal, where are you?
He's right there.
Oh, you're here.
Yeah.
Looking around like, I know Gal. I don't see him. The rest of the growth, by the way, comes from the factors that I talked about that are growing the core business at 14%, right? The overall market continues to expand. We will make inroads further into the 45% of the target accounts for EVAs. We'll continue to grow there. Of course, into the two you talked about, legacy and piracy. I'm reluctant to give you specific numbers for each of those three. All of those is what's contributed to a pretty steady state growth rate of the core business. 14% growth in subs from 2016 to 2017, 14% again from 2017 to 2018. It's a pretty consistent market at this point. All those trends that are driving it today will continue to drive it out through fiscal 2023.
To give you some context, by the time we enter fiscal year 2021, we would've been 4 years. Anybody who didn't come along with us during the end of professional would be in their fourth year of being. It's very difficult to operate in an ecosystem when you're 4 years behind the rest of the ecosystem you operate with. You're going to see those effects putting pressure on the users, where they're actually trying to get more current. We're also, what you saw is that in products capability that Steve just rolled out, that's completely new to our infrastructure as of this year. If you look at some of the other players in the space that have gone through similar transitions, they've actually started to see increased effectiveness from programs that are exactly like that.
We're not beneath copying what other people did successfully, and in some of the things we're doing with piracy conversion and things associated with that, we're really just taking best-in-class capabilities from other people. We expect to see the same kind of increase in efficiency from those activities as we move out beyond FY 2020.
Exactly. Just as a follow-up on the BIM market. Majority of design today started to convert from 2D to 3D, but there's still a bunch of people who still design in two dimensions. If you want to do BIM properly, you obviously need them to move as well. How far along are you, especially when you talk about not commercial, maybe residential, how much of that should happen? Is that a tailwind for your ARPS because people will have to upgrade to Revit as well?
Well, look, it's always good to have more of a market to grow into, and you're absolutely right, there's still a lot of 2D in the AEC space. That's changing significantly because of BIM mandates. We have a long runway in terms of growth of BIM data in the market, especially in the residential space, which is highly dominated by 2D processes. We've definitely got a long way to go. The good news is it's growing rapidly right now, and it's continuing to grow rapidly inside a particular geography and geographic expansion-wise, as more and more BIM mandates are put out there in the market and more and more companies are driving, more owners are mandating BIM. We've got years of growth with the Building Information Model.
The good news is that what we're doing with construction, remember, we have a long view that the model's going to be the currency. Our long view is complemented by a pragmatic view, where the tools we're rolling into construction take advantage of the 2D information as well.
Yeah.
We've got a long view about, well, you know what? Inevitably, everything's going to be model-based. That'll take quite a while to get there. It's great to be the purveyor of the model in that situation. Short-term, we still respect the drawings and the process where they need to be respected. We got lots of room to grow.
Did you want to add something to that, Lisa? I've got a comment. I'll tag onto it.
Yeah. Just to emphasize what Andrew said, I'll just give you a specific example when I was out visiting some customers. A lot of these contractors, they're global, they're competing globally. What some of them have told me is they actually can't compete and win on these bids because if they're competing with a 2D process and there's a BIM mandate, like the U.K. has a BIM mandate, you actually can't compete. It's impacting their business. I'm starting to see some of those customers are saying, "Look, we have no choice. We have to move, or else we're going to miss out on all of these projects that are being funded because there's these BIM mandates that are being driven by national governments.
What I was going to add is, at the end, set as a tailwind to Revit, perhaps. I do think as we look at the opportunity here, it's not just in cloud, right? As Andrew said, BIM runs through the entire process. That will drive more Revit as well, or AEC collection on the front end. I do think it's a tailwind across the board, and it's why I kind of made the comment that I'm not sure how much longer we keep saying, "Well, here's maintenance. Here's subscription plan, and here's cloud," because no one buys cloud, right? They buy, "I have got a construction issue. I need to use the BIM model." At some point, we're going to need to re-vector that.
Won't change the numbers, but just change the way we talk about it and the way we think about it, because the growth rate is pretty substantial in that market.
Well, Revit itself is a key element what's driving our core growth in sales.
That's right.
To Lisa's point, these mandates, they take a long time to get instantiated into the system. Then there's years of growth of everybody adopting it overall. There's a long runway of opportunity still here.
One question right before lunch. Where do you guys feel like you stand, this is Richard Davis at Canaccord, with regard to kind of tying in the cost and you build something, an engineer builds something, but how do you tie that into the cost? That's important both in manufacturing and in construction. In fact, that's kind of been one of the holy grails. Can you link back into the general ledger? If you could complete that last chunk, that would be pretty slick. Where are you now? Where do you think you'll be?
I'm going to let Scott comment on that.
Yes.
Reese.
Reese.
Got my name right.
yeah, Scott's finance team actually has done a really good job of partnering with our engineering teams to understand exactly what the types of capabilities we're building, what they cost. Some of these things, frankly, are compute intense, some are storage intense.
No, they're not talking about that. No, his question is their costs.
The customer's costs.
how do we tie the cost estimation?
Quantity takeoff, that kind of stuff. Yeah.
I see.
Resolution.
I see. Yeah. Sorry for misinterpreting the question. Yeah. Cost, that's another place where the BIM model plays directly into that planning process, right? That's where we've seen BIM 360 Design, what we previously called C4R, Collaboration for Revit, play a big, big role in helping make that BIM information accessible to more and more people. I think that plays a big role in driving the pre-construction or the planning process, which cost is a big piece of that. From there, obviously, it parlays into, okay, well, what are the actual costs out on the job site? That's why it's so critical to have the BIM information or the design intent drive that entire process. Because as soon as that becomes disconnected, you don't even have the design playing a role in what's going out on the job site itself.
That's why we believe taking a different approach than just a point tool, tackling cost alone, or any one of these pieces by itself, just won't work. That's why we think the BIM model has to play a role across the entire AEC process.
I'll add one more thing. Don't underestimate the power of Forge here. What we've done with Forge is we've already got third parties on the construction ERP side that integrate into our APIs and start extracting information. We've got a company that we work with that's doing a lot of work around quantity takeoff, and then connecting it to the construction ERP system to drive cost estimation into the construction companies see it as a competitive advantage to tie model data closer to their cost estimation tools. All right. We absolutely are seeing partners integrating through our APIs, doing some of this.
I think it's a little bit more complicated in manufacturing, but I'll tell you what we're able to do with manufacturing, and you've only gotten a slight flavor of it, but you'll get more flavor of it, is that because our generative algorithms are being connected to a particular method of making, we can generate geometry now that ties to a specific subtractive machine or a specific additive machine. You're actually able to estimate the quantity of material you need to subtract from or the amount of material you're going to add from, and the time it's going to take to do that. You can actually get very accurate costing information on a part level from some of these generative algorithms. That's going to make a big difference because it's intimately connected right back to the machine that's doing it.
When the geometry's done, you actually know exactly what you're going to cut into metal.
That is one of the most lubricating things we can do for the manufacturing customers in terms of giving them accurate costing information to feed into their processes.
Yeah, that's where generative plays a big role in manufacturing-
Sure
Is that if you say, "I'm going to sell this thing for less than $10," then you better know what that thing's going to cost you. That will impact which manufacturing technique that you are going to use. As you define the problem, you define your cost envelope, then that factors in and simulates and predicts what the thing's going to cost.
Thanks. Steve Koenig with Wedbush. Thanks for the presentations today. Very crisp, good layout of all the information. I want to ask one question in the weeds and then maybe elevate it with a follow-up question. Just more in the weeds here. You guys, you've been surprised a little bit on pricing and subscriber additions last several quarters. Your ARR has been very rock solid, however. Generalizing from the different things that have happened in the last few quarters, what has allowed you to keep that ARR steady despite surprises in your model on those other lines? Then looking forward, where could you be surprised? And in particular, I guess one little concern I would have would be on the, you're about to raise maintenance pricing 10% and then 20%.
I'm not hearing that you've assumed a lot more attrition, I guess that would worry me a little bit. Maybe, Scott, if you could address that. Then I do want to ask one follow-up on the product side, if you don't mind.
Okay.
Scott, why don't you go ahead?
Yeah. Sorry, you got to remind me of the first part of your question.
Yeah. ARR has been pretty rock solid.
ARR. Yeah, I got it now. Steve talked about this, actually, Steve, as he was going through his presentation, and it's something you've heard me say on the call, the machine, right? If I think of Steve's team as a machine that drives sales, that machine is built to deliver ACV as a proxy for ARR. The P times Q underneath that, yeah, they forecast that, and obviously, we use that to put into our guidance as we looked at it historically. That machine is built to deliver ACV and ARR. That's why we've been able to continue to deliver what I think is the most important metric in terms of ARR growth, even though some of the P times Q underneath the covers has not been what we expected.
By the way, as we look at this year, it's another thing that I'll hearken back to something else Steve said. Some of that's built into the way we pay our channel partners now. ACV will start to get built into some of their back-end incentive. You start to see that that's how we continue to deliver on the top line because that's what's built into the entire management system of Steve and the rest of the way down. Anything you want to add to that?
No, other than, I think some of the changes that you're referring to are also in the cloud part of the business that has a much lesser impact on overall ARR at this point in time anyway.
Good point.
We do have everybody thinking about driving ARR, first by delivering value to the customer, but then how it actually impacts ARR or annual contract value. We tell our folks, ARR is basically annual contract value and timing. Close the deal with the highest annual contract value and close it as early in the year as possible, and we get the most ARR out of it. That's what our partners are now doing, and that's what our sales teams have been doing.
I want to comment real quick on this surprise notion. There's really only been two, quote, "surprises" in the system, and they've all been positive surprises. I want to be very clear on this. One was, hey, these low-end cloud applications, seeding them, it's just not a productive effort. We're going to move away from that. Which, by the way, affected our ability to forecast what our cloud numbers, because we didn't know how many of those would attrit out post the seeding program. That was one. The other one was M2S has moved a lot faster than we expected, which, by the way, is a good thing. We're excited about it because the faster we move that base, the less we have to worry about churn in the maintenance base. Because of that, it's moved quicker than expected.
A phenomenon we always expected has become more material than we expected it to be at this point, which, by the way, was the collections upsell effect and what happened as a result when people would maintain the same number of users but have fewer subscriptions, which I illustrated on the call. These are good things. Moving away from these low-value seeding programs that distract us from selling the higher-value applications with our important and highly paid sales force is much more of a valuable activity, and accelerating M2S is absolutely a valuable activity. When we say surprises, they were positive surprises that accelerated effects we knew we were going to have. I want to be very clear about that so that we get that into place.
By the way, just so we get ahead of your comment and so you know what we are assuming, we do assume that there will be some additional maintenance churn associated with the price increases than we always have.
Yeah. What I'd just add to that, because that's what I was going to come back to, is we've got years of experience on price elasticity, really, which is what would drive that kind of behavior. We know with certain price increases, with a pretty high degree of fidelity, what impact it's going to have in terms of the Q and the P times Q equation. We have built that into our expectations both for this year and for next on the maintenance side. What's been surprising is what I've talked about, and I just want to go back to it quickly, the example on the earnings call. 42 subs that went down to 20, ARR went up 10%. That looks like 42 maintenance subs that were not renewed.
Yep.
Right?
Yep.
Thank you very much. One quick follow-up then on the product side. As you're
Getting Forge out into the market. Can you give us some color on what you might expect in terms of ability to monetize that through partners? Any thinking of how you would maybe go to market with partners on applications, et cetera? How can that evolve over time?
Scott, why don't you go ahead?
Yeah. Currently what we're doing today is we sell a Forge subscription, which gives a partner access to the platform, and then we partner with them closely to understand their usage patterns. But we see our relationship is really with the partner, the more that they sell, the more that they pay, right? It'll be a consumption business in the future. There are some really good models out there, and we're paying attention to those models. We're understanding the linkage to value. You would assume that the more that the partner sells, the more they use, and the more value they're delivering. We're just following those models, and right now, I would say we're in learning mode, to get to the end state of that model.
One thing. Oh, sorry. Go ahead.
Let us add, we talked about earlier that we're telling our partners there's a huge opportunity for them in services, in service delivery, adding value on top. Forge is absolutely one of the keys to them building out service value. We just talked a lot about this at OTC. We're guiding our partners on how to take Forge and how to start layering app value on top of that to solve specific customer problems and differentiate themselves in the market overall. By the way, some of the biggest developers on Forge are also our customers who are solving their own problems. Forge opens up a huge brand-new ecosystem for us overall.
I think you're going to see our best partners who you talk with, if they're not starting to talk about Forge, you ought to be asking them why they aren't, because that's something they absolutely need to be tapping into. Anyway, sorry, just wanted to add that.
Go ahead, Kent.
I think this is clear, but let's just be even more clear about what Forge is. You think about our 35-year history, we've accumulated a lot of intellectual property. In the past, the only way for you to interact with and glean value from that was through one of our applications. What Forge gives us the ability to do is to take that intellectual property and make it available as a service to anyone to do things that they were just simply never able to do in the past. We're really just unlocking a way for all of our customers and our partners to get at our intellectual property in ways that just simply weren't possible in the past. It's a big opportunity for our customers and partners as well.
Good. We have time for one more question.
Jay goes first, Cash goes last. Hey, guys, thank you so much.
Anything but Jay, my good friend.
You just want the last word, Cash.
When you look at the makeup of your subs, you've been predominantly pivoted towards A versus C, architecture, construction, and manufacturing, more towards design versus manufacturing. As I look at your long-term goal of 7 million subs, are you pivoting away and more towards the construction side and the manufacturing side? If so, what ASPs or ARPS do these two new products represent versus your average that you quoted? Secondly, and finally, if it's okay, the people that are not renewing maintenance and not buying the subscription, who are these people? Why? Thank you.
Kent, you want to take a crack at the first part of it, and then if you want to talk about strategy. As we talk about going after these new opportunities in construction and in manufacturing, we're not leaving the design market behind, right? The design space is what's driving that core. If you remember the money slide, the how we get there slide, remember, blue was still a big chunk of that bar, and blue was a big chunk of the growth from 20 to 23. That is largely driven by the design market. It's not an either/or, either do this or do that. We're going to continue to do what we're doing in the design markets, but then tap into these new opportunities for growth.
Now to answer your last question, right? First off, this is not a large percentage of the maintenance base that behaves this way or does this. Okay? To be very clear about it. We have a pretty good sense for how large it is. Typically what happens with these customers, they're either doing one of two things. They're either taking a gamble or they know how long they have to coast through their job. For instance, in some cases, they're just saying, "Look, you know what? Eh, I need this perpetual license for another five years.
After that I'm either retired or I'm going to be going on to something else, or that project's gone, and I'm not going to need it anymore." They just say, "Eh, I'm just going to zero out for that time." Or they're taking a gamble, which I would highly recommend they don't take, where they're saying, "Hey, I'll just drop off of maintenance now, because you know, that M2S price, they're just going to ratchet that up to the regular subscription price anyway, and they'll have a promo for me in three years and I'll be fine." Okay? Those customers, and that's a very small minority of customers, they've made a mistake. We try to talk to those customers and tell them, "Don't make that mistake," but sometimes they do, and that's a choice they make.
Actually make them say, "I know I'm giving up the opportunity.
We've done that, too. Are you aware? What we know is three years from now, it's not going to be their mistake, it's going to be ours.
It's something we did.
Because they're going to come back and say, "Well, I can't believe you don't have an offer for me." Well, we did. It was called Maintenance to Subscription. By the way, those people are still paying a lot less than normal subscription because it was an advantage path. People make those kind of gambles because they just anticipate something that happened in the past might happen again in the future.
Cash, the other kind of color I'd give you on that when you say, "Who are these people?" If you look at the demographic by customer size, how big are these guys, it's exactly what you'd expect. You see a higher either renewal or conversion rate with the biggest customers and lower with the smallest customers.
Good.
That concludes our formal presentation and Q&A. Again, welcome you to join us for lunch. Thanks.