Good day, ladies and gentlemen, and welcome to the Autodesk Q3 Fiscal 2016 Earnings Conference Call. At this time, all participants are in a listen-only mode. Later, we will conduct a question-and-answer session, and instructions will follow at that time. If anyone should require assistance during the conference, please press star then zero on your touch-tone phone. As a reminder, this conference is being recorded. I would like to introduce your host for today's conference, Mr. David Gennarelli, Senior Director of Investor Relations. Sir, please begin.
Thanks, operator. Good afternoon. Thank you for joining our conference call to discuss our results for the third quarter of Fiscal 2016. Also on the line is Carl Bass, our CEO, and Scott Herren, our CFO. Today's conference call is being broadcast live via webcast. In addition, a replay of the call will be available at autodesk.com/investor. As noted in our press release, we have published our prepared remarks on our website in advance of this call. Those remarks are intended to serve in place of extended formal comments, we will not repeat them on this call.
During the course of this conference call, we will make forward-looking statements regarding future events and the anticipated future performance of the company, such as our guidance for the fourth quarter and full year Fiscal 2016, our long-term financial model guidance, the factors we use to estimate our guidance, including currency headwinds, our transition to new business models, our market opportunities and strategies, and trends for various products, geographies, and industries. We caution you that such statements reflect our best judgment based on factors currently known to us and that actual events or results could differ materially.
Please refer to the documents we file from time to time with the SEC, specifically our Form 10-K for the Fiscal Year 2015, our Form 10-Q for the periods ended April 30 and July 31, 2015, our current reports on Form 8-K, including the Form 8-K filed with today's press release and prepared remarks. Those documents contain and identify important risks and other factors that may cause our actual results to differ from those contained in our forward-looking statements. Forward-looking statements made during the call are being made as of today. If this call is replayed or reviewed after today, the information presented during the call may not contain current or accurate information. Autodesk disclaims any obligation to update or revise any forward-looking statements.
We will provide guidance on today's call but will not provide any further guidance or updates on our performance during the quarter unless we do so in a public forum. During the call, we will also discuss non-GAAP financial measures. These non-GAAP measures are not prepared in accordance with generally accepted accounting principles. A reconciliation of our GAAP and non-GAAP results is provided in today's press release, prepared remarks, and on the investor relations section of our website. We will quote a number of numeric or growth changes as we discuss our financial performance, unless otherwise noted, each such reference represents a year-over-year comparison. Now, I would like to turn the call over to Carl.
Thanks, Dave, and good afternoon, everyone. At our investor day in late September, we laid out the vision and course for our business model transition. Overall, this quarter was in line with the plan we outlined. The performance of traditional P&L metrics like revenue, operating margin, and EPS, most of which were ahead of our expectations, were driven by slightly more perpetual license sales than expected. Metrics like billings and total subscription additions were below our expectations. We'll be in this hybrid model for the next few quarters. As a reminder, we'll stop selling new perpetual licenses for individual products on February 1st, and most of the rest of the products, including suites, on August 1st of next year. At that point, metrics like subscription additions and ARR will begin to be more predictable.
Most importantly, we are pleased with the total unit volume of licenses sold in the quarter, which was in line with historical ranges, and we're on track to add approximately 600,000 or more units for the fiscal year. This number is critically important for two reasons. First, it is the best indicator of the health of the business while customers have two different ways to buy our products. Secondly, it drives new subscription additions starting in Q3 when perpetual licenses are no longer offered. We were also pleased with the continued growth of our desktop subscription sales and growth in new model ARR. From an operational perspective, we continue to make critical investments. Infrastructure for the business model transition and our cloud platform and services remain the top priorities.
At the same time, we've been focused on managing our overall spend and creating increased efficiencies throughout the organization, which led to only 1% growth in total spend for the quarter. This is part of a much larger plan to ratchet down expenses as we move through the transition. One particular cost-saving action we have taken recently is our decision to exit the hardware business at the end of this quarter for our Creative Finishing products within our Media and Entertainment group. This is a small portion of our overall business but comes with substantially lower margins. It's really a win-win, as customers have been asking for hardware freedom, channel partners have wanted to sell it, and it will benefit our margins over time. Another area where we expect to achieve cost savings and margin benefits is with our eStore.
We just launched our new North American eStore this month after years of it being hosted by a third party. The new site creates a better customer experience and reduces the cost of these online transactions. This is important because we believe we can meaningfully increase the volume of business we do online. Finally, during Q3, we opportunistically took advantage of the low stock price to increase our stock buyback. We purchased 3.2 million shares at an average cost of $46.33, bringing our total buyback to 6.9 million shares year-to-date. Let me spend a little time on subscriptions. Once again, subscription additions were led by new model types, which increased over 130% and represented over half of the subscription additions. Desktop subscriptions by the new model type additions. The volume of desktop subscriptions coming through our channel partners continued to increase.
In Q3, channel volume of desktop subscriptions was over 60% of total desktop subs. That's up 44% in the third quarter last year. It's clear that our channel partners are actively engaged with us through this transition and are making desktop subscriptions an important part of their business mix. Looking ahead, we remain focused on driving new business, as well as converting the sizable base of non-subscribers to subscribers. Recall that in the second quarter, we discontinued selling perpetual licenses of LT in Australia and New Zealand, and the results were positive. We took that a step further in Q3 when we discontinued selling perpetual licenses of LT in the rest of Asia Pacific, with the exception of Japan. Once again, we experienced an increase in the unit volume of LT. It's another data point that gives us confidence going into the next few quarters.
Attach and renewal rates for maintenance subscription remained strong. Last year's highly successful upgrade program made comparisons for new maintenance subscriptions challenging. Total ARR increased 18% in constant currency, with maintenance ARR growing 9% and new model ARR more than doubling year-on-year at constant currency. Total recurring revenue increased to 56% of revenue, compared to 48% in the third quarter last year. Taking a closer look at our industry segments, let's start with AEC. We continue to win new business in AEC. One of the most prestigious architectural companies in the world just decided to standardize on Revit, the world's leading BIM software. This particular firm has been using competitive products for the past 20 years. We are encouraged by the continued adoption of our cloud-based AEC tools. BIM 360, our cloud-based software for construction management, continues to grow rapidly.
We recently signed an agreement with one of the top international contractors where BIM 360 won out over competitive on-premise software. BIM 360 is being established as the primary construction field reporting and management tool for this firm's construction business units. A360 collaboration for Revit also experienced strong sequential growth in subscribers. Turning to our manufacturing business. We continue to gain ground in all sectors, but particularly in automotive, aerospace, and industrial machinery. Inventor, Alias, and VRED continue to be very strong performers. Fusion 360, the first cloud-based 3D CAD system, experienced strong growth in adoption and usage. We clearly have the pole position in terms of bringing engineering software to the cloud. We're particularly encouraged by the use of Fusion in large companies switching from legacy systems such as SolidWorks.
We're very encouraged by strong adoption in enterprise accounts, even in industries that have been more cautious about use of cloud-based product design software. This comprehensive functionality built into a cloud-based platform has proven to be very compelling for companies looking for the next-generation product development software. The strength of our simulation portfolio was evidenced this quarter by competitive wins at several high-profile enterprise accounts that have long been strongholds for the competition. For these customers, our focus on simulation early in the design process and close alignment to design tools has proven to be a unique value driver. Our move to the cloud has proven to be a differentiated competitive advantage, especially with our SMB customers. We had another solid quarter with PLM 360. Speed and scale for deployment gives us competitive differentiation versus legacy on-premise solutions.
Both new logos and total PLM deals increased 50% in the quarter. We continue to have a steady stream of existing PLM customers come back to expand their original contracts. Going forward, we're excited about scaling this business, starting with its strong pipeline going into Q4. Turning back to our model transition. I want to revisit the main themes we spoke about at our Investor Day last quarter. We're now a little over two months away from a major milestone in our transition, the end of sale of perpetual license offerings for individual products. We're targeting a 20% compound annual growth rate in subscriptions and a 24% compound annual growth rate of ARR from the end of this fiscal year to the end of FY 2020. As we indicated, the model hinges on growth in subscriptions.
It's helpful to talk about what we expect to be the four primary drivers of subscriptions. First, starting in the third quarter of next year, virtually every license that we sell will be a subscription as we end of sale all new perpetual licenses. As you may recall from our disclosures a few years ago, the typical run rate of licensed unit volume has ranged from 500,000 to nearly 800,000 per year. On top of typical run rate of licenses, we are seeing a multiplier as customers move from perpetual to subscription. This multiplier ranges anywhere from a 10%-30% uplift in units. The subscription price point makes sense for them to purchase additional seats.
Second is converting a minimum of 30% of the 2.8 million users out there who are active but not on subscription. The third piece is capturing a larger share of casual pirates. Finally, the TAM expansion of cloud-based products and services, which brings in new users. The 20% subs CAGR will drive the 24% ARR CAGR. We also talked about a 3% CAGR in annualized revenue per subscription. That number is heavily influenced by the mix of subscriptions. LT and cloud subs have a lower ARPS relative to our other offerings and will depress total ARPS but are great for total ARR growth. I also want to be clear that these CAGRs anchor on the end of FY 2016. We expect some bumpiness over the next three quarters as we end the sale of new perpetual licenses.
This does not change our conviction or our assumptions or the end state of this transition one bit. In fact, the early proof points around subscription adoptions have been very encouraging. I alluded to this at the beginning of the call, I'll reiterate here. Our business model transition will not be perfectly linear, the amount of business that we will transition, the number of subscription additions, the mix of subscription additions will fluctuate. The fourth quarter is always the biggest quarter of the year for us. Gauging the level of activity around customers buying the last perpetual licenses for individual products is particularly difficult to project. As such, we've taken a more conservative view and reduced our billings and subscriptions outlook for the remainder of the year. This is partially related to our Q3 results and partially related to our revised Q4 outlook.
We're also factoring in about a $10 million reduction in billings related to exiting the Creative Finishing hardware business. Global economic conditions remain uneven, with strong headwinds continuing in most emerging markets. FX headwinds remain persistent, they haven't gotten much worse than the first three quarters of the year. To wrap things up, we're really excited to be one step further along and maintain our strong conviction in the model transition. We're heading down a path that will provide our customers with greater flexibility and a better user experience while creating a more predictable, recurring, and profitable business for Autodesk in the years to come. Operator, we're now ready to take questions.
Thank you. Ladies and gentlemen, at this time, if you have a question, please press the star then the number one key on your touch-tone phone. If your question has been answered or you wish to remove yourself from queue, you can do so by pressing the pound key. If you do have a question, please press star then one. We also ask that you limit your question to one initial and one follow-up. Our first question is from Keith Weiss of Morgan Stanley. Your line is open.
Hey, guys. Good afternoon. This is actually Stan Zlotnik sitting in for Keith. Thank you for taking my question. Wanted to kick it off with the long-term model. You're saying that nothing has changed, we're essentially taking down subscriber targets for this year, we're lowering the starting point. Does that implicitly say that you are saying that the growth rate in subs will be unchanged, the endpoint will be different?
Hi, Stan. It's actually a really small amount we're taking it down if you project it out over years. This is just a tiny adjustment from where we started. The important thing to look at here is to remember that what we talk about here in those numbers is net subscriber adds. The number of subscriptions we added was in keeping with exactly what we wanted. The net was different. As I expect there'll be a bunch of questions, I'll go into this more deeply. To answer your question directly, yeah, the net's down and for a number of reasons. In general, we were right on track, and that's why I gave the unit numbers about licenses so you could gauge it.
If you put this over the total base, the number really makes no difference in the model as you project it out any number of years.
Okay, fair enough.
We're talking about tens of thousands versus millions.
Got it. Does that mean that you saw slightly higher churn in the quarter than you were expecting? That's it for me. Thank you.
Yeah. I didn't even get to the next questioner before we got more questions about it, so we might as well just walk through it. The first thing is, let's just remember that the 80,000 is net adds. The unit volume was nearly twice that in terms of subscriptions. There are two things that go to it. One is a maintenance base that is now very big, approximately 2.1 million subscribers, and a renewal rate that's less than 100%. Any small change there, or any change in timing, or any change in buying behavior, even at a small amount, changes the reading at this moment in time when we measure it. The other thing is the attach rates are less than the renewal rate, as we've disclosed over the years.
A little less influential, but it did come from a little bit from attach rate and a little bit from churn in the maintenance base on the existing model.
Okay. Thank you, guys.
Let me just finish on the other part of it is, if you don't do the net of it, subscription adds were just where we anticipated they would be.
Thank you. Our next question from Saket Kalia of Barclays. Your line is open.
Hi, guys. Thanks for taking my questions here. Might as well stay on the subscription line of questioning here, Carl. The reduction for the year, I guess that 310-330 versus the 375-425, maybe you can walk us through where the lion's share of that reduction is coming from. Is that coming from, to your point, that lower attach and renew rate that maybe you expected in the maintenance base? Or is the new model base maybe a little bit less than what you expected as you wrap up the year?
It's about 50/50. At least this quarter was. Once again, I try to highlight it in my remarks. I can tell you exactly what happened in Q3, and in that case, it's 50/50. If you look going forward, because of the difficulty in projecting our customer behavior precisely about whether they'll choose to stock up on perpetual licenses or move to desktop subscription, it's a little bit harder to project. We just thought we'd be a little bit more cautious and conservative, particularly with the nets. I want to draw two pictures. We're more cautious about the net adds. We gave you the other facts and figures to show really that the unit volume is steady and that the demand is high.
I don't want to confuse those two issues, but I can tell you Q3, really 50/50, Q4 and going into next year, a little bit harder to be precise about.
Yeah, Saket, if what you're trying to bridge is the midpoint of the previous guidance for the full year to the midpoint of the current guidance for the full year, what I'd say is that a big chunk of that is where we see ourselves coming out of Q3. If you just add the numbers up this year, we've added about 236,000 net subs. When you peel back how that compared to where we thought we'd be at this point, the bigger driver of the delta is what you said. It's on the maintenance side versus on the new model side. The new model growth has been good. We continue to see it looking strong, particularly year-over-year, but it continues to grow strongly sequentially as well. When you look at Q4, I think what Carl said is spot on.
What we're trying to do is take a somewhat conservative view of what we think the net adds will be. There's a significant opportunity as we hit the first milestone or the next milestone in our business model transition and end-of-sale perpetual licenses for individual products, and trying to gauge what that does in terms of driving additional demand relative to a very large maintenance base. If you go back several years, Q4 has always been our biggest quarter for perpetual license sales, which means it's also our biggest quarter for renewals. There's an enormous maintenance renewal amount that gets under in the quarter, and it's balancing those two off. That's what led us to where we are.
Got it. Then the follow-up, maybe if we shift over to the Creative Finishing part of the business. Carl, I think you said roughly a $10 million impact, I think, in the fourth quarter on billing. Is it fair to think about that as maybe a $40 million a year revenue run rate? What's the sort of gross margin on that you're now going to be sort of offloading?
Yeah. Saket, that's right. To be clear, what we did in Creative Finishing, to this point, we sold that product on hardware. It's software license, and then we built it, we actually packaged it on the hardware and shipped it out. What we did in Q4, or at the end of Q3 is, based on really feedback from partners and customers that wanted the flexibility to use their own hardware, is converted Creative Finishing software to a software-only offering. What that pulled out of our revenue stream is the hardware that was effectively a pass-through. We bought it, packaged the software on it, and passed it through. As you would imagine, it was pretty low-margin business.
Now, Q4 is the biggest quarter for that as well, I wouldn't take the $10 million that we see the impact in Q4 of this year and just multiply it by four, but it's somewhere between $25 million-$40 million is what would come out. Low-margin business, but all hardware-only markup.
Got it. Very helpful. That's it for me. Thanks.
Thanks. Great.
Thank you. Our next question is from Heather Bellini of Goldman Sachs. Your line is open.
Hi, this is Shateel Alam in for Heather Bellini. Just one question on maintenance and what you're seeing there. You added 33,000 maintenance subscribers, but ARR stayed flat. Is that what you're expecting? Could you comment on the attach rate you're seeing on maintenance versus prior quarters? For next quarter, with license ending, are you expecting a strong maintenance ARR uptick while this is the last chance to buy standalone licenses?
You asked a lot of questions, Shateel. Let me jump in, and if I miss one, remind me. The first question, 33,000 net new maintenance adds for the quarter, but billings being effectively flat sequentially. I think the thing you got to bear in mind is that's 33,000 on a base of about 2.1 million. The growth of the overall base, which of course is what drives that, is not nearly as big as 33,000 would sound. On the attach rates, what we are seeing, though, is we continue to see attach rates growing. It still is below the overall renewal rate for maintenance, but the attach rate's growing. It's growing on LT, which is still a step function below the other products, but it's growing on the other products as well. Renewals continue to be strong through the quarter.
We're not seeing any significant change there.
For 4Q, are you expecting an uptick in maintenance ARR?
Yeah. The hard part about calling that is calling the number of customers during Q4 that actually take advantage of the last chance to buy perpetual license. To the extent that that's a big number, that a lot of people look at this and say, "Hey, I want to buy my perpetual license while I can." With the attach rates we're having, that will drive a significant uptick in maintenance ARR to the extent that we continue to see what we're beginning to see is a lot of customers saying, "Okay, I've now gotten comfortable with desktop subscription." We see it growing at substantial rates year-on-year, but growing sequentially as well. To the extent that more people say, "I'm ready to make that shift," that would again drive desktop subscriptions, but only when they actually need it. There's no end of sale on desktop subs.
Better news for us long term, the more people that accept the shift over to desktop subs, but there's no compelling event in Q4 for that set of customers. It's really trying to predict the mix of people that want to take advantage versus the mix of people who say, "No, I'm ready to make the shift.
Got it. Thank you.
Thank you. Our next question is from Matthew Hedberg of RBC Capital Markets. Your line is open.
Yeah. Hi, it's Daniel Bergstrom for Matthew Hedberg. Could you talk a little bit about bringing the eStore in-house? How does that enable you to increase the volume of the business you do online? Thanks.
Yeah. I think there are two different dynamics going on. I think customers, particularly small customers, are choosing to buy more and more online as they get more comfortable with it. The factor that we get to influence is the experience. Outsourcing it has its limitations in terms of the experience you give your customers in buying and maintaining their things online. We felt like we had grown that business to a point where we needed to control it ourselves. Second important thing about it is we felt like we were paying too much money for the service to do that. We wanted to bring it in so we could lower the cost of the transactions and improve the experience.
Thank you. Our next question is from Sterling Auty of JP Morgan. Your line is open.
Yeah. Thanks. Carl, you had an activist investor actually file recently. I'm curious if you had any conversations with them, what's going to be your approach? Will you be open to discussions and to suggestions they might have to improve shareholder value?
Yeah. I probably found out about the filing the same way you did. I happened to be in Japan at the time, somebody sent me a news article. When I saw that, part of the filing was that Sachem, I'm not sure if I'm pronouncing it correctly, Sachem had filed. I reached out. I spoke with Scott. We didn't say anything of substance, we agreed to get together. Just as I've done with many investors. We continue to be open to talk to all of our investors, whether they file or not.
Got it. Back to, Scott, I hate to do this to you, but back to maintenance. I guess I wasn't clear. It sounded like you said that attach rates are continuing to improve and renewal rates are steady. I thought in an earlier comment that when we look at the change in the net subscription expectations for the year, that the maintenance was definitely 50% of the component. I guess I'm still missing which component is coming in less than what you expected. Is it the churn or is it the attach?
No, it's actually neither. In order to attach a maintenance agreement, you need a new licensed sale. When we looked at new license sales for the year this year as we were laying out the plan, it's a little bit like doing a channel check where you look at a small sample size and try and extrapolate to the rest of the world. What we had based our expectations on the end of sale of perpetual and what that would do in terms of driving new license sales is we based it on some of the tests that we had done, and we based it on the experience we had with the end of sale of upgrades. What we've seen is attach rates continue to be strong, and strengthening, and renewal rates have stayed strong for maintenance.
The number of people who are buying that last perpetual license are less than we expected when we set out the plan for the year.
That makes sense. Maybe just to put a finer point on that, do you think that's macro, or do you think it's still understanding the model transition or some other factor?
I think there's two things going on. I think macro is relatively steady. We have not seen a dramatic change. I know this will come up, we might as well just cover it now. The Americas, U.S. in particular, continues to be relatively strong, as does Central and Northern Europe. We see weakness in Brazil and Russia and the Middle East. China, I actually feel is slightly stronger than I think generally people are giving it credit for. Japan is weaker than I think anybody would want. I think that's the roundup of what the macro looks like. In general, I think macro is not a huge factor. It's neither particularly positive nor particularly negative. FX continues to be negative as it is for everybody, and it looks like it might continue for a while.
That's kind of the roundup of what I would see about macro. I don't think the competitive situation has changed very much in the mainstream businesses. There's different dynamics in some of our cloud businesses, where we're kind of leading the way. Other than that, not a huge amount of change.
Okay, thank you, guys.
You're welcome.
Thank you. Our next question is from Brent Thill of UBS. Your line is open.
Hey, Carl. Just on the channel, many of the channel partners we speak with, it seems like there's a divergence. The larger ones get it. They want to be on board. They understand the power of it. The smaller ones are still struggling, which you would expect given this. I guess just, do you feel like you've got the channel guys now really ready to go? They're on board. Can you just give us a sense of how that education process is going at a high level?
Yes.
Just a follow-up, too, is on the deal you mentioned, nice to see a 20-year competitive switch. What drove that in terms of their decision-making? Do you see other deals like that in the queue that are ready because of either an architectural limitation? Just give us a little more color would be helpful.
First channel partners, then the switch. If you look at channel partners. Let's just back up for a second and maybe have a little discourse on channel partners and checking with channel partners. The first thing, just to remind everybody, is if you look at it broadly, 30% of our business, 20%-30% is the very top tier less than a tenth of a percent of our customers. At the so-called bottom of the pyramid, people going through lower-cost channels, like the eStore, that's 20%-30% of our business. When you get to the middle part of our business, that is our traditional good partners, one of the things also is that business is geographically aligned like the rest of our revenue, with about 25% in the U.S. and 75% outside.
It's one of the reasons, just on a kind of an aside, why often the channel checks don't correlate as well. That if you're only starting with 40% of the business or 50% of the business, and you talk to a sampling of 25%, you're sampling 12.5% of the people that you would like to. Overall, we have more than 2,000 channel partners. Use that as a backdrop to this. I would totally agree with what you're saying, Brent. I think the larger partners understand the transition better, have moved more quickly, have prepared their business, as you would expect, but I wouldn't say it's uniform. There are also smaller partners who are some of our best partners, who have understood this, who have always had really healthy businesses that provide great services. If I look by and large, I think your characterization is correct.
If you look at it from the business point of view, we gave you all kinds of metrics during my remarks about why the channel partners are selling it. If anything, I would say our channel partners, the traditional channel partners, are more enthusiastic than desktop subscription than we originally planned. We knew from the beginning that this is a tough transition for cash flow-sensitive businesses. We've tried to do our best to ease it. We gave plenty of warning. When others were urging us to make the transition happen faster, we thought this was the prudent thing to do in not only getting a company with our kind of balance sheet through it, but most importantly, getting those 2,000 channel partners through it. Having said that, I don't think every channel partner will make it through in an identical form. There will probably be some consolidation out there.
There will be some change in business, slightly more than the natural churn that we see there. I would just generally say, channel partners get it. They have gotten on. It reminds me a little bit when we first brought out subscription, and as I listen to the names who have asked questions so far, many of you were there when we did that. Our channel partners thought subscription, if you had listened to them in the moment, was the end of their business. I think almost everyone you ask today, they would say it is the best part of their business. I think in the same way, we have been very considerate and deliberate in how we move not only our business, but an important part of moving their business as well. The second thing was, it was really gratifying. This was a very prestigious firm.
What has happened around the world where we have a strong market position in architecture, engineering, and construction software. People don't realize there are many competitive products. There are strong local products. There are many reasons why people choose other things. We had been working on the periphery with this firm for a while. The thing that put them over the edge is really how prevalent Building Information Modeling, and particularly as Revit as the tool of choice is for Building Information Modeling. Let me just digress for a second. People don't realize this about our business sometimes. While we are in a very complex technology business, technology adoption doesn't happen overnight. For those who may have forgotten, we bought Revit back in 2002, I believe. We are 14 years later, and we still think there is lots of market penetration.
If you were to look at a graph of the revenue, the first five years would not have impressed you. It is really gratifying to see how long this is, and that at this point in time, we are winning the business of some of the best and most creative firms in the world. I expect that trend to continue. As you have seen, it is not only in the architectural space, but in the construction space, we continue to gain market share. We think it is an important part of what we are doing, investing in our cloud-based businesses, where really the future of this industry is.
Thanks, Carl.
Yeah, you're welcome, Brent.
Thank you. Our next question is from Philip Winslow of Credit Suisse. Your line is open.
Hi. Thanks for taking my question. Carl, I know you touched from a geographic perspective on macro, when you look at the industry verticals that you guys service, commercial construction, civil, the different parts of manufacturing, and obviously the media side. When you look by vertical, is there any sort of tone change across those vertical industries? Who hasn't changed, who maybe is changing, and just how you're kind of contemplating that or thinking about this over the next few quarters here, what you're keeping your eyes on?
Yeah. It's surprising little change in the business environment around us. I don't see dramatic changes. I think there's a lot of kicking and complaining about interest rates, even though I think it's fully baked into the market, and for the most part. People doing things like development projects or capital improvements in manufacturing have figured this in for a while. Maybe it's gotten slightly more certain. Surprisingly, in our end markets, we're not seeing huge amounts of change, either geographically, or by industry. It's kind of steady as she goes. Certainly sitting here in San Francisco, I have a little bit of concern about bubble valuations, in down round IPOs. That may be a much more local phenomena, and I wouldn't extrapolate from there. That feels to me like a bad channel check. I would just say, business is pretty steady as she goes.
We're monitoring it. I think the other indicator for me was the demand that we talked about in the total unit volume, kind of indicates to me the same thing. No big change, Scott?
Yeah, Phil, I was going to say the same thing. When you look at the underlying demand, in other words, the units that we sold during the quarter, it's right in line with what our expectations were and where it's been historically. If you slice it the other way by industry vertical, and I'm sure we've got this in the prepared remarks, the revenue growth rates in AEC were around 4%, and in manufacturing it was the same, about 3%-4%. We're not really seeing any bifurcation by vertical at this point in here.
Got it. Actually, just a quick follow-up for Scott. Carl mentioned that, as he did back at the Analyst Day a couple of years ago, the 500,000-800,000 license units. Wondering if you could give us a sense for sort of just what last year was or maybe just the past couple of years, if you have that, just so we can kind of get a sense for, because that 500,000-800,000 range is pretty wide, sort of what the most recent cadence was like.
Sure. Last year would've been toward the higher end of that because we had a very successful upgrade set of sales last year. It's a bit of an anomaly. It definitely would've been at the high end of that range. I'd say if you go back before that, and really before we, in earnest, got into the business model transition, it was right around the midpoint of that range, between 500 and 800, and that's what we're seeing again this year, just in terms of net demand. In other words, net sales into the market, right about the midpoint.
Got it. Thanks, guys.
Thanks, Phil.
Thank you. Our next question's from Gregg Moskowitz of Cowen and Company. Your line is open.
Okay. Thank you very much. Carl, wondering if you could just give us a little color on the sort of activity, meaning perpetual and subscription, that you saw this quarter in Asia-Pac, ex-Japan, and ANZ. How did that change as the quarter progressed and as the deadline for buying perpetual license approached?
Yeah, Greg. We saw what you'd expect to see. To be clear, we didn't end of sale all perpetual licenses in APAC this quarter, ex-Japan. It was only LT, only the low end. Very similar to what we had done in Australia and New Zealand in the prior quarter. We saw the same thing that you'd expect. We saw an increase in the volume of LT sales as we approached that end of sale. It was right in line with our expectations.
One thing I'd say we're seeing with these end of sales is they are definitely more back-end loaded in the quarter than our traditional run rate. It makes it a little bit more challenging in forecasting. I used to talk about, or almost brag about, you could look at the first three weeks, and there was a statistical phenomenon to predict week 13. These promotions, and I expect them to continue through the next three quarters, and possibly four, are just a little bit harder because we really are creating these events, and it's just human nature not to act until the last moment.
Okay. That makes sense. Just for Scott, Carl made a comment that you're looking to ratchet down expenses as you move through the transition, exiting the hardware business for Creative Finishing being an example of that. Are you still expecting a spending increase of 5%-6% per year going forward, or has your view on that changed to some degree? Thanks.
Yeah, Greg, I'll start and then actually let Carl jump in on this too. If you look at the quarter we just announced, the total spend growth was 1%. If you looked at just OpEx, because COGS is obviously going up as we build out the cloud infrastructure. If you just narrow it in from total spend to OpEx, the quarter we just announced was flat in OpEx year-on-year. If you look at the guidance we've given for Q4 and back into the spend growth that would equate to, it's in that same 1%-2% range in Q4. We have really put a lot of focus into effectively managing down spend and becoming more efficient across the board. Carl, you want to?
Yeah. I'd say two things. One is the plan is to continue to ratchet it down going into next year. There's a little bit of the effect of currency, but on a constant currency basis, I'd say it would be coming down into the range we mentioned. The other thing was, it was definitely a change, and certainly a change from Investor Day. Truthfully, we've been speaking a lot over the last six months with many of our long-term constructive investors, and they totally understand and certainly believe in the model transition and the strength of the franchise. There was definitely a hankering for greater cost control. We listened carefully, and you saw, we put plans into place earlier. We did something in Q3, doing the same kind of thing in Q4. Some things take a little longer to put in flight.
It's a combination of ongoing operating expenses and then structural things like we did with the eStore. Which again, ironically, sometimes take an investment in order to do that. That one in particular is an investment in our IT infrastructure to do it. In the long run, it pans out, but some of the fluctuation you see in costs are due to things like that have a benefit over a longer time horizon. We're continuing to make some of those changes. Let's just back up from the expense question a little bit to look at, there are really two things that we're doing in parallel right now. In each of those cases, we're running dual models. On the business model, we really have business systems from front to back end that are for the old traditional model.
We have been building out a complete set of front end and back end to service new customers in new ways. On the technology side, while most of our revenue is on traditional Windows-based machines with installed software, by far the fastest-growing and certainly the future of the company is on cloud computing. We've been building out not only infrastructure, but products and services for that, and we strongly believe in the future there. The reason why we have confidence to not only do it during the transition but beyond, is that some of those costs will obviously go away. They are necessary now, but they become redundant over time. We took strong action to do it, and we'll be doing more and talking about it more. For now, that's probably enough.
Very helpful. Thank you.
Sure.
Thank you. Our next question is from Brendan Barnicle of Pacific Crest Securities. Your line is open.
Thanks so much. Carl, you know you should never let us know it's not, or you're not going to tell us enough, I'm going to have to ask more about this expense piece. I was still modeling, I think a lot of people were still kind of at 5% that you guys have been talking about. As we think about more of a 1%-2%, where should we be looking for that leverage to come from?
As we talked about before, we think there are a couple areas. In general, I would look across the board and I would look at if you peel that back, what you would see is some of the investments we made, eStore is like a perfect example of where those were expenses that were in there. Now that we're rolling it out, we will see the benefit of it in lower cost structure for people who transact that way. We've done it in North America. We will continue to do it in other geos over time and lower the cost structure there. As these places where we made investments for the health of the future of the business come to fruition, we'll start doing it. In some ways, looking at it by function, I know that's the convenient way for you, is easier.
In some ways, it's chunkier because a lot of it is around initiatives, whether it's the Creative Finishing hardware, what we're doing there. The bulk of it will come from redundant stuff we're doing as we're able to move more activity to the new models in both the business transition and the technology transition and invest less in the traditional things. That's all I have to say on the subject, Brendan. You need more, I do not.
Well, thanks for helping us out.
Okay.
At the Analyst Day, you pointed out the, I think it was like 40% of desktop subscribers were new to Autodesk. I was wondering if you had any new stats on kind of those net new customers that you've been picking up in the model transition.
No, Brendan, we're not seeing any significant change in that yet. The one update that I would give, and Carl actually had it referenced at least in his opening commentary. If you remember in that same set of slides that Steven Blum went through, he talked about the percentage of our total desktop sales that are being driven by the channel, and we showed what that trend had looked like over the last six quarters. We did see that, and it kind of gets at the question earlier that Phil had. Brent asked about the channel. We are seeing that continue to grow. If you remember that slide, we said at the end of Q2, 58% of our desktop subscription sales came through the channel. The quarter we just closed, that was up about two to three more points.
It's the majority of our desktop sales now that are coming through channel partners, and it's continuing to increase.
Are those channel partners generally bringing you net new Autodesk customers?
It's a mix of both. I mean, it also.
That maybe kind of could be in your install base.
It also varies by product, right? You remember we also went through for some of the cloud products, Fusion in particular, the statistic of new customers that product, of all the customers they brought in, what % are new, I think it was north of 85% on Fusion. It's less whether it's coming from the channel or is it just desktop, and it's more are these products actually getting to increasing the TAM that Amar talked about, and really getting to a new customer set. There hasn't been any change from all the data that we just showed you about six weeks ago, but we are continuing to see a very positive trend of new customers coming in.
Yeah. We rightfully spent a lot of time on things like subscriptions, but let me just take a moment here and go through where these new customers are, because I think sometimes we lose track of the health of the business is going to be about these new customers. When you look at most of our cloud-based offerings, the first thing I'd say is that we're way ahead of the competition. The customers are coming along quickly. We might've been ahead of the customer slightly, but not for very long. When you look at things like BIM 360, these are either brand-new customers to Autodesk or new parts of a company that had no use for our other kinds of tools. If you look at things like PLM 360, almost entirely, these are new customers to Autodesk.
Our new Internet of Things, the IoT product, is almost all new customers to Autodesk. As Scott referenced, and we've talked about repeatedly, Fusion is clearly one where we're winning market share from existing competitors. Those are all places that are bringing in new customers, and I expect that trend to accelerate as we continue to gain ground and the world moves towards more cloud-based tools.
Well, Carl, I totally agree. That's where we want to focus on. I'm wondering if there's any way we could get some sort of quarterly metric around that, just to reinforce that point, because that was such a compelling one at the Analyst Day, and certainly one we'd love to continue to hear more about.
Damn, I never should've gotten into that speech, should I? Yeah. Well, let us go back and figure out how we can best convey what we're doing. I think both in two things. Tell me if this would be helpful, is the number of units and something about the percentage of new customers.
That would be, I think that'd be very valuable. Yeah.
Okay. Thanks.
Thanks, guys.
Our next question is from Jay Vleeschhouwer of Griffin Securities. Your line is open.
Thanks. Good evening. Carl, I'd like to ask a question that gets to both the cost structure of the company and the saleability of the products. At the analyst meeting, you mentioned that one of your opportunities is dealing with the complexity of the portfolio. The question I'd like to ask is, how are you thinking or planning around simplifying the portfolio, given the large number of SKUs that you have, particularly in terms of suites, just as a matter of improving sales efficiency and cost, versus, on the other hand, your need to maintain some kind of segmentation of the product line, particularly for the different verticals?
Yeah. It's a place where we've been spending a lot of time thinking and planning for it because you're absolutely right. Complexity and particularly unintended complexity drives costs. This is a huge opportunity for us to simplify and reduce the costs. It also makes for a better experience for our customers. We think we can make our customers get more value and pay us more. We've talked about this a lot with our enterprise customers. When we've given them a more flexible licensing mechanism, they've used more software. They paid us more per year, and I would say, at least anecdotally, they're happier about it because they really understand where the use is. I think in the last two investor days, we've talked about that dynamic.
The biggest thing in my mind we can do is extend that flexibility that we now have really only sold directly by us to our largest customers, to the bulk of our customers in the part that's serviced by our traditional channel partners. We really want our mid-tier customers, the small and medium businesses, to get that same kind of flexibility. You're absolutely right, Jay. Handing them all of our products is not particularly helpful, but also breaking it down into 10,000 SKUs is not helpful either. We're looking at the minimal set that we need so that they can pay an appropriate price and get the set of tools that they might need.
In addition, what we're going to do with that is layering cloud services so that they can really pay on-demand for the things they may or may not need. It's really about that flexibility, which is where we feel the way to simplify it and to make it a better experience for them and to make it more lucrative for us.
To the second part of your question there, if you're getting at, are we still going to have a differentiation between AEC and manufacturing and M&E, that's clearly part of the thinking. Andrew was getting a little bit of this in his presentation at Investor Day of how do we simplify that but continue to have something that is targeted and makes sense. For a use case in the various verticals. We'll continue to maintain that at some level, but greatly simplify the way we get it to them.
Sooner or later, and hopefully sooner, you'll deploy what would be in effect a Creative Cloud-like sign-on for the SMB market, with what you've been calling anyway to date, Subscribe to Autodesk. You'd have to come up with a better name than that, but is that a FY 2017 objective still?
First of all, Jay, we appreciate the suggestions that you've published, and we've rejected them all, first of all.
Thank you.
Yes, stick to your day job. Yes, it's an important thing for us to offer, and we will be offering it. Whether it stays that name or not, stay tuned. Yeah, it really is that flexible offering. Remember, it's not only just simplifying and the cost structure. This is slightly more complicated, but the question has come up, what do you do about the 2.1 million users, and do they move to desktop subscription? I'll reiterate here, those 2.1 million customers have historically been our best customers. They still are our best customers today. We think the way to move them over, so that they get more value and pay more, is to move them to flexible license offerings at their discretion. We have a number of knobs and dials to do that.
What we really believe is the essence of it is having a more compelling offering. Everything we've been shown to date is some combination of access to more desktop products and access to the cloud services is the thing that is compelling. You'll hear more about it. We'll introduce it. At this point, I feel like we have enough moving parts in the model that we don't need to introduce a lot more. You will hear more about it during the year.
Lastly, if I may, in the future of Autodesk land, once the model gets steady state and so forth, what do you think the role of your rebates, promotions, TLOs, and so forth will be? Do you think you'll have to employ those kinds of levers as often as you've done in the past under the old model?
I have a point of view. It is not widely shared by my executive management team. I feel like promotions are important for one reason, and that is to create a compelling event for people to choose to buy. For the most part, us as well as our competitors, we sell on the value of the offerings, the quality of the offerings, the suitability offerings, the completeness of the offerings is what makes people buy it. When they buy is influenced by everybody likes a deal. If you look around the world, the number of companies who never sell with a discount is tiny in the technology business and might be nonexistent with the exception of Apple. I see it as being a way to control the timing.
Of course, you want to minimize this as much as possible and only do it to create compelling events. We have a disciplined approach to promotions. They're centrally controlled. We're trying to influence behavior. We're trying to influence behavior not only of our customers, but of our partners. I think it is one of the tools we have. I would be reluctant to give it up. You need to use it judiciously. There's an art to using it. I think the more dogmatic point of view in which there's never a promotion and you only sell at this price, doesn't fit reality. I think it takes one of the valuable tools out of your hand.
Thank you.
You're welcome, Jay.
Thank you. At this time, I'd like to turn the call over to David Gennarelli for any closing comments.
Thanks, operator. Well, that concludes our call today. We also have our Autodesk University event coming up on December 1st. If you'd like to attend that and you're not signed up yet, please email or call me. We'll also be at the Credit Suisse conference on December 2nd and the Barclays Conference on December 9th. You can reach me at 415-507-6033. Thank you.
Ladies and gentlemen, thank you for your participation in today's conference. This concludes your program. You may now disconnect. Everyone have a great day.