Columbus A/S (COLUM) Earnings Call Transcript

Welcome to today's presentation where we have the pleasure to present Columbus. As very much seen here on the front page, the financial results for your Q3 didn't look like much market recovery in the results, but I think I read between the lines a little bit optimism about the coming quarters. So maybe we'll get into that. To help us through today's presentation, we are joined by CEO and President, Soren Krogh; and CFO, Brian Iversen. [Operator Instructions] But for now, I'll hand the call over to you, Soren.

Thank you very much, Michael, and good afternoon or good morning to everybody, depending on where you're sitting. Yes, as Michael was just saying, I'll take you through the Q3 results and some year-to-date results for the first 3 quarters of 2025. I'll start with a brief run-through of the financial and some operational highlights. I prepared a few, I would say, use case, customer case examples to explain some of the things we are working on for customers, then Brian will take over, go into more financial details run through, and then we'll have the outlook presented by Brian and finally, the Q&A session. So the quick disclaimer for you to run through about forward-looking statements. And then -- I have a little bit of an issue. All right. I think we're there, right? So okay. So our revenue in Q3 of 2025 declined by 7%. Market conditions in Denmark and Norway primarily remain challenging. But we're starting to see some interesting signs of recovery in the key market of Sweden, which is our largest market. Brian will show you these numbers in detail. So you'll be able to see that the decline that we see for the year-to-date figures is bigger than the last quarter, and we've seen that. So basically, we are approaching a steady state. And hopefully, that will take us into the positive territory of renewed growth as we saw a couple of years ago. Our EBITDA declined by 18% when you adjust for other operational income and expenses. So those we will get back to, but that's mainly related to a capacity adjustment we've taken in this quarter, and where we've made an accrual to offset that. But overall, these weaker results are mainly a result of a lower operational efficiency than what we're normally operating at. So that's simply the build-out rate of all our consultants is a little bit lower, and I'll show you those numbers in details. On the other hand, our contribution margin ended at 23%, which is approximately the same level actually as last year. And that's again when you adjust for this extraordinary redundancy provision of DKK 11 million. So this tells you something about that the contracts that we engage in and we operate and that we deliver on. We have maintained our contract efficiency, the quality. We have no additional free work or guarantee work. So that is not the issue for us. We need to aim for a higher top line. That's the main story we're telling. Positive note, the M3 business division, our second largest division, where we have developed a lot over the past two years. We continue that positive journey, 8% growth in the quarter. We can see a pretty stark pickup in the contribution margin. So we're not in an investment phase. We're actually starting to see them, have a positive contribution. They're nowhere near where they will end up being. I think they should be compared to where Dynamics is when that's running high. It's a little bit smaller, but it's still large enough to operate broadly at the same metrics, and we will follow them through next year, see efficiency go even a little bit higher and the contribution margin going up further. I'm having a little bit of an issue here, but I hope -- all right. So a few operational highlights. As I said, we're still seeing this caution in the market, although I will say that it's starting to ease up a bit. We simply have more to bid for than we have had for a long time, which means that we bid for very large projects for very large customers. So the sell-in process has always been long. But in these years, particularly '25, but also '24, they've been longer than ever, more than twice as long as what we're used to seeing. So currently, as you can see in my last point there, November has the largest number of RFPs, proposals that are reaching decision points. There's no guarantee for winning, but there's something to play for. And we're very pleased with that. It is the first sign of recovery, it's basically that customers make decisions and then it's up to us to win. And as you see again in the top point there, even in this difficult time where we've had less to bid for, we have been able to maintain our win rate when we bid for new projects. So we track that very meticulously. So how many things do we bid for, and what is the win rate we have. We don't publicize that, but we know exactly what it is, and we track it closely. We'd like to say that it's -- some people think that, that should be as high as possible. That's not the case. There are different philosophies, but if you basically win 100% of what you bid for, you should bid for more. But on the other hand, if you only have 10%, it's definitely too low. And then I'm not going to say much more about my philosophy about where I want to be in terms of developing the business. Customer closeness, there are many of our competitors that try new customer segments, new industry verticals because of this lack of demand that we're seeing. We don't believe that -- in that. We know the customer segments. We stick to the strategy. We have our full verticals. We know the customers appreciate us. We know that they still have digital ambitions. We just need to work through this with them. So our strategy is very much to keep developing these big implementation strategies with them. Some are 6 months postponed, some even 12, but we know that they will eventually revert to their digital ambition, and we will then develop on that. On the right side, efficiency. So this shows you the problem in a nutshell. Q1, Q2 operating at 62% and 63%. We then see the last quarter of Q3 falling to 58%. There is the problem. We've shown you the actuals for October, which is the first month, obviously, of Q4. There, we've already seen 61%, so a 3-percentage-point increase, which is good, but it's nowhere near enough to deliver the improvements we want to see going into next year. And then we have an additional 2.5%, which is from this capacity adjustment exercise. We don't benefit from that yet because those 89 employees that we have to exit are still on our salary books, and we maintain -- we show them as part of our efficiency until they exit our salary books. So that they will -- depending on which country and how long the runoff period is, that starts to benefit us in December, and then we have the true effect as of January 1st. We've also made other organizational adjustments. It's important for me to say that whilst we've had to make this drawdown of 89, we are not in a situation where we have a hiring freeze or anything like that. We keep developing our organization. We've hired quite heavily into new data and AI areas. We have some geographies, as you will see in the U.K. and the U.S. that are still moving forward. So we're not sort of in a situation where we're just reducing everywhere. But the adjustment was 89 that we had to do across those harder hit units. Final point here is a temperature check on the workforce, which is quite positive for us. We started the year with this strategic review, which does trigger some questions amongst our employees. We've had to do this capacity adjustment, but also, in general, very busy consultants are happy consultants, and we -- so it has an effect on them when we work our way through a sluggish period. So we're very pleased to see we've just carried out this brilliance survey, a very deep employee motivation and engagement survey, and it gives us actually still a very high score, way above the benchmark that we're up against, both for consulting and tech consulting. So I think we've been able to, through communication, carefully explaining the strategy to our employees and yes, staying close to them, they remain highly engaged and have a big faith in what we're doing. I'll just take one of the questions that was -- came in first, that said, how do you succeed with hiring these AI resources? I think this is one of them. This spreads from our employees to their colleagues in other companies where we hire from that we -- even though we go through a little bit of a difficult period, we have a strategy. We don't stray all the time. We don't panic. We know what we're doing. We build for tomorrow, and they like that message. And then I think the other one, just so that we've answered that question fully, is the technology partners that we work with. It's important to them that we have a set of partners that are considered amongst -- because those are -- they are the developers, the core platform developers, Microsoft, Infor, but also a range of other smaller ones that are seen as the ones that are front runners, so that if you work with Columbus, you get to work on the platforms that are platforms of tomorrow. And finally, who are the customers that we serve. We have the big ones in interesting verticals that have an interest in investing in this because we have, I think, the most interesting. But I would say that, of course, because we mainly focus on these 4 sectors that have a lot of physical operations that you can optimize on. Moving on to the next one, talking a little bit about this AI life cycle. So let me try to bring this slide alive to you. I would like to start on the far right side and actually read it right to left. But as you can see, we have the typical hype cycle. On the right side, we have those areas of AI that have already transitioned through the hype cycle, left to right. You will all recognize the AI personal boost, whether you use -- probably most of you will use ChatGPT, but it could also be other tools, Copilot, whatever. All of us benefiting already. I don't think there's anybody left that doesn't recognize the boost this gives you in terms of capacity. Horizon2, going one more to the left is the professional boost. The main difference here is that you give the LLM access to not just web-based resources, but the company's internal repositories. It's a huge change. We have made it. Many others have made it, but by far, not all companies have worked their way through that. It actually is quite complicated in terms of governance and security. But that is just about to happen. It's happening as we speak. It's a no-brainer, nothing in particular to say about that. But then we come to Horizon3 and 4, which are very different in nature. And here, basically 3 and 4 are about different true functional boost and eventually agentic boost, where you -- in the H3, you have sort of semi-autonomous processes. You have very complex models where you build systems together. It may not do the full execution and decision-making yet. And in H4, you have then the full autonomous process where the agentic takes over. And that I actually think is at the peak of dissolution right now. So my message with this slide is we see from our customers right now, impatience. Two years of marketing speak has led the Board to expect more than has currently been delivered. They feel that they have made the investments financially and time-wise from their organizations. They do recognize the potential of the technology. They do recognize the 5- and 10-year vision of it, but they want to see something now. So we think now is sort of a rubber meets the road moment, where it's much more beneficial for us to present smaller use cases that we can make work within 2, 3 months, and then we can scale from there. So that's where we're standing. And I've brought 2 sort of short examples, which I will just briefly run through with you, so we don't overrun. First one here is about managing an incentive program across a company, which is in many geographies, many different business units, has a lot of different variable payment plans for the employees. Those are linked to different financial metrics. They are governed by different local law in terms of employment law, et cetera, et cetera. It's very, very time consuming to manage these plans. So the solution has been basically developing an agent in Copilot Studio, drawing on this data, continuously refining it and making it more and more automated. Eventually, perhaps it can even become fully autonomous. What's the saving? Huge amount of manpower from the salary and other IT and finance departments that have to do this analysis. Learnings, first of all, overwhelmingly positive. It was possible. It creates this big benefit. Second one, perhaps not so positive. We talk a lot about democratization of agentic, basically saying any employee could develop an agent that could take over its work. Yes, I can say for sure, we're not there yet. It's fairly complex to develop this. So it requires -- Brian, no offense, but you will not be doing this yourself anytime soon. So some interesting learning there. Third one, hidden cost, be careful. The agent communicates a lot. It draws a lot in the system. It triggers Azure consumption or whatever, AWS consumption or whatever. So it's data usage, and you're going to pay for that. So if you set it up incorrectly, it's going to generate a lot of traffic, which we'll be paying for, and that will be offset against your use case. And then the final one, governance and validation. We have seen that if you set this up correctly, it will work very efficiently, providing the wrong answers. So there's a great need for validation and security governance when you work on that. My final one, and then it's over to Brian, a different example, very simple to understand. A company -- customer has -- still has a significant order inflow through e-mails, not always consistent format, which have been manually [indiscernible] into systems, creating the order and fairly simple to just create a power automation of this workflow, getting it in there. It's not autonomous yet, could be. The fault rate is actually lower because there's no human tipping interface, but we still have a human in control for the time being. And then the next step could then be to do it. And as you'll see here also, we have -- the third point here on the right is that we set it up as multiple agents. And that sort of in terms of adjusting and revalidating the system has some advantages so that we don't have a too complex setup. So we are developing this on a daily basis for many customers and working our way up in complexity and then in scalability. So I hope this gives a little bit of an example of some of these more mundane. I know that these don't knock your socks off. It's not curing cancer or anything, but this is where a lot of hours can be saved or throughput can be maximized by -- I'm not going to say 10x or anything like that. But you just generate a heck of a more throughput because this can work 24/7, doesn't take weekends and all of these things. So it has some clear benefits. Over to you, Brian.

Good. Thank you, Soren. Then I will just spend 5, 7 minutes to run through the financials as long as I'm still here and not an agent at the end of the day. Q3 business line revenue, as you can see from the page here, Dynamics, again, I tend to say, showing a decline compared to the same quarter last year. And I think it's important here to stress and that was also one of the questions. It is primarily due to hesitation and not that we are losing projects or losing pipeline, but it's simply -- it's dragging and it's just taking longer time. And that is what we're looking into at the moment. And Dynamics is also the business line where we had the biggest adjustment of the 89 FTEs that we unfortunately had to to lay off. Secondly, I'll just comment on this page, M3, as Soren also mentioned, is back in a growth journey. I see strong uptake, especially in Sweden and U.S., mentioned some countries where they see a good uptake in new projects, where they are extremely strong within bigger production corporations, where the system M3 that is not known by that many is extremely user-friendly and well structured. If we move to the bottom line, or how we measure the profitability of our business lines, what we call it contribution or contribution margin. Dynamics is down with 8 percentage points, and this is primarily due to -- or a big chunk of that is linked to the dismissal and the provision for the people that we had to lay off. Those would be around 25% in the quarter, still below, and it's not on the level where we normally see it and would like it. On the other hand, M3 is moving upwards. And of course, when you have a growing turnover, if you have a better efficiency and usage of your resources, and that does very fast to be able to see that on the bottom line. Digital Commerce, still on a too low level, if you look at the percentage, but at least it's stabilizing. And we are, as we have mentioned previous quarters, still turning this business lines towards the level where we would like it. Data and AI, we continue to invest in this business line, and we do see a strong uptake. But that, of course, as it's a minor business line, it impacts the bottom line if you have too many new people coming in that need to be trained and upskilled to be able to deliver for the future. Market unit revenue in Q3. As Soren mentioned, Sweden is starting to level out. It's not in the green yet, but roughly a flat development compared to the same quarter last year. Our two headaches, where we see the headwind is still Denmark and Norway as previous quarters, that is something that we are working very hard on and also have some light at the end of the tunnel that I feel pretty comfortable about that, but not there yet. And then finally, maybe some of you would notice that U.K. for the first time in 4 years has had a negative quarter, 7% adjusted for FX, which we don't use, but the U.K. or the pound did have a bit. It's around 4% down in the quarter, not something that we expect will continue at all, but they also see some shift in major projects that sometimes can influence the quarterly revenue. Year-to-date, revenue on business line, Dynamics down with 7%, a bit less than in the quarter. So they are trending a bit downwards at the moment. But still if our biggest business line is, let's say, feeling -- having some headwind, it does impact the full company. Dynamics is still around 60% of our combined revenue. M3, starting to be in a plus again. They started the year with a negative growth, but start to show good signs, as you saw for Q3. So a flat development year-to-date. Digital Commerce, also flat plus/minus a bit. And then our Data and AI is still on a growth path, although on a minor scale, but we are definitely investing in this business line as we see strong discussion and strong pipeline within the Data and AI area. Contribution margin on a year-to-date basis, Dynamics down 2%, doesn't seem like much of 2 percentage points. But it is a big numbers that need to be managed as a business line. Good to see that M3 is starting to get above the 20% year-to-date, 22% where we would like them to be. We always said they should be on the level of Dynamics, and I feel comfortable they will slowly get there. Digital Commerce, we mentioned it before on the right journey back on the bottom line as well and Data and AI affect the investments. One business line or area that we normally don't talk that much about is our other local business, which basically is our AEM business, Enterprise Information Management, which is a business line that is not that big in revenue, but has a very strong profitability and also growing globally. A big part of their bottom line is linked to revenue from product fee. And I think it's worthwhile mentioning that we have a strong will, and that is said with a smile, because they are basically bigger in bottom line definitely than Data and AI, almost likely to converge when you look at the bottom line. Finally, Market Units year-to-date, Sweden, still down, but trending towards [ flat. ] We expect them to start moving upwards again. It is our biggest market. There have also been some serious headwind over there in the past year. Denmark, Norway, tough year, and that's also where we have adjusted some of our employees, especially in these two countries. And U.K. still going strong with a 4% growth year-to-date. And we do expect the full year also to be in positive. U.S., although it's small in revenue at least, it's -- they're growing again after some years with a decline, and we actually see a strong uptake over there. It's a huge country, but we are very selective in the projects and the customer we have, and I think we see a good traction there. Good. That's all for the numbers. Then I will just take the last slide before questions is the outlook. And we maintain our full year outlook as we have announced on the 16th of July, we adjusted a little bit. That means that the revenue will be in the level of last year, roughly DKK 1.7 billion. Last year, it was DKK 1.659 billion. So that is around the level we expect we will end this year as well. And then secondly, our EBITDA margin should land within the range of 7% to 9%. Good.

Let's come into some questions. [ Soren ] you have already answered, why you can attract AI people. I think, Brian, you also alluded to you haven't lost projects, it's postponement. So I think you answered that already. Then a question, not putting a number on it, but how big is the pipeline compared to previous years? I don't know whether you want to go into those details. I don't number we are after...

But I'll just say a few words about why we've brought this up because we had some serious discussions, Brian and me, about whether we should say this because it is forward-looking clearly to talk about pipeline. In Columbus, we always maintain a pipeline, as you can imagine, supported by a CRM system. We track it. It has hundreds and hundreds of opportunities that we're working on. When they have a certain size, it has an interest -- direct interest for me also in terms of being part of presentations. And then it is if this deal bought, as we call it, which we see has something extra in November. It's simply -- the number that reached the level where I'm interested and then the concentration around November, which we find interested. I cannot say we don't disclose sort of how much our pipeline is because we will have to prepare against that. The interesting thing here is, of course, for you to ask, are you absolutely sure that they will be decided in November because that's the first thing because otherwise, you can't win or lose. And I can say that we can never guarantee it, but some have -- one has already been decided. Some -- we are quite sure that we will see some decisions. If they slip, it's going to be into sort of early December. Some simply for technical reasons that we can't go into, have to be decided now. And then obviously, do we win because we're not always alone, and that will be interesting to follow. And we will look for ways to keep you as updated as possible as this going forward. The third thing that you should also be aware of is that when we win, it's not a binary thing where we can then switch on the electricity on the next day. It's a question of -- we roll on an initial team, which is typically a little bit smaller than the full production team or construct team that will enter later. So there's a ramp-up for the very large projects, which we need to go through or some of them we are going through -- yes, I think that's the closest we can get.

Yes. And then actually, capacity adjustments, potential high win rate, we are not taking it into the books. How are you balancing this, if you understand what I mean? And is that because you have a longer ramp-up that you are comfortable even if I should achieve some wins and my efficiency are down here that I can absorb a high win rate if that should happen. I'm not trying to give you any...

Yes. So obviously, we need to balance our belief in the future, but also -- I mean, as you saw, we had 58% in Q3, that's unacceptable. Even with belief, we need to take action on that. On the other hand, if we have some belief, we're not going to -- there is a question here, which we might as well include what's a good level, right? And Brian and I like to be operating somewhere between 65% and 70%, then we're really having fun. And you could cut to that depth, but then we wouldn't have any growth potential left. And essentially, then we wouldn't have any belief improvement, right? So as you could see here, we see the month of October going up to 61%. We have 2.5% from this already conducted capacity adjustment. That takes us somewhere. And then we're going to have to see an uptick in the market. So that's the balance we're trying to strike.