Adobe Inc. (ADBE) Earnings Call Transcript

All right. Welcome, everyone. Good morning, and thank you for taking some time out of a busy earnings week to spend some time with us. I'm with VP of Investor Relations at Adobe, Jonathan Vaas. I'm Matt Swanson. I'm one of the software analysts here at RBC, and just a big welcome to Jonathan.

Thanks. It's always a pleasure hanging out with you, Matt. I'm glad to be doing this, and thanks, everyone, who's tuning in.

Yes. So I think we're going to spend a lot of time today on GenAI, on the product set and a lot of the things you're doing, but from maybe a wider industry perspective right now, it seems like the 2 big topics are the GenAI narrative and the macro sentiment and kind of how those are interrelated. Maybe to start off, can we just talk a little bit about what you guys are seeing from either the macro or the demand environment?

Yes, for sure. It's -- that's been a dynamic one, and it's one where every company pays a lot of attention to what is everybody else saying. And then sometimes you hear from other companies, saying what you've been saying. So I'll just remind folks of what Adobe is kind of -- how we've characterized the macro in Q1, how we characterized it in Q2. I mean it's definitely been an interesting ride, I think, really since -- gosh, I mean, going all the way back to COVID but certainly since 2022. Enterprise software sort of has gone into this phase of when enterprises' growth rates have slowed, they've really been more focused on profitability. And so we've talked about scrutiny around spend, scrutiny around budgets. I think right now, ROI is everything to enterprises, so they're going to be spending on things that have a really good ROI case to not only drive their top line but also produce efficiency. So we've talked about that budget scrutiny for, I want to say, a couple of years now, and that's something we've heard from other companies. Adobe always, when we talk about the enterprise, talks about how we're really diversified -- and when we talk about diversification, we mean geographically. We mean in business models. We mean in customer segments. So there's -- we look at what the pure-play enterprise companies are saying. We look at what some pure-play SMB or consumer companies are saying. Adobe, what's -- makes us resilient as a company is we play across all those segments. So we're kind of able to use a data-driven approach to lean into strength, to lean away from weakness with how we invest and where we focus. Q1, I would say, for Adobe, this year was one where we talked a lot about success in the enterprise. So I think while we would certainly agree with that characterization of a lot of scrutiny around budgets, we've adapted really well to it, and I think enterprise was really kind of a hero segment for us in Q1. And then in Q2, we talked a lot about digital being an area of strength. That means small businesses and consumers. So I think if you look at -- what we said in the Q&A of our Q2 report, that was the first time in a while we talked a lot about some strong demand in small businesses. So I think a lot of people looked at our commentary from Q2 and saw maybe some improvements macro wise or in terms of the demand environment coming from small businesses. We're not out of the woods yet, but I think Adobe has been optimistic. We think if we do get past this high rate, high inflation environment to a declining rate environment, we think that there should be a more accommodating macro where businesses get back to growth investments, and that should be good for our markets and for our demand. So it's been exciting. Resilience for Adobe has been a good thing. And hopefully, that optimism comes to pass here not too long.

Yes, that's great. And just as a reminder to investors on the call, feel free to submit questions through Veracast or e-mail directly. But I'll start with one more on my own kind of echoing some things me and Jonathan were talking about before this call started. Outside the Olympics, the thing we were talking about was, in the past 2 years, we've seen sentiment kind of shift back and forth for investors around GenAI, but internally, I'm sure the investment strategy is much more linear. Could you just talk a little bit about maybe the biggest points of confusion or topics of conversations with investors on the topic of GenAI?

Yes. One thing internally we talk about a lot as we invest, and we've been investing for several years in these technologies, are people tend to overestimate impacts of big inflections in the near term and underestimate how transformative they'll be on the world in the long term. And when we look back at the dawn of the Internet or cloud computing or mobile, I think those things certainly hold. And so this is one where we think there's been a lot of hype and controversy around what AI means, and I think for investors, sometimes that means next quarter or next year. And when we really think about our addressable market and the opportunity, we're thinking a decade out. But we think we have a lot of conviction that this has increased our addressable market and ultimately can be a growth inflection driver for Adobe. So I think our internal commitment in terms of where we're investing and where we place these priorities has been very, very clear, but it's been interesting to see how much confusion there might be every quarter or every announcement or every time you see a new AI model do something cool. And what I would ultimately say is we believe -- you can ask, is GenAI replacement for things that humans do or does it augment and aid things that humans do. We've all played around with ChatGPT or diffusion models for long enough to know. I think by and large, these things are helping us, augmenting us, but the human is the pilot, right? That's why a lot of companies are going to use that word, a copilot. And so I think the biggest point of confusion is when people see a prompt into an AI engine produce something cool, they worry. Is the future going to just be people typing prompts until you get something cool? And our belief is these things -- these AI technologies, if done correctly, can be deeply integrated into applications people use to be super controllable. We have an idea or our client has an idea or our client's CEO has an expectation, right? And people need to be steering to make that happen. And it's too much of a wildcard to just keep typing prompts until you get something you like. Ultimately, though, with the aid of the ideation and the variance that AI can produce, it's a huge augmentation of what's possible. And I think Adobe is way in the lead in terms of thinking about this problem of not just making cool AI outputs to show it in a better -- pick the best 10 that you got, and put it up on the stage and impress people. But make these things super controllable or toolable, so you can tweak in different ways to make humans really in charge of where you're going with this stuff. And I think it's a harder technological problem to make AI engines toolable the way Adobe is really thinking about it. And over time, I think people will see how far we are in the lead in that way and what a big opportunity that unlocks for us.

That's really helpful. I'm going to mash up one of my questions with an investor question here. And thinking about the monetizable benefit, it seems like Document Cloud is one of the areas that we've highlighted as maybe an early beneficiary in the second half. Could you just talk a little bit about what you're doing in Acrobat? And then the investor question specifically was more around what the contribution has been from the generative assistants?

Yes, that's a great one. I'm glad we're hitting Acrobat early because it's Adobe's fastest-growing business and doesn't get enough attention. Think about PDFs as a file format. First of all, to frame this, PDF is one of the most ubiquitous file formats in the world. I often tell investors, if somebody sends you a document that's not a PDF, you assume they either sent you a draft or a virus. So this is the way we exchange documents. This is the way people open documents and consume them. And there's a massive productivity uplift if in the moment of document consumption, AI can just sort of say, like, hey, what do you want to know about this document. Would you like a summary of it? Like would you like to know the top 5 like takeaway points to put right into an e-mail? Adobe uses this now every time we send colleagues PDFs. We just put the AI assistant summary, right, in the e-mail. So it's a huge time saver, and it's a way that people can get much more quick insights from long documents or even collections of documents. So I think the -- it's an example of where owning the workflow, owning the interface of how people interact with a PDF means that we have that strategic high ground to put the productivity uplift right then because could you copy and paste a very long PDF and go to a different large language model and put it in and have an -- you can do these things, but I think you lose the productivity unlock when you go back and forth between tools. So big opportunity. We started actually selling GA this AI assistant in the middle of Q2. It's a $5 add-on. So it's not consumption-based. It's just a -- would you like fries with that model. When they buy Acrobat, we say do you want to check this box. Do you want to give us an extra $5 a month for all of this productivity? And a tremendous amount of people in our digital checkout flow are checking the box. To me, it's really persuasive evidence against this bear case that people aren't going to be willing to pay for generative AI. They just think of it as table stakes. And we're seeing a lot of people willing to pay for it, and that's super exciting. There's this other even bigger opportunity for us as we encourage the user base, this is over 0.5 billion monthly active users, to start playing around. We give a little bit of these conversational AI assistant interfaces away for free every month and then get them understanding the productivity unlock and making that up purchase. So that's a huge TAM. I can't break out the specific number that it contributed in Q2, but I've told investors, if it wasn't there in terms of the net new ARR, you would have missed it. It was enough of a contributor in the quarter that it certainly contributed to a nice beat against our guide, and we're really bullish on where it can go from here.

We'll do another mash up here of an investor question. So the investor question is how AI-related CapEx will impact your return on invested capital and revenues associated. And I'd like to pair that with kind of the talk about generative credit. And just the consumption model, right, would be kind of the natural CapEx offset, but we've taken a little bit of a different strategy. So if you kind of just talk through both those.

Yes. So my expectation for Adobe on CapEx is that our CapEx, which is quite low, grows slower than revenue. And that's in part because we have not been looking to build our own data centers to do all the training and inference. We're a very, very large customer or partner with the big hyperscale cloud providers through our enterprise software. I mean we have such a broad deployment on Azure and AWS and NVIDIA. And so we have very large agreements with all of those, which gives us, I think, good pricing to roll the training cost into R&D for Adobe. So when you look at our R&D, we're certainly investing hard in the people and the GPUs to train these new models. And then as we scale these features and our products and it turns into inferencing, that shows up in COGS, in our gross margin. If you look really closely, you can see those expense line items that are at our scale, I think, the movements on any line item on our income statement are pretty small. And part of the reason is that Adobe has always been really disciplined. So when there's an area where we need to spend more to fund these sorts of things, we really look hard at other parts of our business. And where can we save? Where can we renegotiate deals or do different kinds of payment structures to get better pricing? And also our digital experience business is expanding margin and expanding particularly gross margin over time to help offset some of these costs. So going back to the question. I don't think you should see a material impact over time in terms of return on invested capital. But when you look -- if you just think about our gross margin, it should be able to remain relatively flat as digital experience expands to help fund the inferencing we're going to need to spend on digital media. And on an operating margin basis, there's a lot of leverage that comes out of the increased health of the business of when people use GenAI, the top of funnel is more efficient with product-led growth. Engagement and renewal rates improve, and so we think over time, you'll see a very healthy margin, which still has room to expand.

Another question for investors. Could you talk about the competitive dynamics in the mass market area? So not high-end enterprise, thinking maybe more on the consumer side. they're specifically asking about Canva, but if I could, as always tack on another question to it, something I've been thinking about, the difference between the TAM in terms of headcount in consumer and the TAM in terms of wallet share in consumer. Just kind of that paid versus unpaid and how you think about going both of those and the competitive environment.

Yes. I'm going to start by talking about something people ask about a lot, which are all these other AI models. OpenAI has Sora, which can create videos from text or there's Midjourney, which can create images from text, things like that. But Adobe, we view -- any diffusion model where you can type in something and get content whether it's still content or moving content, we view that as a replacement for cameras. So Adobe has always had deep partnerships with camera manufacturers to make it really easy to get your content from a Canon or a Nikon or Olympus or Panavision, professional video cameras, get it into our products to do editing and complete a project. We view any AI model, including Firefly, as really replacing the need to do a physical shoot, to do a virtual shoot with these virtual cameras. And you will increasingly see Adobe partnering with other AI models to really make our products, our editing platform an ecosystem or a platform where our users have choice. We'll make Firefly models, which are commercially safe, the default, but to allow them to bring in content from other models that they can then take advantage of that controllability and editability that we have in our products. And so we really do view this as a content ecosystem that's expansive. We found, with the proliferation of this video and photo device in everyone's pocket, that more and more content was a good thing for our business. And we think that is going to happen with all of these AI models. So on the margins, Firefly for that inferencing may be competing with other models, but what we really want is people coming to our products to become content creators and learn to edit. And so you'll see, I think, even later this year, Adobe keep -- continuing in that direction of making it easy for our customers to bring different models into our ecosystem. The other place that we really compete at the lower end for the consumer, this expansive TAM, is with Adobe Express, and you mentioned the big competitor there. I think historically, Adobe only really played in the high end, and there were lots of democratized low-end free tools, Microsoft Paint, clip art. If we all think about our own professional lives, most of us, whenever we needed a piece of content, didn't create one or hire a creative professional. We just kind of went online and found a picture and copied and pasted it. So the proliferation of easier-to-use tools like Express and AI, I think, takes all of those people, think of all the knowledge workers of the world or all the Internet users of the world, that's 5 billion people, and potentially activate this creative intent, where instead of going online to search for an image, they can just describe the image they're looking for and AI makes it. And then we can bring them into Express, which is now fully replatformed as an AI-first application, to sort of let them like just think about the basic starting blocks of becoming a content creator. Add your own fun. Add your company's logo. Do simple things and share it on social. So I think it's a new TAM for Adobe. It's one where we weren't -- we weren't first to the space. In terms of getting templates online -- people starting with templates, that's something Canva did really well. And for the last decade, starting with the template and editing, it was kind of the way to go. We think the next decade won't be as much about templates. Templates are kind of slow. A human creator has to create each one. We think AI first is the way that market is going to go. Whoever has the best AI tools to make people be able to take a picture of their cat and put it in outer space and move things around and publish it out to their Facebook page, that's going to be kind of the winning equation. And we think we have the best AI-first application out there, so now we're really going to be aggressive about starting to sell that both in consumers and small businesses and enterprises. So we're excited about having these big new markets that we think are validated enough that we're convinced they can drive a lot of growth. And we're starting to see -- particularly in emerging markets where we never quite have cracked the nut on monetization, we're starting to see a lot of new customers coming to Adobe. So we're excited about where we sit.

And I think you answered this question but just something that comes up sometimes. Have you guys given thought from a TAM perspective of generative images and generative content versus generative editing and kind of where the long-term value line is between those?

Yes. I think the content is mostly in the consumer space, where people are -- the same people that were just scrolling through Google image search, looking for an image to use will now be just typing some prompts to get something good enough to -- for whatever their school project is or their work presentation. I think that's a big new TAM, and that's the 1 where we talk about 5 billion global Internet users are what we're going after. And the place where Adobe historically has competed, which is creative professionals, businesses, enterprise workflows, those folks aren't looking to replace anything they do with AI-generated outputs. Shantanu will often say to investors, "I don't think Steven Spielberg is planning his next film with a bunch of prompt engineers. He's looking for known talented actors, producers, directors and shooting something to match his vision." AI can be a tool in that, where if there's -- I mean, I remember watching all the Lord of the Rings extended features and learning about when they -- when there was a continuity error and they had to -- somebody flipped around a badge on the cloak. They had to go reshoot the whole scene. Well, we can fix that. AI can flip around the badge. And so AI can do a lot -- I mean there was a Game of Thrones, where there was a Starbucks coffee cup on the table. AI can jump in and help out, right? So I think in the world of enterprises and professional content creators, it's very much an editing tool. It's an ideation tool on the front end, but more importantly, it's a tool to drive variance, localization and personalization. So for example, Adobe in our own like fall campaign, I think all of our agencies and creative professionals created around 100, what we call, hero assets. But we -- to do all of our localization and personalization to -- we're not targeting 10 segments. We're targeting 10,000. We turned those 100 hero assets in over 100,000 variants. Firefly services and AI did that. There's not enough creative professionals in the world to do that for every business. And so it enables things that weren't possible before, and then end users can get content that's much better tailored for them. So I think the TAM and the opportunity is totally about the editability, the controllability, the variance. Those are the hero cases right now. And those are the things that businesses are actually asking us for. They're not saying like, "Hey, we don't want to go shoot B-roll of a bird on a tree. We'd rather type prompts until we get a nice one." B-roll is easy, right? That's not -- nobody is trying to solve that. They're trying to solve much more challenging problems with scale and speed.

And when we talk about kind of the strength of your editing suite, which has been a long-standing part of the Creative Cloud, one of the questions we got from investors were basically talking about that embedded knowledge base you have within the industry. But as AI makes these tools easier to use, does that enhance or hurt your competitive moat? Basically, does the 20 years someone spent on Adobe make it more entrenched? Or does it make it easier for them to switch to someone else that the onramps are easier?

We found as we make -- so this isn't theoretical. We've done it in Photoshop. Photoshop is a complex product. It's one I've been using since I was 17, and I still learn more about it every day. But when we made Photoshop much easier to use with this AI-assisted onramp, we sold more Photoshop, just period. More people stayed and went from trial to buy. Retention went up. So we know -- it was -- I described what we saw in Photoshop as thesis confirming data that show that if you make these things easier to use, editing easier, people will do more of it, not less. So that democratization is a great thing for us, and we certainly want to bring more and more of that. The other thing I would say -- I mean, there's a lot of ways in which the network of professional creators, the deep knowledge around the file types they use, collaboration between artists of different media types, there's all kinds of ways that our ecosystem really advantages Adobe in this way. So -- but I want to come back to something I said earlier. Making an AI model that's capable of producing some cool things that you can put up on a YouTube video and everyone says, "Wow, look how neat that is," that's one thing. But making those cool outputs super controllable, so people can actually do what they want with it is technologically, it's a different [ product ]. It's not just sort of like I'll spend some administrative time wrapping a tooling interface on top of it. It's hard to get these models to be -- to hone them and control them to do exactly what you want in a tooling ecosystem. And nobody else, to my knowledge, has -- is really trying to solve that other than Adobe. And so I think that's a really important differentiator between us. And I think as this plays out, investors will start to see how important that is to the big spenders, the agencies, the studios, the professionals of the world.

Going to another investor question, and this is maybe a little bit more broad than some of the products we've been looking at. But can you help give us a sense of what's driving confidence and visibility to the Creative Cloud growth in the second half? And what would it take for us to get comfortable with a $2.5 billion net new ARR next year? Is this the right way to think about the Creative Cloud business sustaining in the [ LDD ] zone?

Sure. So let me first just remind folks if they might not be aware. We said in our Q2 call that the Creative component, we have digital media net new ARR, but the Creative component of that will grow based on our expectations in Q3 year-over-year and in Q4 year-over-year. And I think investors have been really excited to see that happen. There's been some concerns. Will it happen? And based on all of these new products and underlying trends that we have that are ramping in the back half of the year and simple dynamics around the timing of some pricing changes in the portfolio. Effectively, we now have exited a tough comp on pricing in the Creative line item from the year ago quarter and have some nice pricing tailwinds with what we've been doing with the AI rollout. So based on the way we've modeled the business, investors are going to see that back to growth year-over-year in Q3 and Q4. I'm not going to get into any kind of foreshadowing on 2025, but I really like the way Adobe's set up. With -- software companies always say we love having new things to sell. The enterprises are excited about, consumers are excited about emerging geography signals are really positive and even some pricing tailwinds. So I think it's a great setup for the back half of the year. And the reason we're investing in all of these things is certainly because we think there's a big opportunity to drive an inflection on growth. And we said we're really focused on -- this is the first year we'll cross $20 billion in revenue. We're really focused on that journey from $20 billion to $30 billion, which we've been talking about for a few years now. So I think there's a -- it's a good spot to be in.

Yes. And you mentioned the pricing point. And to another investor question, obviously, we've kind of talked about Q1 headwind, Q2 tailwind. Q3, I think, was supposed to be even; Q4, to a tailwind. Fast forwarding kind of into FY '25. Can you just give us some color on how we're going to lap prices next year just so we're clear on when the tailwind subsides?

Yes. So Adobe has been pretty thoughtful when we raised prices, first of all, to make sure that it's really driven by customer value. And we always like the benefits of our products to accrue disproportionately to our customers. So I think the price increases have been well received because there's so much value. And then we're also thoughtful about kind of feathering in any pricing-related growth, which we've had pricing-related growth in our ARR since 2018 after we got through the original transition to subscription. We do it in a way that I think avoids any big lumpiness or -- and so for the AI price increases, they started in November 2023. It wasn't a full geographic rollout, and then we sort of extended some of the geographies as we've gone through this year. So the prices that we've announced already will drive net new ARR in 2024 and contribute positively to net new ARR in 2025, so it's -- and then we've said those price increases related to roughly half of the Creative book of business. And so as we operate the business and we're looking at data between new customers coming in and upsell flows, there are other pricing-related changes that we can make as we operate the business or not. So I think we have -- it's something that we've shown as a core competency of being able to use pricing to help be that third leg of the stool to our growth algorithm. But certainly, we're more focused on all the new products and new customers, which are the primary drivers of growth.

Maybe staying on pricing but a little different aspect. A question from an investor was just when do you expect to start seeing some monetization from generative credits. And I know when you originally launched the tiers, the idea was that most people, based on usage trends, wouldn't be going over. And maybe it makes more sense for upsell, right, to the next platform instead of purchasing more. So is this something that more relates to the more compute-intensive models like video 3D vectors that are coming or kind of just what you're seeing right now?

Yes. When we introduced the concept of generative credits, it's kind of we invented a new economy where we control all the levers. We do enforce the free-to-paid generative credit paywall. So if you're going on and playing around Firefly online or in Express, once you use your free allocation of generative credits, you had a paywall. So in that extent, it already is something we enforce that's driving monetization. But within our paid plans right now, we're letting people play as much as they want. And I liken it to a video game arcade 30 years ago, where it was $0.25 for every game, whereas now if you go into an arcade, this one's 8 credits. This one's 6 credits. There's -- they're more complex experiences. And so that inflation is something we'll drive over time. And there's some more complicated models that are more compute intensive, like video [ come out ]. That will also drive inflation. So it's something that we'll, at the right time, decide how we evolve credit allocations and credit enforcement. And also, I imagine, as some of these new models come out, there will be new ways we monetize it. And going back to Acrobat, rather than make it based on credits, we decided to initially make it a very simple $5 a month add-on. I suspect when our video models come out, we'll do some exploration and also have some flat fee SKUs that enable people who want to go really deep in video workflows to do all that complex stuff without thinking about credits per se. So I think there's a lot of different places we can go. But initially, we knew we were planning for proliferation. The analogy I make is we're giving away a lot of free coffee in our apps and getting everybody a little bit hooked. And then eventually, we're going to bring out the higher-priced lattes and espressos and cappuccinos and drive a lot of upsell behavior. And the great thing about coffee is I never begrudge who I'm buying it from because I'm the one who wants it. So when you -- when people are hooked on something they really enjoy, they're in control of when they want to buy more and use more. And I think that certainly applies to what we're seeing in GenAI.

Maybe a natural caveat off of that is without like the direct revenue contribution from credit, what do you think are the right line items for investors to be focused on when they're thinking about GenAI adoption? Is it digital media net new ARR as people are upselling seats and moving up? Or is it RPO for the Experience Cloud, seeing people create larger and more standardized buying patterns?