Okay, thank you, everyone for turning up this morning. Welcome to the UBS Technology Conference. I'm Michael Briest, Head of our European Software Effort, and delighted to have with me Dominik Asam, who is CFO of SAP, as I'm sure you know. So, we've got about half an hour. If you have any questions, there's a QR code on the desk, and you can submit some questions to me, which I can put to Dominik. But obviously, I've got some to start with, and I'm sure you're getting tired of this: every conference starts with a macro question. But you're doing a fantastic job in moving your customers to upgrade ERP. A lot of the industries you serve are financially challenged. There's political uncertainty in much of Europe, including in your domestic market.
Why do you think customers are willing to sign on the dotted line for those sort of multi-year commitments of migrations? What is the value they see in, and is there any more hesitancy today than there was, say, three or six months ago?
So I'd say the major benefit we have is right now just the upgrade cycle that a lot of customers need to move to the next generation of SAP systems. And I think the evangelization on the benefits of the cloud is making good progress. They also see that with AI, being in the cloud has certain advantages they can leverage. Yes, of course, there's macroeconomic pressure, and we do see that also in some segments. Think about Fieldglass, temporary workers, Concur, travel expenses, and then also on the Ariba Network, on top of the macro, which is dampening a little bit the procurement volumes. There's also the kind of transition in our pricing model on that. So that has been a little bit of a headwind on the transactional businesses, but on the subscription businesses, we are really very solid.
And especially on these migrations of the installed base, we're doing extremely well. Let's not forget that more than half of the maintenance base we have still in on-premise on ERP is from prior versions of SAP, so ECC and older, where we have out-of-maintenance states, mostly starting in 2027 and then extended maintenance 2030. So given the complexity of these products, they need to start sooner rather than later, to make sure they have a product that works. So it becomes more and more also a business continuity issue on that front. The other thing we think is currently helping is that we are basically delivering tools for knowledge workers, and personal expense is actually seeing some upward pressure with inflation and with some of the policies that Trump might enact that inflation could also increase.
And if the macro is more difficult and the kind of top-line opportunities are harder to come by, the most natural thing you do is you figure out how can I really push the bottom line and drive productivity. And that is then really also the second to AI. All the management boards are currently challenged by the supervisory boards, the board of directors. What are you doing on AI? And we tend to run quite generic processes like record to report, procure to pay, where there's not so many companies who really wanna invent, reinvent the wheel for themselves because they know that we have tremendous economies of scale of doing that for them. And I'd say we have also a good traction in terms of showcasing opportunities.
So I'd say there is a little bit of a move back on that front to, yeah, ERP systems, which potentially during the heydays of the COVID crisis were a little bit under-invested. And then last but not least, I'd say there is a little bit of a what I would call the return of the suite. If you look back into our previous, kind of on-prem model, we were always burdened by these high degrees of customizations, which triggered a lot of upgrade costs. Customers procrastinated upgrades to not invest too much money because their budget might be constrained, but now that, of course, means that they cannot consume some of the more productivity-enhancing innovation in the near term. And they really are now buying into that move to the Clean Core and the Integration Suite and the BTP.
That kind of puts them in a position to also avoid investing in system integration to stitch all these applications together for whatever they need to do there. So these are the most important factors I wanna call out here. This is why you have seen also in the last quarter that we have an unabated momentum in what we call Cloud ERP Suite, which is basically that transactional engine that you can deploy in a broad base in these customers.
I mean, you've mentioned the transactional business and the headwinds there. Is there anything you're looking for which might be a lead indicator of more positive times ahead? I mean, what would you look to in the business besides maybe transactions that, you know, indication we're getting through the worst of this?
I mean, honestly, we are cautious on speculating on big snap-backs on that front, because we might be proven wrong by macro. And this is why it's not really moving the needle so much more. It's because it's diluted down over time. It has been actually declining in the last quarter versus a Cloud ERP Suite growing. And Cloud ERP Suite is actually including the transactional business. So you can imagine how fast Cloud ERP Suite excluding the transactional business has been growing.
And the good news is that on a year-on-year comparison between what we've seen in 2024, which will be probably a kind of, yeah, 1.5-2 percentage points, dilution of the CCB into revenues because of transactional business, that will be cut in half next year just by virtue of, we believe a normalization on the Ariba business network, which is going through that pricing change. And that should kind of find its spot in this year. And then we don't need any miracles on the macro front, really. So I don't spend like day and night figuring out is the transactional business snapping back. I'd rather kind of take a prudent assumption of what it will do and then do the heavy lifting on the growth with Cloud ERP Suite and the subscription business there, frankly, on cloud.
You've said before that the RISE contracts have a natural ramp in them as a customer is committing to a migration over many years, potentially. Can you give any sense of the visibility you have of, obviously, next year's revenues are coming from this year's CCB, but how those build up over time and maybe 2026? What revenues already sort of on-boarded.
Yeah.
From these migrations, which is,
Yeah.
Have started already?
I mean, you know, we disclose what we call the TCB once per year. And I think last September 31st last year was 39% growth, constant currency growth, and to reach EUR 44 billion. For now, the coming year 2025, the kind of booking rate is very much determined by what we call the CCB. It's the 12 months ahead. We have been quite, quite kind of explicit on what is the existing CCB we anticipate at the beginning of the year. We said that we have entered the year 2024 at 27%. It will not be so dissimilar for the end of the year. Now you've seen that we printed 29% in Q3, but there was also the inclusion of WalkMe in there. It was a good percentage point of accretion, so to speak, on growth.
So actually de-accelerated, so to speak, for that, we had now three quarters in a row at 28%. That is giving us a good base for next year. We have to say Q4 has a little bit more difficult comps because we had a super strong acceleration last year. But if you put it all together and assume that kind of reasonable assumption on CCB growth, you see that actually, on cloud, given that we said we wanna have EUR 21-21.5 billion of cloud revenues, it's constant. No, that was actually 11%.
Yeah.
which is about 21.7, I think, on the rate we had entering the year because we had 109, I think, for the beginning of the year. There's more than 80%, if we hit that type of CCB number end of the year, there's 80%, in that CCB that's already covering. So then the rest has to come from really turns, business, so to speak, in the year from the transactional business, from some other more consumption-based models. And we feel that's kind of giving us the right mix. So it all depends now also on how we close Q4. It's always very back-end loaded. I'd say the pipeline looks reasonable, and we work with that pipeline, but we'll only know for sure post Q4.
But no change to what we said before that we are really tracking just online for that kind of Ambition 2025.
Okay. And a question I sometimes get is the sort of real world, if you like, of these migrations. Some of your larger customers will have tens, you know, many tens of ERP systems. What happens if they miss a go-live that they've committed to either for commercial reasons because they've got cost pressures or practical reasons that the project's gone a little bit more slowly? Would that still, that date still trigger a cloud revenue conversion for you, or do you sort of maybe give them leniency around that? And how good are the customers at managing the sort of non-duplication of revenues for you so that they're not paying you both for maintenance?
Yeah.
Cloud at the same time?
I mean, you have to look at it really kind of installed base, company code by company code. So when they go to the cloud on that transition, they would, over the course of that year, then reduce the maintenance, or eliminate the maintenance and then pay subscription. There is a little bit of an overlap within the year. They changed.
Yeah.
But that's minimal. But in general terms, you can say it's a very gradual journey. And normally there is a plan, a ramp plan agreed upon. So many customers, they just execute on that ramp plan. There are some clauses sometimes that allow the customer to kind of defer a certain portion of the consumption into the following years. And that, of course, would not be considered in our CCB, by the way.
Right.
So as soon as there is some deferral rights of the customer, it's not in CCB. We push it out, and yeah, this is also why we incentivize our sales force much more strongly on adoption to really make sure that this machine is spinning well and ramping. It's actually a big portion of the growth that is triggered by the ramp of what we have signed already in terms of RISE deals, and then off top, on top, you have the net new names, you have the bookings, the conversions, and so forth. This is giving us that very, very high 30% plus growth on the Cloud ERP Suite.
Okay. And, a lot of investors are now looking beyond 2025. I think at Sapphire, you made a comment that even if the rate of decline in the on-premise and services portfolio doubled from about 4% to 8% per annum, that you would continue to grow and accelerate. Is that an actual indication of how you expect the non-cloud portfolio to develop over the next few years?
I just, I mean, we have always said that there will be a slight acceleration of maintenance because as we hit 2027, 2028, when the auto maintenance comes, more people will try to get out of there because it's becoming more expensive with, and then 2030, they have to be moved. So from that perspective, that decline in maintenance that is very gradual, by the way, we always said it will never be like a step function. It will be a kind of, and I mean, yeah, if you double that, that's not a completely stupid assumption in terms of how this non-cloud revenue base will evolve. And the point I made at Sapphire is actually. I was actually saying more than that.
I said that even on the kind of Cloud ERP Suite, we would see, we would be able to absorb some decline there to lower levels and still grow. This is just a mixed effect. I really encourage everyone interested in that story to do a mini spreadsheet and then do what you said, have a reasonable assumption about how the maintenance will evolve and the software will evolve. And then, also look at a kind of typical low growth of services that gives you certain deceleration of the non-cloud revenues. And then, you know, in cloud, there's also the infrastructure service business, which is declining. And as it's becoming smaller and smaller, that headwind will ease too.
And you will actually figure out that in order to keep revenues flat from 2025 to 2026, 2027, which I sense is what sell side at least has, has now in the models, that you can actually afford a pretty massive deceleration on Cloud ERP Suite. And of course, that's not our ambition. And this is why we say, as long as we're not kind of almost imploding on growth on Cloud ERP Suite, we have a pretty good likelihood and statistical likelihood that we can accelerate revenue growth in 2026 and 2027. Those of you who know me, I'm always very spooked about not giving long-term guidance. Normally, I don't do that. But because of that super strong mix effect, we have now 83% of cloud revenues of SAP running at 30% plus for 11 quarters in a row.
And having the type of materials, so to speak, for the next years in terms of what we can still convert, I'd say I have no reason to believe that Cloud ERP Suite will massively decelerate. And given that's the case, I feel confident despite all the imponderabilities we always have as we move out longer term, to say, I mean, it would need to be pretty damn bad if you would not accelerate group revenue growth. So that's the logic. And you can, I think it's really helpful to play with these models and say, okay, how much can we decelerate and what does it mean for the group revenues?
At least through 2027, because we have that ECC maintenance date looming, there's a very good underpinning of growth from the conversion that gives me the faith that Cloud ERP Suite will not kind of fall over a cliff.
Very clear.
Yeah.
I think in the past you've talked about a sort of two to three X multiplier of the maintenance revenues when a customer migrates. I was at a presentation of one of your SI partners recently. Maybe it's macro, maybe it's to do with the urgency of this 2027 deadline. They observed that a lot more customers were doing a sort of quick and dirty sort of lift and shift of ECC to S/4 into the cloud, not doing a full transformation, not going through the cleansing of the custom code. To my mind, maybe that makes them less approachable with the cross-sell and upsell. What are you observing about how customers are migrating? Are they following best practice and giving them?
Maybe, maybe the comment is also a little bit self-serving. It's clear that in some way, our cloud strategy to drive to Clean Core is trying to reduce waste in terms of customizations people do in their existing ERP estates. And it is actually a pretty big evangelization effort to make sure that our system integrators really understand that this is the way to go. And the good news is we are pretty big if you quiz them as to who is their biggest kind of partner implementing stuff on ERP. Probably most of them will say, well, SAP is pretty damn important for them. And we clearly tell them, look, this is not the world we want anymore. Of course, if you have a lot of pressure, for instance, time pressure, then you might just do a lift and shift.
But if you do that lift and shift, normally it doesn't end there. Normally then people start to depollute the core, make sure that they're upgrade ready because they do understand that if they are not able to upgrade fast, they cannot consume all the innovation we bring to bear in particular scenarios we deliver now on AI. So I'd say I'm not surprised that system integrators would say, look, people don't wanna go to standard because honestly, it's bad for their business and why, why should they say anything else? But the truth is that I've not seen any CIO recently I've talked to who said, I wanna just do a lift and shift and I'm happy with what we have.
Everybody is, how should I say, burdened by the curse of a kind of highly, highly customized system and tries to look at ways to become more agile, to become faster with innovation and not have these extremely expensive upgrade cycles, which can be actually as expensive as a new implementation. So in some way, we're of course attacking that profit pool of kind of, how should I say, billing hours, to make that a more scalable product-like approach and then share the benefit of that between our customers and us SAP.
I wanna come on to AI in a moment, but just one question on BTP. To keep out of RISE, it was an important revenue metric you gave up until last year. Can you give any sense of adoption there as, you know, in terms of a platform use case as opposed to an integration? Anything on the number of customers, the growth rates, the number of industry add-ons, something to demonstrate that traction?
BTP is evolving very nicely. I mean, the reason why we stopped reporting it on a very granular fashion is that, of course, just statistically, the quarterly growth numbers are more volatile, for any individual sub-segment of our business. And then I don't wanna have a discussion about why is that 3 percentage points up, the other is 3 percentage point down. And of course, the investors then always pick, pick, pick the ones which is down and quiz me is anything now, is there a catastrophe looming on that front?
Also we have developing business models where the customer has the right to swap between consumption on BTP and other lines of business and the ERP system, which is a big competitive advantage if you think about it because the problem the customer has today is they're really scared when they go on the cloud journey that they're kind of captured in the cloud, that they're locked in. This is also why you see the TCB growth being higher than the CCB growth because they're trying to stretch out the duration of the contracts. It's actually not what we want. It's actually the customers who push for that to have visibility. Based on that kind of, but the counter argument to that is, oh, I don't know how much of what I really consume when.
And we have a unique opportunity by virtue of the breadth of our portfolio to say, don't worry, while with any niche customers, you're locked in that you need to consume over such a long period of time, that with us, your demand might change over time and you can actually reallocate. This is what we call commit to consume models. And in these commit to consume models, it's really, really difficult then, to isolate the numbers. And it's also getting more difficult from an auditing point of view. And this is why we said all of that is Cloud ERP Suite. And as long as that metric is roaring and up and running, people shouldn't be concerned. And we don't wanna, frankly, waste time on saying why is the percentage point less on that and percentage point more on that because people can shift.
It's a competitive advantage we have because we have a very broad portfolio. This is also a reason why people say, okay, if I'm underconsuming on something, you mentioned, isn't there risk that some kind of cloud solutions will be consumed more slowly? Yes. But then, well, we have an opportunity to say, yeah, why don't you kind of throw out this competitor and consume more of this other thing and then you can absorb your credits? So this is the background of that. I can kind of affirm that the ecosystem is coming along nicely. We have really thousands of partners now coding on BTP. It's an extremely high attach rate, especially on the more complex enterprises. So most of the journeys, I mean, almost all the journeys have actually a BTP component to it.
I mean, RISE without BTP is a little bit sad, yeah. So, it doesn't make so much sense. So it's really a very high attach rate.
Commit to consume, is that a big part of the business now or the new business you're writing?
No. It's significant, it's becoming bigger as we speak, yeah. But it's also more flexible, it's also harder, frankly, from an accounting point of view to keep it separate what's going where. And I do believe that it's a trend that people wanna have both predictability of their budgets and the flexibility because they don't wanna be boxed in in niche spend, which they can't get rid of. So I would say yes, it's a product that in the portfolio will grow.
Okay. I don't know, I think on the last earnings call, Christian said that 30%, of the orders include an AI element of some form. What are the capabilities you offer today and plan to release, say, by the end of next year? And is there any profile to the customer that makes them stand out as why they're using AI or early adopters?
I would say to start with the second part of the question, the biggest question is how much personal expense does the customer have in the knowledge worker space? And the bigger that number, the better for us because then with inflationary pressure every year, they need to find some mitigants against that. And there are not too many mitigants except for AI these days. So how do you make these people more productive on these processes that are typically run by SAP systems? How do we avoid waste of time? And this is why we set out to say we wanna basically bring the 80 most frequently used analysis and transactional cases to cloud, to Joule, by the end of this year. There's a good trajectory. There's 100 use cases defined.
We have actually reached that earlier than the end of the year, which was the ambition, and I'd say the big pushes we have are of the following nature. The first one is really Joule in itself. For me, it's just the next step. I mean, everything we do in Cloud ERP Suite is transactional in nature. It's not about, it's not about some funky thing that's nice. I mean, I've seen from other AI providers some demos of what they feature in terms of great stuff we can do with AI, and one was I can simulate a mechanics book with a basketball kind of analogy, and the other one is, how I'm preparing my trip, to go, in a tent somewhere in, in the winter and what do I need to take along? I mean, who, who on earth is interested in that in enterprise applications? Nobody.
People are interested in getting shit done. And so what we are doing is we're really tailoring our stuff around exactly these tasks. And with Joule, you just accelerate the transaction. The same way with e-commerce, you didn't have to park your car, go to the store, wait in line for being served, and then figuring out that the product is not there and going to the next store, you could go on an internet platform. Today, we still have that inefficiency in the system, which we can kill with Joule, which is there's a guy who transacts. That guy wants to have some data before he decides to transact something, and he wants to transact quickly. And that's not actually happening on the internet, on the typical layers you have in terms of these clicks you have to do.
First, you go on a search, you click your way through somewhere, and then you have to click your way to execute something. And Joule should actually kind of shortcut that whole thing in an order, do this, and this will give you the data and then execute it for you. So this is not even full automation here. This is just making sure that a good copilot is really doing the job of a copilot. And then, of course, there's the full automation where we use large process models to really figure out with these foundation models, well, if somebody's sitting in a shared service center, this is what they do today. This is what it's about to do now. What are the next steps in that to really automate more? And then there is the enablement of our entire partner ecosystem.
When I did the transition to S/4, I had about 60 people from a system integrator, and we had to fire half of them and replace them because they didn't do their job. Now, with Joule for Consultant, I can make these guys black belts, each of them, because they have at their fingertips all the collective knowledge of the best system integrators on the planet. Think about the huge ABAP code. We talked about going to Clean Core. What you need to do very, very tangibly, if you wanna go to Clean Core, is you have to look at these tons of customized ABAP code in your on-prem stack and figure out how can I move that to ABAP Cloud in a Clean Core approach. The same way there's a GitHub Copilot for other languages, we do that now for Joule.
So these are just examples of how we wanna really bring tangible use cases to the market. I'd say there should be a lot of vibes, and I can only encourage everyone to plan to come to Sapphire. We will then show what we are doing because ABAP Copilot is coming, for instance, Q1. Then there is also the knowledge graph. One thing you have to understand is that our transactions are in the HANA database, basically represented in lines of a massive database, with a lot of attributes. And what we do with knowledge graph is really figuring out how all these things in real processes in companies are linked together and how I can actually expedite AI to deliver to the foundation model, the context, the process context of all this data, which is super valuable. So a lot of good stuff happening.
Yes, you can always say it takes a little bit longer until you really see the big revenues from that. Honestly, for 2025, to hit the ambition 2025, it's not kind of based on fantastic AI growth. But I'd say, from the assets we have in this area, we are really privileged that we can combine the data with the context of the process, the semantics around it, that we have a chance to make it relevant. I mean, I mentioned what does it help you to know how to what you'd bring along on a hiking tour in an enterprise; it doesn't matter. I don't need a basketball analogy for mechanics book, frankly. It has to be reliable, and hallucinations cannot be afforded. Then there's the responsibility. Think about the stringent contextual access rights in an SAP system.
How should somebody just egressing data left, right, and center preserve that to run AI? It's not possible. So these are the unfair advantages we currently enjoy. And customers understand that. And this is why also customers converting quickly to the cloud versions of SAP. And this is what was driving that kind of Cloud ERP Suite growth we are showing every quarter.
Okay. And a couple of questions on margins, just in terms of the go-to-market function, obviously the change of leadership at this middle of the year. What are the, say, the two most important things that are gonna be changing or two or three, whatever, over the next 12, 18 months to get you in the right place?
I mean, one thing is really to leverage the ecosystem better, for pushing the standardized, multi-tenant, what we call grow public cloud product. We are finally in a position to really tackle smaller customers with that because you can really procure that at a much lower cost, and we see, in particular, in greenfield deployments where customers are basically starting from scratch that it's getting adopted, but frankly, we cannot sustain the high acquisition cost of a large enterprise go-to-market approach, white glove, hand-holding team of reps running around, which have been, frankly, created, also in the context of the acquisitions we've done. So we've put a lot of kind of specialized sellers next to each other, and we need a more streamlined, partner-enabled, model, so we will also design sales territories in a much more stringent way, give incentives, give the marketing material.
And this is something that's currently really happening to also turbocharge the growth and GROW. On the large enterprises, the conversion of installed base, we actually also need to come to a much more data architecture-driven approach. So it's not about selling one single line of business anymore, like the solutions we have acquired by all these acquisitions. It's about being the trusted advisor of the CIO, being able to understand the enterprise architecture. What's the role of SAP? How do our data estates, how do they federate these estates with other data estates? So it's a much more sophisticated tech job, actually, as opposed to it's a gardening tech job as opposed to pushing stuff down the throat of the customer.
And this is also a different approach where we actually need less people, but probably a little bit more technical people who have more the view of the full suite as opposed to one line of business. And this is the beauty now. I mean, I always say jokingly to Christian Klein, now we have a dictatorship. You are CEO and sales at one. You can do what you want. And we are really now implementing what actually is only common sense and what some investors have already complained about for years, what we're doing differently and probably not as well as others. And we have an extremely high degree of confidence that there's no resistance on that and that will be implemented. And honestly, if you look at our selling expenses for new acquisitions, they are so high, there's not much to lose.
So the risk-return profile of that activity is very attractive because if we can get reasonably close to what we are gonna achieve, we can eradicate so much waste that the risks in kind of execution are quite moderate. And also with the lion's share of the kind of growth now in the box with the conversions, these are big journeys where we have a lot of people on the customer side, on the SAP side. It's not dependent on one single guy not being happy about certain changes, but it's really an almost institutional journey that two companies have embarked on. So I think it's right, the right point in time to do that now.
We are very, very optimistic that, with these two main levers and there's other smaller ones we can discuss, we will make good progress in bringing the selling expenses as % of revenues down while hopefully on top, accelerating the effectiveness of what we do here.
We're almost up on time, but just you, you brought up GROW and public cloud or multi-tenant cloud there. How important is the adoption of multi-tenant cloud by your, you know, ERP S/4 customers in the years after 2025 to the margin vision you have?
Yeah.
Maybe you can say a bit about how much of the S/4 Cloud today is actually public cloud as opposed to private?
Yeah. I mean, it don't give data on it, but what I generically see, it's kind of the private cloud is one order of magnitude bigger than the public cloud. So if you know one order of magnitude in mathematical terms means 10X, yeah? So that's the kind of very rough, don't write any report. It is 10X. It's just one order of magnitude more, and what kind of ballpark that is, you can figure out from that kind of logic. It's almost like an engineering logic, and that means that it's not so important in the near term, but it will be very, very important in the outer years of the planning horizon.
And this is why these changes I mentioned that you cannot kind of sell GROW for, I don't know, 100, 100K ACV with a white glove, seven people strong handball team. So you need a more dedicated, more automated go-to-market. It is not the heavy lifting on the margin. I mean, you're seeing currently the growth is still predominantly driven by the private cloud on S/4. And still we are seeing the margin expand. So the biggest lever for gross profit expansion to the levels we, for instance, guided for 2025 is the margin expansion in private cloud itself, where we have plenty of opportunities. Then, of course, there's the mix effect because you know that public cloud has a much higher margin than private cloud because you have a much more efficient model. It's more attractive for the customer, and it's more attractive for us.
So it will become important in the years. I'd say in the outer years of a planning horizon. And this is why we are really now, accelerating and put all the ducks in a row to make sure that this next leg up is really well sustained.
Okay. We're up on time. We're gonna have a very quick one. Obviously, head of sales or customer success and chief technology officer roles are currently sitting with Christian Klein. Is there a timeline for replacements, and is there a view on internal versus external?
Yeah. Actually, this is in our corporate governance setup, is the role of the supervisory board to decide that. But of course, the company's working feverishly, and these, the people in these, supervisory board committees in charge are working feverishly. But I would still say it's more important to have the best guy than, so we are focusing on that. And I'm really convinced that where we stand today with what we have to offer, anyone who wants to play big in ERP and, who looks for the leading company in ERP, probably has a desire to join. So process is running, but I can't now give you an update concretely on when exactly this will be terminated because negotiation where it takes two to dance and so forth.
Understood.
I'm optimistic that this will be solved. For the time being, we are not suffering from it because Christian is working like hell, himself, which also many customers appreciate to be kind of personally served by the CEO for a certain period of time. On the transformation, it makes things so easier if you don't have to discuss between sales and CEO what's happening, but it's one person.
Yeah.
This kind of positive dictatorship is really helping.
Good to hear that. Thank you very much, Dominik.
Yeah.
Thank you, Ambassador.