SAP SE (SAP) Earnings Call Transcript & Summary

Okay. Perfect. Let me start this next session off. I'm Adam Wood. I look after European tech research at Morgan Stanley. It's a great pleasure to have Dominik Asam, CFO of SAP. Dominik, thank you very much for joining us.

Thanks for having me, Adam.

It's a great pleasure. So look, let me start things off. I think the last -- actually, probably a couple of years, we have seen a relatively tough spending environment for software companies, and yet, in your business, we've seen the cloud ERP suites even accelerate. I think investors are well aware of the 2027 and 2030 deadlines, as I would call them as light sticks to get people to move. You might disagree. But could we talk a little bit about what are the carats? What are customers getting excited about and seeing as the positive opportunities from moving to S/4?

Yes. Firstly, I want to say that it's not that we go completely unscathed with the macro because you've seen in our transaction businesses that we did see a dampening effect of the macro on businesses like Fieldglass, which are temporary workers, Concur Travel Expenses or Ariba, where it's about procurement volumes. We do actually indeed enjoy that secular growth boost from converting our huge install base very gradually into cloud customers. And whenever we do that, we have an opportunity to expand the revenue base quite massively. The factor we always mentioned in the context is somewhere between 2 and 3, to give you a ballpark number, and that is really firing on all cylinders because, as you mentioned, some legacy systems like ECCs have an out of maintenance deadline, mainstream maintenance end of '27, extended maintenance 2030. But that's not all of it. There's also that kind of scrambling for productivity. People are burdened by increasing personnel expenses with what we see now happening in the U.S. Maybe inflation will be there for a little longer, and that has an effect on productivity of the employees in terms of cost per job done. So everything that's more related to make companies more competitive, productive is right now in high demand. And maybe the focus is gravitating a little bit away from other areas. Also we have a very strong upsell and cross-sell opportunity, which we can see in our renewal rates -- in the net renewal rates, which are really supporting us. And the integration we have invested in very heavily over the last 2, 3 years is now starting to pay off, that customers realize the synergies if they can put stuff on one platform and they don't have to waste a lot of money stitching them together with expensive resources from system integrators. So all these things come basically together, plus, of course, us entering something what used to be kind of a new market for us, which is the mid-market with the multi-tenant public cloud grow motion, which is, of course, the ultimate destination for many customers. But for the smaller customers, it's easier to go there right away in one step. And all these things are coming together and give us a certain robustness, resilience in cloud ERP suite.

And maybe just on that theme of visibility, could you talk a little bit about how companies contract with you when they sign these S/4 contracts and what visibility that gives you into '25, '26, '27?

Right. Basically, when we sign the RISE contract, which, by the way, might not be purely RISE, there's also some parts of the business that might go GROW. Some examples we can mention here where that's the case. When we do that, it's for the large enterprises, which are the traditional incumbent customer base, a multiyear journey. So we talk about maybe 5-year journeys. You can also see it a little bit in the backlog we have. If you look at CCB versus TCV, you see kind of ratio is that order of magnitude. On the lines of businesses outside that as for journey, it's shorter, tends to be shorter because customers have no choice. One reason why people want to have a long contract duration is also that not only does it take them time to convert one instance after the other to a cloud environment, but they also want to have predictability of cost. And they think if they have a long contract, the kind of pricing is locked in for a long time, depending on what CPI clauses we have there. So it is actually for the lion's share of the large customers, which are also doing the heavy lifting on the revenue. It's more -- it tends to be a longer-term undertaking. And this gives us this beautiful granular portfolio. So it's diversified in terms of when -- what customers ramping, diversified by geography, diversified by industries. And this is why that number is kind of remarkably stable, that cloud ERP suite number, because it's such a tremendously diversified portfolio, which is gradually sticking up to more and more revenues.

You see the visibility from customers. You know year 2, you have to kind of step up on businesses.

Yes, we have the ramp plan with the customer, how to roll out the systems.

Very helpful. And could you just talk a little bit about, we mentioned RISE, how much more attractive it is for you to sell a RISE contract versus a customer negotiating separately and choosing infrastructure, maybe the run of the environment?

I think with every customer is, of course, the discussion. I mean if you put yourselves in the shoes of a customer, you would basically deny all the upside from AI and using phrase that we delivered all the innovation we're going to bring to them and they would say, "Look, I don't care about this funky stuff going forward. What I care about is total cost of ownership. And I have my software license from you, which, by the way, you sold way too cheap to us." So it's a tough -- my biggest competition is software we sold too cheaply previously. And then we have an infrastructure contract and then we have the managed services on top. Now we have some unfair advantages. I mean first of all, the economies of scale are huge. We are much bigger than any of our customers in terms of how much hyperscaler capacity we procure. There might be some exception. I don't know whether U.S. government might be at a similar scale also. But any normal industry company is a fraction of our demand. We have 3 of them we use, but we have all these customers on the planet. Even the largest customers, we have a small fraction of our portfolio. Then the managed services, you can do so many things in one go, so to speak, and ensure all the compliance requirements, security requirements, all of that for many customers at the same time where if you outsource that for your company only to an Infosys or so, well, it's a much -- it's more economy of scale, basically. So if you fold these topics together, and also our commercial motivation that we say we are willing to put up with a slightly lower gross margin on cloud than on-prem, we can actually -- despite the fact that the customer might be TCO neutral, and obviously there's no reason for us to go beyond TCO -- or neutral because there's tons of icing on the cake for the customer once they are on a modern platform. So we try to do the contrary, we try to say, "Look, TCO is the wrong compassion. We want to have actually more than that because you get value from us on top." We can convert the revenues at 2x to 3x because then, also, there is an incentive -- also in the initial motion, sometimes, we can then sold in other products like BTP and so forth. So if you put it all together, this is what gives us the lift, and there is very few customers who are willing to do deals with us when there is no good business case for them, too.

Right. That makes sense. I think probably, the question I get asked most from investors is where are we in the cycle? We know that you've got a big installed base on support. They're moving to the cloud, going to S/4. Where are we on this journey? When does it peak? I think you've talked about this in terms of that support base and how that converts. Could you help us with that journey?

If you talk about conversion, you have to start from somewhere, and which is basically on-prem. And on on-prem, you pay maintenance, which, by the way, there's a little bit of a misnomer. Maintenance sounds like you are basically trying to reestablish the original status quo when you sold the software. No, it's a continuous development. And when we develop the software, it's actually changing quite dramatically over time. So a large portion of the code line over time is changed, so I prefer to call it unofficially continuous development of the product. And then we look at where that -- these revenues come from and who are the customers who are already paying both maintenance and subscription in the cloud. That's easy to figure out in our ERP system. And we see that -- that's actually kind of very roughly 1/4 of the revenue base in maintenance of the S/4 and ERP maintenance. There's a little bit of maintenance outside because in the lines of business, we tend to be cloud already, because we acquired these companies and we're predominantly cloud companies. But on ERP, that's the way you can get some feeling. And what it tells you is that, first, there is roundabout 1/4, that is only roundabout 1/4, that's paying both, and it still gives us now the kind of platform to deliver to our Ambition 2025. So it shows you with how little of that installed base you've been already going very far. It also shows you it's not all about conversion. There's a little of net new in there, there's a lot of cross and upsell in there. And 3 quarters is basically not paying any cloud revenues yet. And we have 2 paths for them. If you are on S/4 ready, it's a quite easy path to go to the cloud. So there, it's more about convincing them TCO is attractive, and you get all the functionality that you can't get in on-prem. So why don't you do it now? But it's not a business continuity issue. It's more about, do I need to leverage these tools now or do I wait for later and wait for other customers to toy around with it first. And so there's different philosophies in different companies. On the legacy systems, first and foremost, ECC, there is a little of a business continuity topic moving because we cannot support them beyond 2030. You know that ECC is running in third-party databases. We -- actually, the licenses for these databases are expiring in 2030. So this will -- they will be, well, then on their own, which is a big call. If you think about that system being a nervous system of your company, if there's any breakdown in the system, any security breach, any mishap, legal changes are not properly implemented, you have compliance violations, I mean, I would feel very uncomfortable as a CFO. If I ran any major parts of my business, it's like an insurance policy that SAP is really taking care of all these issues for you and just saying, "I don't need that anymore," and very few people tend to do that. And we even see that the guys, many people who have left to do third-party maintenance are actually coming back because they have some hiccups, and it's not the same as our kind of continuous development as I say that maintenance is really kind of plugging the hole in the system as opposed to continuous development the way we do it. So there, you have these both highly plausible transformation scenario, one being a low activation energy, high upside with new functionality from S/4 to cloud. And the other one being a pure business continuity issue, where, basically, they have to move either to our system or a competitor system. But then if you look at how much money it costs to really completely implement some other systems, it's even more expensive. And most of those who procrastinate tend to do it for financial reasons and they say, "I don't want to do it now because I don't want to spend the budget." But if they go somewhere else, it's becoming even more expensive. So I'm not very nervous about these people. They can, if they want, but it's -- I think we have a very attractive offer to help them with what they need to overcome.

And realistically, there are very, very few options for large enterprises in the market in any case even if they were to think about that.

But we have competition. We have a red company being competitive. But it's -- yes, you have also trained a lot of people on the way the system works. So it's a big effort. It's possible, but it's a big effort.

And on that 1/4 of the base that are paying maintenance and cloud at the minute, do you have any idea of where they are on their journey? Are they nearly done, halfway through?

I would say, my gut feeling from the data I see is that they are probably not even halfway through, even the ones who started because we talked about the time line, and when we started the whole thing, so it's already plausible from that end. And then -- and if you look at the relative size of the revenues they generate in subscriptions and they generate in maintenance, and then you compare it to 2x to 3x conversion factors, you see, okay, there's still a lot remain to do. And so even on that kind of, you could argue, already converted base, the revenues are by far not converted, yes.

So if we've got 75% still to go on and the 25% that have gone and not halfway, we're nowhere close to a peak?

Yes. I mean we are starting -- we are in the early innings of the transformation, which is not a transformation that for ECC customers, at least, will take forever because there will be some reaction needed at some point in time.

So when we do the modeling and we put 2x to 3x in the model, that obviously gives us quite a big range of outcomes. Could you help us? Maybe is there an average or could you maybe help us with what drives the upper end versus what drives the lower?

Really, it's also hard for me to prognosticate where exactly on that range we end up over the coming years because it depends on many factors. It depends also on the competitive dynamics. For us, revenue growth and margin are a little bit of communicating tubes. And margin, of course, that has also a certain impact on revenues because if I give bigger discounts, I have lower revenues, I have lower margin. And we have a certain aspiration to convert quickly because as long as we are not converted with the customer, there's still a risk that you have an on-prem customer with highly customized software, which is looking for a solution now in some areas. And I have these situations today with many customers. And then they might kind of plug in a competitor solution, cloud native still here. And I want to avoid that. I want to get as many customers as quickly. So I'm willing to sacrifice or give higher discounts if I can convert faster. And still, this is also why we created a little bit of freedom for us by saying, "Don't look too much at margin, look at absolute gross profit we generate," because I want to have that freedom. If I see that, I can accelerate more on revenues, and the margin is going down a little, but I have more absolute euros of gross profit, we're still going to do that. And -- so to cut a long story short, I don't want to be more specific on that because I don't know yet. And I need that flexibility within that range to really optimize the long-term profitability in absolute euro terms on gross margin. And it might -- sometimes, I might say, "Okay, it's running well, so I can aspire to lower discounts, higher revenues, higher conversion factor." Sometimes, I might need more pressure on the system, so I want to play with that to really make sure we keep the right speed at which we convert that customer base.

But the price for you of making sure they start that journey and then you can keep competitors out in areas is much bigger than what the initial discounts is?

Yes. That's the kind of prisoner's dilemma or bargaining lever with the customer. They know that when they move to the cloud, we will make more money with them. And so they can leverage that. And theoretically, I probably can negotiate it down to the point where you make the same money you make on-prem, and then you start kind of arm wrestling. And then the 2x to 3x, where you end up there, is a little bit of a factor of where you end up with every individual customer on that.

We talked a little bit about competition, but maybe let's dig in a little bit deeper. I get a view from investors that, yes, you've got the Workdays, and Salesforce is competing in the application layer. But actually, the more important battle for SAP is at the platform layer as we go forward over the next 5 years and ensuring customers are running BTP, so that you can cross-sell more effectively, you control how the data is managed. How do you think about the competitive landscape? Which is more important? And how do you manage that?

I think on BTP, the attach rates are very high, especially on RISE where we have more complicated customers with a lot of needs and different lines of business. And so I'm not nervous about that. I think you're right that to prepare for this AI game, it's all about managing the data in a smart way. And we think that we need to find ways that the data of different applications in our customer systems can be federated to keep all the semantics, all the context of the data. I always remind people that data without semantics is worth nothing. The answer, yes, is worth nothing unless you know the question. There's another complexity in our world, which is we have laws around cloud compliance, data privacy. In finance, we have some authorization rights, what people can see what. And not everyone can see everything in our SAP system. There will be a catastrophe from confidentiality or some other obligations. So you want to create -- you want to keep the data gravity in the system itself to preserve that context of the data to really do great analytics, which means you have to federate data from different sources. And there is data also we at SAP would like to have outside the SAP system, which would be very useful for us. For instance, if you have a -- some world disaster somewhere and you want to replan a supply chain, you need some data on earthquakes and stuff, which is not what we do. And we are in discussions with many vendors to negotiate these partnerships because the important thing is, of course, then the monetization, I mean, who is then getting what share of the pie? But for great analytics, you need highly structured data with good context. The access right discussion sounds trivial and not so important, but it's super important. If you really want to go to the extreme on AI, you have 3 requirements on AI. It has to be relevant. It has to solve a real problem. It's nice if you create nice videos and funky stuff. But if you can't make people more productive doing their job, it doesn't help. Then it needs to be reliable. In business, it has to be reliable. We cannot afford hallucinations. It has to be at least better than human beings in terms of being reliable. And then it has to be responsible, i.e., cloud compliance and all the things around it. And these are actually contradicting targets, if you think about it. And if you just egress data from all over the place into a data lake and do analytics on top, you have a big problem because the slightest pollution of that kind of analytics with something you have no right to see is killing the whole work. So you pollute the whole system, you have stopped doing it. On the other hand, if you don't use any data because you're scared that you're violating access rights, you have no relevance and you have no reliability because you have no data. So it's actually that important to really have a system that allows you to preserve that type of structure even into the analytics layer, and you completely destroy that when you just egress data. You can try to reconstruct the semantics, but that's a big effort, and some people try to do that. So I agree with you that, that part of the equation will become more important in our discussion. But I must also say, you see me sitting here with a lot of confidence because the gravity of data in our system is great. We have so many interesting sets of data that people want to see to run their analytics, and we have very smart systems to administer access rights. We have one sort of true systems, and we have a lot to contribute, which makes our bargaining position in these discussions really good.

And by the way, I think there's another question that I think you just answered, which is, can companies use AI to construct systems themselves and replicate what you do using a data lake? And that repudiation of the complexity of how you do that, it's just so enormous, that's a really interesting...

It's interesting that you mention that. I've been confronted with that thought. I come -- my kind of socialization in large and manufacturing companies was such that I was initially groomed with seeing an ERP system, it's clearly needed. And then at some point in time, I was confronted with the idea why couldn't you not put an analytics layer on top, you don't even need an ERP system. And I've seen companies toying around with that. I can tell you, I'm not aware of any company that has been ever able to do it. And I'm not very nervous about AI getting that done.

So maybe let's dig into the AI side. You've announced your CoPilot. Could you just talk about use cases that are resonating for customers?

Look, it's a whole bunch of use cases. What I like is a lot where we have millions of people sitting in shared service centers and doing stuff and then making them more productive, and that's one big lever. So to really make repetitive tasks much more automated and faster. And take a typical example in an accounts payable department we have a screwed-up process where the number of goods received is wrong, the prices don't match the contract and stuff like that. And if you see how people grapple with that today, they spend a lot of time on every single case. And with AI, you can analyze way more data and make them smarter. I always use one -- the other thing which -- where AI can play a huge role is when people actually don't do it so frequently. They're not very sophisticated doing it, and they are not qualified enough, and the system is basically babysitting them through the process. And this sounds a little bit innocent, but I'll give you one example I have been going through, how it works today and where I have a vision that this will be completely revolutionized, which is my business assistant had a whole bunch of access rights from me. I am the CFO, I have all the access rights, of course. And I need to transfer that for certain jobs to my business assistant. And then a new business assistant came. I had to transfer these access rights to the new business assistant. And the way it worked for me in the first place was we had a data and diary, the guy from IT came. Old assistant, new assistant, they are sitting next to me. We're clicking through the SAP system, the Microsoft Outlook, and we changed all these access rights. It takes us about 10 minutes. I say, "Well, that's incredibly wasteful." If I had Joule, I'll just say, "Joule, can you please transfer these access rights from old person to a new person?" It's done. Now the truth is, I was then told, "Dominik, you're wrong, it's not 10 minutes. We've actually spent hours this morning to figure out what access rights the prior guy actually had." So this is a nice example of where hours of work can be transformed into seconds. So for me, it's a little bit the next step of the Internet revolution. The Internet revolution, you as a consumer overcame the burden of going in a bricks-and-mortar shop, seeing where the product is there, being told it's not there, maybe waiting before in line to be told it's not there, going to another shop. So we now have mobile Internet, so you do it all. And it's the same journey in -- when you use these tools, you have an interface, which is hierarchical. You click through things. You need certain data from elsewhere. And so what we do now is we kind of shortcut that massively. And this is what makes me so excited, and now whether it takes half a year longer or not is not what I'm caring about. What I care about is how fast and who is best positioned to ultimately deliver that Joule or CoPilot of any nature. You know we have great partnerships on that. We are also integrating, for instance, Microsoft, by now. When we schedule a trip with our Conquer tool, boom, it's ending up in the calendar directly and stuff like that without the person touching it. So this is a big opportunity. People sometimes feel because they toy around with Joule, with CoPilots of other people, and they draft some e-mails. And when you reply to an e-mail, I find it very hard to reply to an e-mail using the CoPilot because it -- I know exactly what I want to write and I want to be concise. And boom, it -- I sometimes think it takes me longer to write the prompt than just writing it myself. But on these transactional tasks, I think with my example I gave you, there are things you can do where you can tremendously accelerate. And it's not all of that. There is also the process -- large process models we are currently developing that we are really looking at what is the guy in a shared service center doing. Anyhow that's independent from Joule, it's just figuring out what is the next step in the system. The same way on large language models, you say, if somebody said that, the next thing he's going to say this and that. You say, if somebody in shared service center is currently looking at that, probably the next thing is she or he is doing this, that. And then all the ecosystem we have, we have a huge -- we talked about the transformation. We have to transform our on-prem customers to a cloud compliant, upgrade-ready setup, which means they have to all rewrite their custom code from on-prem into ABAP cloud. We now train our CoPilot for ABAP. So if 1 million people coding ABAP and now they will become much faster writing that code. And right now, we have a huge demand for that because people have to rewrite their code to make it cloud compliant. And we can embed the cloud compliance in the way we train the CoPilot. Then we have all these consultants. I remember when I did S/4 transition myself on the customer side, we had 70 people from the system integrator. And frankly, half of them were not any good. We had to replace them. Probably than half today, if they have a kind of Copilot for consultants from SAP, probably they are good enough to do the job done because they have the kind of collective wisdom of all the other consultants at their fingertips. So there's all these things we can play off and -- on all these activities, we're investing big time. So I have the very comfortable situation that for the Ambition 2025, I don't need tons of monetization on these opportunities. I have also the conviction that once this is mature to a level where it really works, it has tremendous potential. We have 300 million users already today in the cloud, and we're just scratching the surface. I have estimated that the salary hike of that 300 million user base, and many of them using SAP a lot, is hundreds of billions. So if you can only get a small part of that productivity, and it's only the merit hike, it's not the salary base. So -- and you also know the order of magnitude of revenues we generate, it's kind of hopefully in the kind of outer years of a planning horizon, some mid-double-digit billion revenue number. So you know that 1% of growth is only $0.5 billion, and you compare it to these huge opportunities, and so why wouldn't that be something that could add several percentage points of growth in a 5-year planning horizon? I'm deeply convinced, if we do the right things and better than competition, and I try to explain why I think we have a little bit of an unfair advantage on the stuff we do, ERPs, we can certainly add some good growth momentum there.

So is the focus today, is it more getting usage and getting people into the cloud and using AI as something to pull them in rather than trying to monetize heavily today?

No, no, we also tried to monetize. But honestly, like in every early inning of a motion, you start to incentivize for people to get hooked up on it and say, "Oh, I need that." And then you can kind of groom the pricing. And over time, it's not the key to really extract the maximum value in '25, '26. It's really the key that we -- that basically, everyone says, if you want to have the most efficient back offices in the world, you need to run on SAP system, and there's no alternative to that. That's where we need to get to. And for that kind of word to be spread, you need a lot of users and -- who feel it every day. And this is what we're currently pushing.

And you just said if we do a better job than competition, then I'm very confident that we've got a lot of advantages, I can't resist following up on that. Could you talk a little bit around why you think you're much better positioned to do this then?

Well, it's really the kind of 50 years plus of experience on how these processes are running, the intimacy with the customer, the contextual awareness, the access to the data, the system itself. I'd say the S/4 -- the HANA database was the right tool that has invented. You have to think about it like a huge transactional universal ledger where every single transaction, be it the hiring a person, be it paying something, is registered in one source of truth. And the beauty of that, it's highly scalable. So on every transaction, you can put as many attributes as you like. And these attributes will later on help you a lot to be fast in analytics and not having to run expensive AI. I mean there was this myth some years back from some companies in the kind of analytics business, don't care about structured data. You can do everything with unstructured data. And it's probably right. But now you learn that the kind of -- it's like entropy in physics. When you've create a mess, it is quite cumbersome and compute-intensive to bring order back to the mess. So our philosophy is why don't we create order in the first place, and HANA is a perfect repository for all that data to bring perfect order in -- and perfect order in stuff that can be extremely complex with really almost an unlimited number of attributes to bring order into it. And I think the combination of all these assets is making us quite unique, and this is also what the customers buy in. And I think the numbers speak for themselves. I don't see any ERP vendor running at 30%-plus growth rates right now in the cloud. And even some other vendors who have been very high on cloud revenue growth, we are currently in the very best neighborhood in terms of what we can generate in cloud revenues.

So it's been a great discussion about the technology, and it's great to see such excitement from what's coming. But with the CFO, I suppose I have to ask a little bit about profitability and what's going on below the top line. You said you think SAP should be a Rule of 40 company. Could you define that? And could you talk about how SAP gets there?

Yes. I clearly said that if you look at benchmarks, a Rule of 40 is a good benchmark. And the way we define it is, basically, you look at the free cash flow, divide it by the revenues, and that's the free flow margin. And we add it to the consolidated revenue numbers -- growth numbers. I think we jumped off like '27 -- 2023. In 2025, you can calculate if you materialize it. We are pretty much more than halfway there. Now I have to make one caveat there. Some competitors are running around pounding their chests and saying, "I'm actually a Rule of 50 company." Few who really are, maybe Microsoft is better than that, but...

You're quite hard on yourself doing free cash flow, by the way...

By the way, I'm even harder if I do internal benchmarking here. When I do internal benchmarking, I say to make apples to apples, I also look at the equity settle stock-based comp and convert it to cash. And then some of the guys who pound their chest and say, "I'm a Rule of 50 company." Aren't actually, because they have 10 percentage points more stock-based compensation to sales than me, but it's different story. But I take the standard definition, and you see that we are covering more than half of the gap in a couple of years' time by virtue of our restructuring program. The rest has to come from accelerated revenue growth. So we did say that because of the very, very strong mix effect, with cloud ERPs being 83% of cloud revenues, with that kind of center of gravity of our business running at 30%-plus growth rate for more than -- now it's 11 quarters in a row, and the dilutive effect from on-prem licenses, dilutive effect from Infrastructure-as-a-Service business, which, contrary to our competitor, we're actually deemphasizing because we play the hyperscaler's competition against to get a good value for our customers. We can even afford a deceleration in cloud ERP suite growth numbers to still have an uplift on cloud -- on total revenue growth. So a little bit more growth on revenues beyond what is embarked in our ambition is a 12% midpoint outlook 2024 to Ambition 2025, with a 12% growth embarked. And we think we can do better than that in '26, '27. And then we said we want to contain the total expense growth relative to the revenue growth in the kind of 80% to 90% range, which we also see in our competitors. So it's simply saying we want to be as good as our competitors. And we have, I think, a whole bunch of measures in place that will enable us to do that. So the restructuring program we are currently completing, we are about to complete not only has the goal of making a big leap forward on that Rule of 40 logic, but even putting a gradient in place that in the following years, we can accelerate. Now I don't want to be nailed when exactly that will happen because between 80% and 90% fall-through in the cost, it's a big difference when you can reach it. Whether you add 1 percentage point acceleration from '25 to '26 or more makes a big difference. And I don't know yet. And -- but this is definitely something that it doesn't look stupid. And I also remind our people internally that our competition is not standing still. So maybe once we are there, maybe our competitors are even higher. So we have to keep on running all the time.

Perfect. I see -- just aware of the time, I'll maybe open up to the floor to see any questions to the audience. Take one from -- maybe let me carry on then. I think another question I get regularly is there obviously is a big restricting program this year, there's been some change at the management level at SAP. Could you just talk a little bit about how you manage that amount of change in the company in one go, the risks that some things that otherwise would have been spotted fall through the cracks? I think most people feel there is room for a lot of productivity improvements in SAP. But people aren't doing 0, they're contributing something...

No, no. It's a huge challenge to have so many balls in the air, so a lot of reconfiguration happening as we speak. You might have seen that. The way we think about our staff is that we are going to eliminate, by the end of this restructuring program, 9,000 to 10,000 positions. We'll rehire at a similar tune, so we will end up actually with slightly more people. Now we also did an acquisition, WalkMe, which gave us 800 incremental people. So we will end up probably even higher than that. So that's the massive effort. And of course, things can fall through the cracks when you do that. But I'd say there are some improvement areas within SAP, whether this functionality penalty is so high that the risk-return profile of tackling them is very attractive because you can allow quite some mistakes in execution and still be better off doing it. And I think I'll give you one example where I'm really convinced it's of ultimate importance to our long-term future and that we need to take them now is we have a great product now with GROW and our public cloud, multi-tenant offering. But our go-to-market motion is still quite heavy in terms of old style with a lot of people from lines of businesses. How do you want to sell a kind of 0.5 million TCV mid-market deal with the type of large enterprise go-to-market way of doing things, where you have to feed the mouth of expensive sellers in each line of business of SAP that doesn't work? You have to really do something radically differently, and it's not rocket science, frankly. We can look at our competitors who are more on that market and see how they do it, and we see a big gap. Now the issue is you might be bogged down in the day-to-day and don't find a time for that, and this is why it's actually very good. I think that now Christian Klein, our CEO, is taking over temporarily that sales function because he said, "Look, I know it's tough." We are not on so much pressure on some things that can screw up our numbers on sales. The biggest sensitivity on cash flow and revenue and profit is still license income. If you have tons of license revenues or not, we have been prudent in planning it. We have done very well so far. So we don't need miracles in the Q4, to put it mildly. So we've derisked that in the way we think about it. So we've created the room for a little bit of a disruption on some areas. We tell the people, "Don't worry about it. Let's go on cloud, that's not kind of -- if a license deal is slipping into Q1, not a problem." And so -- and we are forcing through these changes now. Changes are always unpopular. I have to confess you've seen a little bit our employee engagement index dipped. But I'm also confident that this will come back. I mean every transformation has a price. And I'm -- the whole team is motivated. The good news is why we are still also quite resilient despite some disruption is that the lion's share of the heavy lifting right now on the growth is coming from RISE journeys with large customers, right? And there, you have big teams on both sides. And if 1 or 2 people disappear from the team, there is enough kind of inertia, so to speak, to keep pulling the project on. It's not depending on one rainmaker salespeople now pushing the button. So I think I'm not too worried about that. I'm, to the contrary, much more excited about eradicating some dysfunctionalities, which were blatantly obvious to us and have not been tackled because we were also busy rushing for the quarter. And now it's the time to do it. And Christian is on top of it. So nobody dares to say it's the wrong thing because it's the CEO who's running the show now. So it's about transformation, as we peak. And I think it will -- I mean, there are so many good things that I observed since I come, how things have improved. For instance, in my shop, a huge issue is the quotes we give to customers. We want to double them over the next 3 years. And before, it was virtually impossible between sales and finance to converge a view, which is a global optimum, how you manage complexity in Ts and Cs and how you price and make it more algorithmic. Now -- and this was because everyone had their own resources to do that, and they all optimize the local optimum. Now we integrated all these operations in one cluster, and they go for end-to-end process and they optimize the global optimum, and it's tough, and I have good confidence that it will work. And I asked the team, actually, who of you thinks that we would have been able to get that level of confidence to be much more automated and algorithmic on that in the old setup. And people said, "Zero chance." They said, "It's not a done deal yet in a new system, but had we been stuck where we were, zero chance." So I see there's many things falling into place. This is -- I see the transformation as a much bigger opportunity than -- of course, there will be some disruption, but it's a very good risk-return profile in terms of tackling it.

Perfect. I'm bumping up against time there. Do we have one -- could you just take one question? Do you want to just -- I'll share the mic...

You shout it out, and he can repeat it.

I'll repeat it.

Can you talk about Asia Pacific at a particularly strong quarter? And what's your visibility for the growth in India, Japan, and you called out China as well?

Okay. Great question. You've seen that we did really well in terms of growth rates there. These are very, very different markets. I mean Japan, I think the opportunity is a super fragmented market with a lot of domestic tiny suppliers. I do believe that with AI, ERP becomes more winner-takes-it-all thing because what you need for AI is tons of data, and the smaller suppliers don't have the data to train their models on. So I'm excited about Japan. We've actually decided to invest more in marketing in Japan because we are not so big. But it's still the fourth or the fifth biggest market of SAP. India is, of course, how do we go in the mid-market, very many, many customers. So there, it's very relevant what I told you about reconfiguring the sales machine. I can tell you, with a large enterprise sales machine like SAP has it today, India is just off limits. And you mentioned China. China is a geopolitical topic, probably the trickiest one. So I don't think we can deal with that and zero...

We can offline.

But we are not too exposed to China in terms of revenue base. We have a lot of business with multinational companies who have their subsidiaries in China. Many of the Chinese customers or state-owned enterprises, they do a lot themselves. But it's not -- I think there's more upside than downside because we start from a low base relative to our competitive strength in other regions of the world.

Perfect. Well, I'll close there. Dominik, great discussion. Thank you, again. Really appreciate it.

Thank you.