So I wanted to start off, you know, you've been in the SAP CFO seat since March this year, so several months into the job. Could you maybe just give us a little bit of a perspective from an outsider's view coming in, you know, where do you see the two or three main opportunities to make a difference, maybe one or two of the big strengths of the company that you've found, just to help us out with your experiences so far?
Sure. I mean, I've been somewhat familiar with the company, as a long-standing customer in actually all the companies I used to work for before. You alluded to my experience at Infineon, where I was also in charge of IT, and we did a very, very large S/4 transition. I think the overall impression I have is actually, if anything, to the positive because we have very strong customer traction. I think you can very nicely see that in the cloud revenue growth and the CCB growth, 25%, over the last quarters, and if you compare it to market data, it's clear that we are recovering market share on the ERP front.
I think part of the reason why that's the case is that with the advent of AI and the shadow AIs casting on our customers, they really wanna make sure that they have the right tool in place to leverage all the treasuries they have in the data in the ERP system. So that's one thing. The other thing is, we have that ambition for 2025, and you have to ask yourselves, what are the opportunities we have to protect that ambition to make sure that we can do the right things on the cost side. I think there is a good understanding by now what can be done, which gives me even better confidence that execution will come along properly. So, I think the best is yet to come.
We have, of course, still this huge installed base of ECC customers and on-prem S/4 customers. You know, for ECC, maintenance is running out. The official maintenance period is running out in 2027, and we're currently really working with these customers on the transition plans. Of course, sometimes there's a propensity to procrastinate when the macro is difficult, but given the macro backdrop, I think our ability to convert that base is quite strong, and you've seen it over the last quarter, so my confidence is actually very high.
Perfect. That's helpful. So you mentioned S/4 migration. Maybe let's dig in a little bit to that. So, you know, we've seen some macro headwinds over the last 12-18 months. Other software companies have suffered. I think what's been particularly impressive for investors is the strength of that S/4 business, how the backlog's grown, how the revenues have grown. Could you just talk a little bit about why companies are willing to invest in this product, why are they starting with it, and could you give a couple of examples of the benefits they get out of it?
Yeah. Yeah. Let me use a little bit of an analogy, which is maybe fitting in this room here, and it's not quite perfect, but it's the best I could find why it's I would call it a killer application. We discussed the value of data, and if you wanna analyze data, you need to have access to that data. All of you are probably very familiar with spreadsheets. Spreadsheet is much simpler than an ERP system, but there are some similarities. You probably build models of SAP and other companies, and depending on how much time you invest, the sophistication of these models can be quite high. There's a lot of intelligence embedded in the formulas, in the links to other spreadsheets, macros to update for certain scenarios, and all that type of stuff.
And now, you take the kind of income statement, balance sheet, and cash flow statement in that spreadsheet, and you copy that. You open a new spreadsheet, and you paste the values. And now I ask you the question, you as a human being, if you wanna run analytics intelligence on the tool, would you prefer that kind of paste value spreadsheet, or do you wanna have the original spreadsheet? And I'd say most people who have some sense would say, "I wanna have the original spreadsheet because there's much more granularity, intelligence, drill-down capability." Now, our ERP systems are much more complicated, but there are some similarities. For instance, this thing changes every millisecond. Every single booking is flowing into that thing. In every cell, there is a lot of other data embarked, which you don't have normally in spreadsheets.
You have to regulate who can look at what cell and manipulate it in a very granular way, so it's a much more sophisticated tool, but for analytics and artificial intelligence and reasons, customers see that having the ability to go into that tool and drill down, if need be, to the transaction level to understand what's really happening, to slice and dice the data by geography, by product, by even salesperson, is very useful, and to connect the dots in that ERP system to really glean insights, automate processing of the data, and all these things, so that's the beauty of the tool, and I think that's the secret sauce behind why people invest a lot of time and effort to do these transitions, because they just wanna have the assurance that when they need all that data, they can access it, and they're not missing it.
There was, I think, a conviction five years ago. You take these kind of pasted-value spreadsheets, and you run some analytics engine on top. With the power of the tools we see today, it's just not doing the job. Most of these approaches have actually failed. What we also have gotten much better at, at SAP is to say, "Well, the analogy is the link to the spreadsheet." If you need some data from Google or from other Databricks or Snowflake, what have you, and you wanna federate that with our data, we actually give you the tools to do that without you having to do the cut and paste. We federate the data. We will ourselves provide you very powerful use cases. If you wanna consume them quickly, you have to do it in the cloud model, of course.
By the way, we also have to do it in the cloud model so that our models can be trained. Otherwise, I don't have access if it's kind of hidden in an on-prem installation, and so from that perspective, this is, I think, what they also see, that we offer them a journey to bring their completely very complicated legacy structures by the Clean Core approach, by segregating what's customer-specific versus the Clean Core of SAP in a kind of,
I'd say, upward-compatible mode, so whenever we upgrade, they don't need to change all that stuff, but there are application programming interfaces that can link to that, and this is the overall package we give to them, which they find extremely compelling, and it's not only lip service that this is a vision we have. We can show them it's working today, not perfectly yet today, but that we are in the pole position to make sure that two, three years down the road, it is actually perfect.
That's a pretty compelling sales pitch for me to get my SAP model into S/4 then. The other thing that's interesting is, you know, we've.
Doesn't fit, but it's good.
You've got some enormous customers that are on a multi-year journey. You know, we look at the annual contract value that you sign, but you're doing multi-year deals that give you much more visibility into what customers are doing. Could you help us a little bit looking further out? What visibility do you have? How do those deals ramp up that we can't see?
Yeah, so, you know, we have this ambition for 2025, and we have a guidance now for this year. And there is a remain to do in terms of growth. And, I'd say that everything that we've contracted now, plus the transactional business, which is also quite sticky, where we don't need to regenerate new bookings, but it's more dependent on how much people are traveling and so forth, plus some reasonable assumption on renewal, if you embark that and then say, "How much incremental new bookings we need from now till then to hit that 2025 guidance?" Basically, what we have in the books today is more or less the revenue base of 2023, sustained into 2025. So the growth part of it is something we need to do more or less with new bookings, so very roughly. But these new bookings are not black magic either.
It's the conversion of the installed base. It's the growth. So it's two big levers. The biggest lever by far is the conversion of the installed base. You see it from the maintenance base, which has actually been remarkably resilient. There's still a lot of fish to fry there. And a good chunk of that is on the famous ECC, the old version of SAP, which will run all the maintenance in 2027. So in discussions with customers, I don't really see a discussion about, "Well, we don't have to do that. We like other products better, and it's easier for us to go to a competitor." It's all the same discussion. It's, "Look, we're it's still working, and I don't have a budget now.
Let's do it later," or, "I have just invested in S/4 upgrade on-prem, EUR 300 million, EUR 280 million for the system integrator to customize, and EUR 20 million for SAP for the software. And why should I now invest? And and I, I can wait later. And why by the way, wouldn't you extend maintenance anyhow?" And I say it here again, personally, only over my dead body, yeah? And I'm very clear on that front. And I tell every SAP employee who's seen customers before leaving the room to say that to the customer, because sometimes I get blamed by customers that they don't. We've not told them clear enough. So I say, "We tell them now clear enough again and again." It's more about when they're taking that step.
Right now, the new bookings is really about convincing customers that now is the right time to go on the journey. Don't wait till 2029. And that might mean that sometimes we give some support in terms of, yeah, helping out with our IT services to make it easier or even commercial support, because we don't wanna have that J Curve ahead of us, in 2027. So I feel we have a good pipeline. The more we move out now, we start seeing kind of the degree of implementation tracking on the conversion for 2024 now.
So we still see the leads for 2024. We start making models about Q1, Q2, Q3. And we have certain KPIs at what level of maturity we have to have each opportunity, each lead to convert. So this is why I'm not really nervous about adding that kind of revenue base. But it's big heavy lifting, because people, again, in the macro environment, such as the one we have currently, they tend to procrastinate. But so far, we have been able to track along the plan.
The other part of the discussion we've talked about moving to S/4, starting that upgrade process. When we speak to partners over the last 12-18 months, the other feedback we've had is the product integration across the SAP portfolios improved a lot. Could you just talk a little bit around what are you starting to see around the cross-selling and how, when you get in with S/4, you can move and sell other products?
Yeah. Maybe to put that my answer to that question into context, if I'm coming from the outside as a customer look at SAP, I think it's important to understand the historic context, yeah? We have been, for very good reasons, that kind of ERP powerhouse on-prem, then we had an activist investor in 2019, which was pushing us to do some homework on cost and improve the profitability, and that kind of led to a procrastination of the move to the cloud, because that move to the cloud is very expensive, and if anything, pushes revenues to the right, so it creates a J Curve, so we had, I would say, almost a near-death experience with the transition to the cloud.
And luckily, some people, and I was not there, so I can't take any credit for that, in October 2020 said, "Okay, look, I mean, it will not be appreciated by capital markets, but that's what we have to do. We have to invest in the future. We have to bring that to the cloud." And it was right. Why? Because you're doomed to die on an on-prem model. And we have suffered from that in the past, because our customers, as I mentioned, have been heavily customizing our software. They have taken the software, customized it heavily. And whenever we had to upgrade it to use new features, they did it like five years later than we have delivered it to the market. That meant that they were looking at an old SAP product on-prem versus a fresh, brand new, cloud product from an agile competitor.
This is how we lost market share around the fringes of our ERP system. Sometimes we've even completely left the terrain to some other competitors. Now we solve this problem with the Clean Core strategy. By the way, I'm really preparing some more communication with capital markets on how exactly that works, so it's clearer. The BTP, which means the partitioning into Clean Core, and then the ability to put that customization extension stuff into a compartment, which is kind of almost upward-compatible with our upgrade. So we kind of shed that slow break we had on our own velocity of innovation with the customer.
And now we are trying to turn the tide around and say, "Look, and there are so many synergies." And I always like to use, for instance, our so-called SAP Analytics Cloud tool, yeah, where we have an analytics layer on top of SAP, where, frankly, a lot of customers have gone to competitors of ours, doing that, because they had an agile, fast, modern cloud solution, and we were stuck on-prem. And now we turned this around and tried to expand on that core, yeah? And that's the key strategy we are pushing here, so that is what we call land and expand. So the land is really to make sure that the customer is embarking on the right journey. S/4, BTP, and you see the platform as a service growth numbers showing you how fast we grow there. But that's not the whole story.
There is a whole bunch of very transactional, close-to-ERP solutions where we have lost some ground to competition, where we're currently playing catch-up. And, I think our journey in terms of also trying to make it visible is that we'll talk more about that, how we kind of integrate that. And now integration has been a key focus the last year. So it used to be, I'd say, back in 2019, 2020, lip service. So, we have bought all kinds of applications and never integrated them properly. Didn't have a BTP, didn't have a clean core strategy, didn't have Datasphere. Now we've invested like hell from October 2020 till now to have these tools. So we can land now and expand and can offer very compelling value to the customer.
And there's one value item, which is quite easy to understand, which is that we can also give them a pool of credits and say, "Well, you might not exactly know where you consume it. And I give you some flexibility to shift between applications, within that core ERP offering." But for that, it's good if you take it all from us, because then you are flexible. You get some more flexibility. And that's something very few, to put it mildly, can do. So we're really trying to build that kind of cohesive land and expense strategy bundle and try to fight back on the fringes where basically competitors have been eating into our lunch.
You say very powerful against those point vendors that can only do that on one thing.
Exactly. Exactly. Exactly. Because they would not, for instance, be able to say, "I give you a bunch of credits, and then if you underconsume here, you can put it on that one," and also, there is some benefits in the product, because integration per se costs money and is kind of reducing your capability to run AI use cases, actually. It creates breaks in the system.
It's a little bit like to stay with my spreadsheet story when you have two spreadsheets which are actually from a different vendor, and they don't talk to each other. Then somebody has to build a glue logic between them. This is why I think that's the next stage of the rocket that we can demonstrate. Well, the land, I think, is quite visible now, but the expand, I think, we can give you even more meat at some point in time to show that it's also working.
That's working.
Yeah.
Yeah. That would be helpful. On the transactional business within the cloud, you know, that's been a drag. I think you were saying it was only slightly growing in Q2. It was down in Q3. Could you give us a little bit of a feel for how big that transactional business is within the cloud and what the drivers are for that to reaccelerate?
So it's a kind of high triple-digit million number, and we don't disclose precise data, but to give you an order of magnitude. So it's not huge. But of course, if you have, I don't know, EUR 800 million-ish of revenues, which are basically flattish, it means that it dilutes the overall cloud revenue growth. And you have to understand that this business in some pockets is actually quite non-linear. So if you take Concur, there's a subscription part to it, but there's also some overages. And as the word overage already implies, it means that if you're doing a little bit more, you can have a healthy uplift on revenues. But if you then consume a little bit less, it can have a disproportionate reduction. On the contingent workforce, we already commented. We have this product called Fieldglass.
Even on Business Network, where we have also a little bit of a kind of business model transition, which is a little bit like this J Curve we had on the ERP system, it's lower. And of course, on procurement, it's also depending on GDP growth. I mean, if you procure less, then it's less transaction revenues. So I would say that that business, on average, is growing maybe not as fast as our standard subscription model, but growing quite healthy. So normally, it would be not a very noticeable dilution. But if it's because of macro really flattish, it is a certain dilution. And it kind of is, besides the famous Litmos divestiture, which we had last year, where we basically divested the business last year, which is making the comps more difficult.
The other reason why, it's quite easy to bridge from our cloud revenue growth, which is about 23% in Q3, to the CCB, the current cloud backlog growth, which is a forward 12-month forward-looking subscription-only growth indicator. So if you take into account the kind of dilutive effect of the transaction business, which we think will reverse at some point in time, plus the difficult comps, because we've divested the business in the prior period, you see that actually, it's quite consistent with the CCB growth, yeah? So you can say the subscription business is actually growing at that kind of mid-20s growth rate.
It would be Q2 next year that you have the base comparison start to ease from that transactional business point of view.
The transactional part is really dependent on macro. So, will it? I mean, it's tough to imagine it staying like zero forever, because that would mean really bad macro environment.
That's hopeful.
Then I think we have a different story to discuss. And then on this Litmos divestiture, I think it closed early December. So, I think, in Q1, it will be already a clean comparison. And that headwind in comparison will be out.
Perfect. You mentioned this before, that the support business at SAP has been very resilient, and I think a lot more resilient than we, in the investment community have expected. Could you help us understand that a little bit? And I think, you know, one issue would be, is there a risk that people are paying for premium and double maintenance, paying maintenance on subscription while they go through an upgrade process?
No. I mean, when we structure these journeys from the on-prem world into the cloud world, private cloud many times first, we exactly give the customer a very granular conversion schedule on when they can kind of reduce maintenance payments and then switch to cloud. And the customer is not one big instance where you suddenly have, day one, everything on-prem, and then the next day, everything on cloud. It's different parts of the company which are gradually rolling over. So it's a very nice granular portfolio. Inconceivable that there is, like, some cliff or something in that business. What matters, of course, is also the evolution of the pricing on maintenance. You might have seen us increasing prices 3.3% last year. Now, for first of January 2024, we have set max 5% CPI, but capped at 5%, which is completely compliant with our contracts.
I must confess that sometimes in commercial negotiations, some customers have been arm wrestling and then gotten some concessions temporarily on that. So it's not necessarily 100%, but a very high share of our customers. And so it's a lagging thing. So we adjust 1st of January for what happened like half a year back. Now, with inflation falling in most of these countries, I'm not expecting that to continue. But it will also mean that on the cost side, we have less uplift. I mean, we're currently also discussing next salary rounds. And of course, when we discuss salary rounds, I want to discuss what is the expected inflation in January of 2024 and not what it is now, because it's coming down like hell. And so it's not something that will die quickly, but it's a very gradual thing.
And we had actually also stronger license revenue. So, some customers are still very, very stubborn and say, "Well, having everything under my control is more valuable to me than all your nice pitches on cloud." And of course, we support them. And by the way, we have to, because otherwise, it would be also not the right way to deal with customers. So it's remarkably resilient, but it's not that we have double payments with customers. That's exactly the commercial deal that we say that maintenance payment is converted as a certain factor to cloud. So if you take three times the cloud revenues from us for the maintenance, you can we will you will not pay maintenance on that anymore. But it's normally not happening one day, but when the different parts of the company are switching different modules of SAP and different legal entities and so forth.
I guess that reflects your installed base. These are big, complex companies that are going to go on a multi-year journey.
Not all of them. We have also some famous startups, but they'll be more on the public cloud journey, which is a different game where it's normally greenfield, really, that they start. We source that. And I'm also very proud that we have great traction on the public cloud with the big professional services system integrators, which I would argue should know what's the best product in the market. And so that's another growth avenue is really to offer these for these less complex business models. And of course, our ambition is to make our public cloud better and better so that then even more complicated companies can use it. And we have some first cases.
I think we have go-live in February, with one huge conglomerate, for instance, which decided, "I want to go public cloud in one region," and we are pretty convinced this will work well, and the more evidence we can show to the customers, cool, and the more process performance improvement we can demonstrate to the customer that the public cloud solution fully standardized is giving them some real efficiency, it's cheaper because it's on a multi-tenant installation, which by definition is much cheaper than a private cloud, so that kind of journey will continue to the public cloud for most of them, but it will take a long time.
You talked about competition before and some of the people kind of at the edges of ERP, you're losing share to them. But I think, you know, when I speak.
Have been losing.
Have been. Sorry. Yeah. Fair. But I think when I speak with investors, a lot of people feel that the competitive landscape and the battleground's going to shift. And instead of competing in that application layer, the bigger battle is going to be at the platform layer. So where you have BTP, you're trying to control data analytics, insights. And then you have people like Microsoft and other hyperscalers who are also trying to control that platform layer and data and information and insights. Could you talk a little bit about how you see that, how do you agree with that and how we can track that battle and how well you're doing in executing and getting your platform?
I'm always paranoid about competition, and I try to be as humble as possible. So I'd say we will continue to see strong competition also on the application level. But what you're saying is right, that the real fundamental debate right now for our customers is how do I future-proof the company? And it's like building a house. You build the base first. And that's about data management and all the things you described, and the analytics capabilities. And this is why I tried to use this spreadsheet analogy before to tell you why I think it's pretty damn hard for customers to deny that what is in the ERP system should be dealt with by SAP, yeah? And but of course, you can then say, do I do the planning on top of the ERP system with a different tool?
And the competitors you mentioned, they have their tools, or the visualization of this data in a different tool, because I use that tool for manufacturing, for all kinds of other things in the company in one uniform layer of interaction, or the copilot discussion, yeah? I mean, it's a little bit like in the old days of the internet when you start, we discussed about the start page. Is it, do we put, can I make sure that the employee is kind of going on the SAP start page, or is it he's or are they on Microsoft Teams, and then they use the copilot to talk to my system? But the good news is we are actually cooperating with these guys because they need our stuff. And so we have some leverage in discussing how to monetize this.
And we have our own ambitions. For instance, on simple stuff, you can argue you take a Microsoft Copilot and then, you, you talk to this thing and say, "Please feed the following data into the SAP system to generate, I don't know, whatever document." And then the large language model can use that and stick the data using an application programming interface of SAP into our SuccessFactors, if it's HR, and then create something, yeah? We, of course, say, "No, no, you want to talk to our Joule copilot, and you tell that guy what to do." I think on easier cases like drafting something, you can also use another tool and then use the API to, to connect. But if it's really complicated stuff on the process landscape, what's the benefit of going through that bypass just to have a uniform interface?
I think I feel quite comfortable. Not comfortable is the wrong word. Competent that we have a super strong position in that battle. I like it, actually. You see it in our numbers in the platform as a service in the SaaS business. This is why we acquired Signavio a couple of years ago. This is why we acquired LeanIX recently, is to really be very strongly embedded in the customer, understanding the tech stack of the customer, being a little. We are a little bit more customized. I think the intimacy we have in terms of what's the specific need of the customer is higher normally than general standard Outlook product, yeah, or Office product. This is what we play there, that we say, "We know your company inside out. We know every process.
If you want to run use cases about around these processes, why do we go through an intermediary? That's it. It's like you have an architect. You build your home, and you want to change something. It's easier to do it with that guy than to call somebody else to come in and then try to figure out where's the plumbing and what do we do with it. We're going to try to leverage that hard, yeah. I feel pretty good about what we can achieve on this.
Yeah. That advantage of understanding the processes of the companies is obviously critical, yeah? That the discussion around copilots moves us nicely into a little bit of a discussion around AI. Could you talk a little bit around that? You, you've talked around pricing with maybe being able to get a 30% premium on apps. What is the pricing strategy? Maybe let's start with that.
So, the famous 30% premium is a premium bundle we have put together where we do two things. We basically expand the kind of minimum core S/4 offering to something that is allowing the customer to leverage the tool in a more comprehensive way. We add the journey on the Green Ledger, so the sustainability reporting, plus the AI.
Right.
So it's all linked together. And honestly, I'm not expecting wild leaps, and we don't need miracles on that front for our 2025 guidance. It's a longer-term thing. I think it will be a tremendous growth opportunity in the second half of the decade. What matters now to kind of leverage AI for our revenue growth to get to 25 is exactly the platform discussion that people say, "I'm betting on that SAP platform. It's future-proof. It's the right company. They know what they do." And that, of course, puts us in a pole position then to create that value on top of that platform. We have to, of course, do some early showcases so the customer has the faith, "When I buy this platform, I want to see a little bit what I can do with it," yeah?
Of course, we make some revenues already, and we have already 70, 80 use cases, and customers use them. We do that also internally. For instance, I always bring the fabulous example of three-way matching. So if the supplier sends us an invoice, we have the procurement contract, we have the purchase order, we look at the goods receipts, we compare it, and we then pay, yeah? And it's not so easy, but the so-called bypass ratio that we don't have human touch in this process anymore is increasing. So these are the type of things the customers want to see that they believe in us. It's not a huge monetization opportunity in the near term, but it will exponentially inflect. And there are, I'd say, five, at least five big levers on AI, which I think we can play in a powerful way.
I mentioned automation already, simply getting human beings out of the way and optimizing the processes. There's, of course, the coding, yeah? I mean, it's an opportunity for productivity for us. We need less people to code the same code. But it's also a big opportunity for our customers who we have to program in the old times in ABAP, now BTP. These are quite complicated, very sophisticated tools. And you used to need a black belt guy for that. And now we democratize that and say, "With AI tools, somebody can express, the needs of the customer." And then based on all the kind of other customers that had similar problems, they're coding something. And then, maybe one black belt guy is still needed to fine-tune, but it can boost the productivity of these people massively. And, you know, these resources are scarce.
There is a topic around assurance, internal controls. We often forget that SAP has a lot of very sensitive topics, on governance, on accounting. And the regulatory fantasy of what type of reporting and, also monitoring efforts you impose on people is humongous. And if you're not automating like hell, your overhead will explode. So that's the other thing. Then we mentioned the processes. What I always find extremely compelling is to show a use case to the customer and say, "This is your process end to end. This is what you do today. This is the target stage." And this should not be static. We should really, use the richness, the wealth of data we have on these topics with the customer to train our own models on how to optimize the customer process, which is very much about finding benchmarks.
Mm-hmm.
That are as close as possible to the customer's real IT stack and the processes. So this is why we have Signavio and LeanIX, so we can analyze that.
Right.
Get the data from all the customers. And we have the consent of the majority of customers on cloud to do that. And then tell the customers, "This is what you have today. This is your ugly animal here today. And this is the nice thing you can have in the future. And by the way, that runs on 50 other customers like that. And there's no reason not to believe this." What have I forgotten? Of course, the faster decision-making that you have better analytics and come to a very good, intuitive decision-making, and then the copilot things we discussed. So these are the things. I would argue our problem is not that we don't know what to do with AI.
Our problem is there's so many things that we have a task of prioritizing properly to think what's the lowest hanging fruit and have a good balance between showing some quick wins near term to get the customer on the hook. Say, "Yeah, that sounds great," but also do the J Curve on the more complicated stuff like the process mining I described, which I think is super powerful but will not be dealt with in kind of one, two years. But it's a longer journey where we have to train models and, do it ourselves, yeah?
We talked a lot about technology. I've got a CFO on stage with me, so she'll probably ask some financial questions as well.
Of course.
Maybe let's start off. The question I get a lot is around the cloud gross margins, and there's a debate around, you know, you have big customers. They've tended to go a lot into private clouds. You've talked about a customer there having low 60s% gross margins, and that makes people nervous that that's the model people go to. You're not going to be able to hit the cloud gross margin targets in 2025. Could you help us with how you get there?
Yeah. I mean, we think we will hit the ambition 2025. Otherwise, we would tell you.
So.
And I think the kind of near-term trajectory diluted for some special projects like the cloud harmonization journey is actually pretty, pretty consistent with that trajectory. We do have a strong growth in private cloud, but we also can achieve by virtue of this strong growth economies of scale on the private cloud and improve the gross margin there. And as I mentioned, we are deeply convinced that the private cloud is not the end of the journey. And then whoever is on private cloud over time will try to get on public cloud step by step.
So, yes, it's a challenge, but it's something we have enough self-help levers, including economies of scale and also better conditions by virtue of larger scale with hyperscalers playing that competition very tough, sourcing more cheaply, harmonizing our own lines of business, though we can not use different infrastructures with different lines of business but using the same. So I think we have enough self-help on the gross margin debate. I'd say that even allowing us to use our commercial cloud to be aggressive with customers in that kind of effort to pull in demand from 2027 to convert, so I think we can have a reasonable confidence level on the gross margin, while not pricing ourselves out of the market, I'd say, and jeopardizing the top line because of that. So it's work, but we have measures, and we track them, and they are on track.
On track. And then the other area I get a lot of questions is around the free cash flow. That needs to grow ahead of EBIT out to 2025. Again, could you help us? You're aware of the areas you're focusing on to improve the cash flow conversion of the business?
Yeah. So, on cash conversion, of course, I mean, cash itself is very much driven by profit. So the first thing we have to do is to improve the profit and hit our guidance for the operating profit, non-IFRS operating profit we've given for 2025. And then there's the cash conversion. There are some super low-hanging fruit. I'm almost blushing if I tell you because you know, like, the lower stock-based compensation, but that's more cosmetics because, frankly, in the end, we have to repurchase the shares. What we really want to improve is stuff like working capital, CapEx to depreciation. And, on working capital, we have, from my taste, too many overdues in the company. We really want to kind of reinforce the collection efforts we have. I mean, people are really dependent on our system, so I don't see why they wouldn't pay us on time.
I think it was not really the focus of the company before. We have been, sometimes not focusing on cash and purely on profit and for very, very low implied returns have been prepaying stuff. And I will stop that. On CapEx, I can give you one example. You will love that. We are probably the only company in Germany or maybe on the planet, I don't know, which are buying the company cars for our employees. And first of all, almost, I mean, the part of the population, which is entitled to a company car, is huge at SAP, much higher than anywhere else. And of course, in the benchmarking, we tend to forget that kind of goody, and then add the salary on top. So there's also a discussion about some of the perks we give to employees.
Then there is the things on our balance sheet. So we have to dole out the cash for buying the whole car. I mean, I understand that IFRS accounting is also putting the value of the car on the balance sheet, but only for the value of the lease, which is, of course, not the full value of the car. And then comes the best, if we then divest the car and in case we have a residual gain on the residual value, the employee gets the money. So, it's not very efficient on cash. It's out of standard in terms of compensation policy.
So these are things which can help. It's just to show that there are things that can be improved, self-help things that are just grinding on working capital, making sure that we say, "What's our priority? Am I a used car sales company or am I a software company?" And there's a cloud company, and we will turn over every stone. And then, I'm pretty confident that the compound effect of all these measures will get us to where we need to be in 2025.
Perfect. We're bumping up against time. Maybe I'll just see if there's one question from the audience to close things out. Anyone got a question on the floor? Can we take one in the corner over there, please?
Yes. Yes. Question on your, let's say, product roadmap. What can we expect after S/4, after, let's say, the 2027 deadline and a lot of customers have migrated to S/4? What can we expect after S/4 in terms of product roadmap? S/4 is very much stretching out into the '30s. S/4 itself is not a dying species. And this out-of-maintenance topic I was describing in '27 is on ECC, yeah? And there are still a lot of customers who have not transitioned to S/4. And frankly, it's a pretty complicated transition. I mean, I've been doing it at Infineon a long time ago. And it's a very comprehensive project, and you need some pretty solid operational guys to get it done. And that's the thing with the out-of-maintenance '27.
On S/4, we also have S/4 customers, where it's not the out-of-maintenance that's the stick. It's really the kind of recognition that as long as they're on-prem, they will not have Green Ledger. They will not have any AI use cases because we're not going to invest in that because it's kind of legacy structures where the whole system doesn't work. I mean, why should I start toying around with AI if I don't, I'm not even able to get access to a lot of customer data, which I can use to train my models? I mean, this is, it kind of doesn't work. So it's about now if you take these ECC customers, what happens in 2027.
There is some prolonged maintenance that which is a limited maintenance at a higher cost. So that's not necessarily bad for us, but they will need to work on the platform. You mentioned it. That's the thing. Do I have a kind of future-proof platform? So they can kind of procrastinate, but sooner or later, I'm really convinced that these companies who will just go ECC, then prolonged maintenance, then third-party maintenance, they will have compliance issues because they will not be able to maintain the software in 70 countries of the world.
And they will not pay their VAT anymore properly. And the audit committee will shout at them. And honestly, they will not be able to use anything, of cloud technology coupled with AI. So if a company had that strategy, I'm really saying but I don't think they will let that happen. The typical constraint is we have not the resources. We don't have the money. It's more about commercial. Where is the business case? They look at the old, on-prem business cases of S/4, where, frankly, the use cases have still been limited.
The power of the technology has not been fully harvested. But now with AI, there is much more potential in that. And it was an inefficient deployment model with system integrators stuffing their pockets with customization, so bad IRRs. And they say, "I look at that IRR. Why should I do it now?" And then we have to evangelize and say, "This is a different model," and show them concrete improvement potential and say, "This is the business case you can run for. This is a transformation project we sell to you where we need the system integrators to do it with you." So I bet they will come.
So far, touch wood, in my career, I heard threats about people going to Oracle, but never any serious ones that ultimately when I then held the line and said, "No," it was not anymore about. My typical response to that is, "Look, do what you think is right for you to do. We want to keep you as a customer, and we can talk about commercial aspects. And I will. You're a great customer, and I will do everything I can for you commercially, but I'm not going to kind of change my strategy for you." And by the way, Oracle, if Oracle doesn't, and by the way, if you want to have the choice between different hyperscalers at the lowest cost, well, we are completely neutral. I don't have an infrastructure to sell to you. Somebody else has.
So I've never seen a customer then saying, "Oh, so it's worth the effort." And don't underestimate that the transition cost from the on-prem installation to the cloud, if you are SAP to SAP or Oracle to Oracle, it's really much bigger if you change the supplier because there are some ways processes run in these tools which are different, and then you have to change even more. So why would the guys who are the most penny-pinching and most budget-constrained then do the more expensive move? So I've really not seen that.
And from that perspective, I think it's not a question about if, but when, and how can we make it smooth for them, and how can we haggle to come to the right trajectories and how we phase the old world into the new world, and convince them about our toolset being the only toolset available to them to efficiently migrate. And this is also something I feel as SAP, we've not done a great job yet in making that visible to you, the investors. I want to make more work on that to show you really how technical that works, because it's quite impressive and unique. And then you can ask competitors, "I mean, how do you do it?" And they will say, "Well, we are cloud, and it's just greenfield cloud." That's it. Yeah. So, sorry, I'm overtime already, but.
No, no. So we're, yeah, unfortunately, we'd love to carry on the conversation. We've bumped up against time. Dominik, thank you again for joining us.
Thanks for having me.
Appreciate your time.
All thanks.