Experian plc (EXPN) Earnings Call Transcript & Summary

Well, good afternoon, everyone. Thank you for joining us. For those of you who haven't met, my name is Annelies Vermeulen. I'm the Head of Business Services Equity Research for Morgan Stanley in London. I'm delighted today to be joined by Brian Cassin, CEO of Experian. Brian has been CEO since 2014, so over a decade, was CFO prior to that. Hopefully, Experian needs no introduction, but in one line, this is a global data and technology company that uses data, analytics, software to help businesses manage their credit risk, prevent fraud and enable individuals to access credit. And over the last decade, Brian has been instrumental in Experian's transformation from a credit bureau to a truly data-enabled technology business, an area that we're going to delve into today. So Brian, welcome.

Great. Thanks for having me. It's nice to be here.

So let's start straight in with the strategy. I think most people, the majority of this audience, likely included, would think of Experian predominantly as a credit bureau. So could you talk a little bit about how the company has evolved over the last decade, your strategy to converge data analytics, decisioning, fraud, marketing into integrated platforms like Ascend and moving beyond those traditional point solutions?

Sure. Okay. So the history of the business, obviously, I think people think about bureau think about data. Data is at the heart of what we do. Some of the largest proprietary data sets in the world, hundreds of millions of businesses, well over 1 billion consumers worldwide, some of the most sensitive data on those entities. So it's a very, very distinct and very powerful proprietary non-replicable data set. So that's the first point. Of course, data is used for something. People needed to either make decisions, do analytics, create business processes around it. So we've always viewed our strategy has been to build the best data sets, but also to leverage those data sets to do more for our customers with that data. That started off in simple things like developing scores, analytics. It morphed into decisioning systems for credit originations, for account management collections and things like that. And you just sort of gradually expand into a whole ecosphere of products and services that actually need data at the core, but do something value-added on top of that data to deliver a function to our customer. And that continues to be the way we expand our addressable market. You referenced one of our products called Ascend, which is, for once, of old money language, big data platform. We introduced it quite some time ago. It's essentially a Workbench product, which allows people to do modeling and analytics. And as we were sort of developing that module, we realized that the customers were coming back and sort of saying, can you use it for this? So we were co-creating with them. And they were actually helping us identify where else this could have applicability, and what other revenue opportunities there are. And that's how we've continued to expand the business. Today, we are a business which has got 20-odd thousand people, 11,000 technologists, 7,000 software engineers. We're actually one of the biggest technology companies in our space. And that's -- we've used that very aggressively for more than a decade to expand the business.

And when you think about those new market opportunities, whether it's insurance, marketplaces, fraud business, reg tech, beyond the traditional credit bureaus, what are you most excited about?

Well, I think really, it depends which vertical you're looking at. We cover -- I should have said we cover a number of different verticals. We cover financial services, obviously the biggest one. We also have a very big presence in health. We're in marketing services. We're in automotive and so on, quite a few others. So when we look at the broad opportunity that we have, we see specific growth areas in each of those verticals. If we take financial services, I think the Ascend platform is -- has been a fantastic growth opportunity -- growth driver for us, and actually gives us a really strong strategic footprint as we think about how the industry evolves going forward and all the capabilities we can continue to add on top of that. So that's one. I think if we flip over to our consumer business, we have a huge platform there, 200-odd million consumers worldwide. We only really started to scratch the surface of the products and services that we can provide to those, and move that business more towards a sort of financial copilot for consumers. Those are very exciting opportunities. We flip across the health. It's an amazing business, amazing opportunity. Also with a technology landscape that's way far behind most of the industries that we operate in. Their use of data is from the stone age. We have so many opportunities there. That's really exciting. So those are just some examples.

And putting all that together, how do you see your competitive position evolving relative to the other traditional credit bureaus?

Well, so I think we're always put in that bracket for obvious reasons. We are substantially larger than our other 2 players. I think we are regarded as the most tech forward, most innovative. We are the ones that actually sort of break out new concepts. Obviously, our competitors wouldn't admit that, but they sort of do in private. And I think we have moved so far beyond just that sort of certainly transactional bureau reference point, and we are much more of a broad-based technology data and analytics company, continue to push that way. I think we see the competitive set being, frankly, much bigger than that as we look across the landscape.

Very clear. We've made it 7 minutes, so it's probably time to talk about AI. So you've always been very clear that AI, poses a very minimal disintermediation risk for the business due to the confidential nature of the data that you hold. So how would you respond to those who are concerned about the potential competitive threat from AI? And specifically, can you see any scenario where AI could allow for better scores to be generated with less data?

No, simply. I mean there is not -- in my lifetime that I don't think that's going to happen. And the simple fact is that you need access to data. AI needs data. You can have -- in fact, we -- if you think back 10 years or so ago, we've always had this question about can the data sets get disrupted? The first one, some people in the audience were asking me these questions 10 years ago. Social media data, Facebook, Google, surely, they have more data than you, they can do. That's -- people simply don't understand how credit scoring is done. You cannot do it without bureau data. Why? Because we have 7 years of history of every financial interaction a person has ever done, every credit card, every loan, every delinquency, every bankruptcy, right? And only we have that data. That is not freely available because it's obviously sensitive private information. Even that's no good to you when somebody shows up looking for a loan today because they might have defaulted in the last 24 hours. So it's both the history and the freshness. We have over 15,000 institutions that contribute data to us. We have to compile all that. I think everybody thinks this data just comes into us in a nicely sort of formatted way. Not true. We have to resolve identities. We have to match all of that data. We have to ensure it's accurate. It's hugely complex. It's highly regulated. It's an incredibly large network of contributed data, and it's not replicable. So the answer is no. And by the way, I think scores have used AI for quite some time. But this is math. It's not language, right? So the algorithms you always use the most advanced mathematical techniques available. That continues, and we continue to experiment with that. But it's not a large language model application. That's actually no good to you in a credit score.

Yes, exactly. Good. I'm glad we've cleared that up because there continues to be some, I think, misinformation around that. So perhaps moving from -- and you referenced there the complexity of the data. So moving from the perceived risk to the potential opportunity. You spoke a lot yesterday about the internal productivity benefits you could see from utilizing AI. So is that more on the cost savings side? Or do you think actually also improved commercial momentum, new product launches, et cetera?

I think it's both. I think every company has, in the last couple of years, talked about productivity as an opportunity using AI. I think that's going to become table stakes. I think if you're not using AI to advance productivity in your business, you're behind. So we do see that, and that's coming through. I think we gave some stat there about how our headcount is relatively static, notwithstanding consistent high organic revenue growth. So -- but we see it's twofold. We see that productivity benefit gives us flexibility to continue to invest in new products. And increasingly, we are turning to our existing product teams, scaling them up from an AI perspective and making sure that AI is an integral part of all of our product road maps. That doesn't mean that every product is going to be an AI product. I mean I'll give you an example of that, which is today, we deliver a lot of our data in API pulls straight into production environments. These things perform at an unbelievable scale, like they -- millions and millions, billions of transactions. They're all instantaneous and they have to be 100% accurate. It's actually hard to see how a different solution would, one, improve that to at lower cost, and it's difficult to see what the benefit of changing that particular product would be. But I think the opportunity comes really around all the processes that go around those functions. And we gave an example yesterday of -- take Ascend Sandbox. This is a sandbox environment that we've built for people to build their models in. So you have experimentation, you have scorecard build, that then goes to determine how people put products in market. When you put a new -- when you build a new model, before you put it into production, you have to make sure that it's regulatory compliant, it's risk compliant and it gets approved internally. The regulations for model governance depending on jurisdiction are enormous. Folders like that big, thousands of pages. So what we've done is we've taken a small language model, ingested all that, trained the AI on that data. Now we have a product that sits on the platform, that actually builds all of the regulatory cases as you are building the model, so that you can actually sort of at the end of it, simply say, prepare the documentation, it prepares the documentation. You obviously still have to have a human reviewing that because today, even still, there's still -- it's not 100% accurate. But the time saving in that is enormous. We're talking months and lots of manual efforts. So there's a really good example of how we can use those capabilities to supercharge process improvement around our products. And actually, we have several of those already in market. So it's not like we're just talking about these things. We've got real products in market now.

Yes. And on that point, when you think back to when you set your medium-term framework, this is probably before all this -- the pace of the technology has accelerated. So actually, when you think about those opportunities, is there some upside risk do you think to that margin progression, which you've already delivered well ahead of?

Yes. So well, we're not changing that today. I think when we set that medium-term framework out, we said growth objective where it was, 30 to 50 basis points improvement over the medium term, that's 3 to 5 years. If you actually look at the organic margin improvement that we've delivered in the times that we've done that, we've already exceeded the entirety of the margin guidance for the medium-term framework. So we're not changing that today. But what we are seeing is increased confidence that we can deliver against that, not just because of productivity, but because we've had dual running of costs in our P&L, because at the same time we've been transitioning a lot of our technology to the cloud. That will start to drop out. And the ability to continue to do what we've always done, which is actually invest back in new product, to drive new opportunities and growth as well as deliver really great financial metrics.

Absolutely. And then thinking about the opportunity, but also the risk because clearly, implementation of AI, we've seen this in other examples, there's a risk of cyber threats, data breaches, hacks also goes up. So how do you think about balancing the investment there between AI-driven product innovation and also security initiatives to protect data?

So one of the first things that we did, and this is not -- this is actually several years old now. The first thing we did was actually to set up our internal processes around how we use AI, how we govern it, what technologies we're going to use and how and how they connect? And so that is -- there's a lot of proprietary technology in there. And we've given tools to all of our developers, which are within that wall, and which are governed and have security as built in a standard. That was the first thing that we did because that is so crucially important to us. So we can build with confidence now. And we're 100% confident of that, that risk is covered.

Yes. Front-footed as always. So then thinking about on the customer side, you've perhaps worth talking about your virtual assistant EVA here as well, which you're rolling out further AI capabilities for. So what are you doing again to persuade, I suppose, customer concerns around how their data is used and what the AI assistant will do in that regard?

So that's a big concern for our B2B customers. But whenever we have a product, a new product, or an existing product and an improvement, because of the type of nature of our customers, we go through very extensive testing with them. They're incredibly rigorous. There's all sorts of security audits around it. So that's just part of the normal procurement process, and it fits into that. Flipping over to the consumer side, to be honest, it's sort of -- it's an interesting question because we do lots of customer surveys and we have a lot of feedback. And if you ask 2 questions of a consumer, which is the first one, are you concerned about your data privacy? They go, yes, 100% of people do. And then they go, would you give up all of your data for a free coffee in Starbucks? 100% say yes as well. So there's a little bit of schizophrenia on this. What we do know, when we started this probably more than a decade ago, there was a view that take, for example, open banking, or access to consumer permission data. Some people said consumers would never give more data to bureaus because they're worried about how they would use that in the credit scoring. I mean it's proved to be 100% inaccurate because what we found is that consumers are more than willing to actually contribute data to us as long, as they perceive some value in that exchange. And in fact, it actually goes a bit further than that. They sort of expected. And the weird thing is that we get more comments back from consumers saying, I got this thing. And that's not relevant to me. And I thought your experience, you should have known that. Why did you send that thing to me? They think we know more about them than we actually do, which is kind of interesting. So I don't think there's any issue about that. And we've been -- we've proved very successfully that, that's a huge opportunity, and it's an asset that's actually working for us. In fact, you may ask about cash flow attributes, product we introduced last week. That is a good example of how we built a B2B product using training on consumer contributed data in our consumer business.

And then continuing on the theme of huge opportunities. When you think about your Ascend platform? And could you perhaps update the audience on the rollout there, the progression? I know it's been a phenomenal story so far. And how is it creating new markets for you? And I suppose what would also be interesting within that is if you're seeing any of your competitors trying to replicate what you're doing?

So take the competition one first. We introduced the Ascend platform probably 6, 7 years ago. And every year since then, I've been asked that question. We -- for this particular product, we are the only people who have any market penetration whatsoever. So we created this category. And it's been very difficult for anybody else to actually displace us anywhere. So a large majority of Tier 1 banks in North America were the first adopters, which will tell you how powerful the solution is because these organizations have thousands and thousands of technologists themselves. We are in the middle of actually rolling that out in the U.K. I think that by next year, we will have virtually 80%, 90% of all lenders in the U.K. using the Ascend platform. So it just gives you an idea of the strategic position that gives us. And of course, as we proved with the originally just the sandbox, now much more. Ascend Ops, Ascend Marketing, Fraud Sandbox. The ability to just continue to add more capability onto that platform, more modules, it's really, really powerful. Yesterday, in the results announcement, I gave an example of a Tier 1 client who has been a long-time customer for our data. And if you look at the customer journey of that, we went from selling data to selling them point software solutions, particularly decisioning. They also use other competitors for software and internal solutions. And then we have now migrated them on to Ascend platform. And within that, we've not only replaced other software vendors, we've actually added capabilities like a product that we have called Aperture, which really helps customers ensure that their data is ready for use in the platform like this. It eliminates some of the problems with siloed data because it's got incredible matching and pinning capability. So that actually helps the data be used in that environment. We then actually now talking about model risk governance. So it just goes to show how that can really build a completely different relationship with clients.

And on that, then you obviously you're bundling all these services and then pricing for that. And as you become more integrated with those clients, your sort of replacement risk also goes down?

So replacement -- that's obviously the objective, like everybody wants to be irreplaceable in their clients' environments. And most clients try and avoid that from happening because they don't want vendor lock. So there's always a bit of a tension around that. I would say we have always been an organization that had many, many products. I use this example. We have one Tier 1 bank in the U.K. that takes 80 different Experian products. So bundling of products is not a new thing for us. But because those products legacy, many of them legacy, they're all sort of individual point solutions. And when you think about it, there's a commercial opportunity for you where you can actually give additional value. But maybe from the client's perspective, they have to integrate with 80 different products. So why is that different from integrating with 80 different vendors? You start to move towards a platform strategy where these products are actually connected to each other. You don't need 80 different integrations. You only need one. And so that's a huge operational saving. What we are seeing across the piece, particularly in areas like fraud, increasing demand from our clients for solutions that help them manage the complexity of dealing with different solutions. So orchestration layers, built on capabilities. That's where things like Ascend, CrossCore, PowerCurve, they've always been really great at that, and we see that as a long-term trend.

Very exciting. So switching now perhaps to Brazil. Again, something you've spoken for a long time about around the opportunity from positive data and so on. So what's the latest there on positive take-up of data and progress?

Yes. So I think we don't really talk about positive data separate anymore because it's sort of everywhere. We actually don't really sell negative data and positive data separately. We did at the outset, but now it's really all integrated. So what has it done? It's massively expanded the addressable market because more consumers are now credit visible. Previously, I don't know how many people in the audience understand the difference between positive and negative data. I'll just explain it. If you have a negative data market, it means that you can only see people who have defaulted, which in Brazil is a lot of people actually. When you have a positive data market, it means you can see everything, which is that your performance against payment against loans and blah, blah, blah. So it's a much richer view. It's a much bigger data set. And of course, it means that you can see more consumers. So that market has expanded quite significantly. All the scores and analytics have to change. It's driven quite a big significant growth for us since that introduction. You can see our business is now much larger. We're really still only in the foothills of that. It's sort of got a bit interrupted by the economic backdrop in Brazil because once you get a contraction in lending, yes, it's fine. You got more positive data. But if you're pulling back and you don't want to lend anyway, it doesn't matter. So we've seen that interest rates have moved up very high. We've seen that softness in the market. But ultimately, when we get a bit more of a positive economic backdrop, I think that growth rate starts to accelerate again.

So perhaps, again, for the benefit of the audience, how far along would you say Brazil is in terms of replicating what you have in the U.S.? And what else would you like to add? Clearly, you've added fraud, you've really expanded that footprint, but what's next?

So a long way to go. And I think -- if you look at our business there, we're now the #1 fraud player in Brazil. We are the #1 player in business information, the #1 player in consumer information. We have a strong decisioning business, probably #1 in analytics as well. So it's an amazing footprint. But actually, if you look at the revenue contributed from those different verticals, still much smaller than places like the U.K. and North America. So for example, the U.K., if you look at our business across data, decisioning, software platforms, analytics, scores, it's very balanced like 20/20/20 fraud as well. Brazil is still more skewed towards data. So the opportunity for those growth areas is still really significant.

Perfect. Then I just wanted to come back on the margin opportunity, which we obviously touched on when we're talking about AI. So specifically on the transition to cloud-based decisioning solutions. So how do you think that will, A, benefit your competitive position, and also drive that margin expansion?

Yes. From a competitive position, I think it's important to just clarify what we mean by transition to cloud. All of our new products are built in the cloud. There's nothing built not on cloud. All we have left, not on cloud, are really some data centers, some legacy systems, the actual repository of the ingestion procedure for some of the bureaus. So we'll move that across. In fact, we're migrating -- we'll be finished migrating all of our U.S. clients by the end of this year. In some cases, they won't actually see any difference. I mean that's the objective, right? We don't want them to see any difference in terms of performance and so on. The benefit for that will be for us because the benefit of building new products in the cloud, we're already doing that. So from a competitive position, it gives us -- eliminates dual running costs, gives us more financial flexibility, you can turn more of our resources back on to product creation. So that's going to accelerate that. In the marketplace, we've already been taking advantage of that technology to build new stuff.

Absolutely. Great. We'll come to audience questions in a moment. So if you have any burning questions, do get them ready. But before that, we need to cover U.S. mortgage as well, which again has been a big topic of discussion for the last couple of months. So perhaps, again, remind everybody, how are you responding to what FICO are doing in terms of the direct license program? And what is your strategic response to mitigate revenue and margin impact?

So I'm just going to take as read that people understand what actually happened here. But -- so I think, as we look at the U.S. mortgage, I think there's a couple of debates that have been knocking around. First of all, FICO traditionally distributed its scores through the bureaus alongside the data. There's a good reason for that because it's actually operationally the much simplest thing to do. And then you have the bureaus calculating the score, you have limited room for differentials between outcomes and so on and so forth. So they have a few weeks ago, made it available as an option that you can get the score direct from them. So what that means is the resellers, so the people who sit in the middle between bureaus and the people who write the mortgages have to now calculate the score. If they want to use that model, they don't have to. So they're going to take on the operational risk, complexity of doing that. And there's lots of things they have to do, such as actually setting up a technology environment to calculate the score, actually getting a different feed from us to do that, which requires security, compliance, blah, blah, blah. There's billing issues. So there's a lot to do. So it's not going to happen just like that. I don't know when that will be ready. I think some of the big resellers will be ready sooner than others, but not all the resellers will move, because I think people don't appreciate that. When you calculate a score in the U.S., you are liable for if that score is right or if that is wrong. We as bureaus have been doing that for a long time. No reseller has yet done that. So it will be interesting to see how that evolves once they meet all the wonderful lawyers that we have to deal with all the time. So look, I think -- so that's going to happen. But at the end of the day, it doesn't really matter to us because the key point here is the debate about where is the value? Is the value in the data, or is the value in the score? And I think here's the fact which I think proves where the value resides irrefutably. The GSEs in the U.S. underwrite 50% of all mortgages. If you look on the GSE website, they're quite explicit. When they do that, they use data from all 3 bureaus. They can't do it without data from all 3 bureaus. They do not use the FICO score. So they do that underwriting without any reference to score whatsoever. The score is attached as part of the sort of package to sell on the loan. So it's embedded in the system. But we know that the value resides in the data. So this change just means that actually our economic value from that process will be concentrated in the data and score processing fees. So we put together a package called the Experience Score bundle, which incorporates price for our data, the processing fees plus the FICO score should they choose to sort of continue to consume it from us, plus the Vantage Score alongside that for free. And we're going to be actively helping people test the Vantage Score against FICO in the market in real time. And we'll see how that evolves over time.

Exactly. And obviously, I mean, Vantage Score is something you've been -- has been in the business for a long time. But do you think -- what's your view around those who would say that FICO will remain the industry standard? And how quickly do you think actually Vantage Score could get adopted?

So there's clearly one narrative which says nothing is going to change. That narrative was several years ago was that the GSEs are never going to approve Vantage Score for approval in mortgage. So that wasn't correct. Now the narrative is it doesn't matter because people are still going to use the FICO score, maybe. But let's look at non-mortgage where FICO didn't have a monopoly. And this is before we never pushed it, right? So Vantage Score has about a 30% share of unsecured lending. About a 50% share in fintech, more than 50% in card, lower amounts of some of those other verticals. And that, to be honest, that's without really, and the bureaus trying, because we weren't really economically incentivized one way or the other because all the scores are coming through us. So that's changed. And Vantage Score is now available for free for testing, and we're going to encourage people to put it in the sandbox. The other bureaus are also doing that as well. So it's a pretty heroic assumption to assume no market share is going to shift. Then what's the feedback that we've got, feedback, people are looking at this and they're looking at it seriously. And there are some pretty big people looking at this. So where that gets to, we'll find out. I think it's true to say it will take time, but I would also say that we have been a bit surprised that some people are more ready than people realize.

Absolutely. And look, I think it's fair to say that it was a bit of a surprise when that hit the tape. So if you think about any other examples in the business where you could see a situation like that, whether it's a regulatory change, whether it's some -- change in regulation that perhaps we weren't expecting or change in policy.

Well, yes, I'm not going to sit here and predict what politicians might come up with. But what I would say is that if you actually look at a lot of the kind of surprises like that have happened over the years. So for example, when the Biden administration came in, there was a lot of talk about a public bureau, and there was a lot of people concerned about that. We were never concerned about that. We know how difficult these things are. The concept that they would actually get their act together and produce something which was useful in the marketplace was a little bit fantastic to believe in. And that's kind of gone. And I think you look at -- I think that was -- it was a surprise when it happened. But when you actually look through the dynamics of what happened, you realize that it actually really doesn't make that much difference to us longer term. And in some senses, our economics could even be improved by it, particularly if you start to get some shift of share towards Vantage Score. So always -- they're always capable, but I think the resilience and strength and position of our business always shines through.

Absolutely. And I guess it hasn't changed your view on mortgage. Your mortgage has always been lower than your competitors deliberately.

Yes. Although -- yes, that's right. Although we -- lots of things happen in the mortgage market. There has been a lot of change, and we've actually been leading quite a lot of innovation around change. Some of the stuff that scores that's changed is to do with cost to consumer. We've been very proactive in helping improve and change the process so that we reduce the cost to our clients for non-closed loans, for example. So we expect that to continue.

Yes, absolutely. And then sort of related on capital allocation, I suppose, what -- could you talk a little bit about your M&A strategy? And I think this is an interesting point here, especially also around the margins. So some of your recent acquisitions, whether it's ClearSale or NeuroID, how those integrate with your platform? What is it you're looking for with deals and how you see that progressing?

So the acquisitions sort of fit into a few categories. We like to buy bureaus wherever we can. And we'd like to consolidate bureaus wherever we can. That's the Australian acquisition did last year. Take the #2, #3 position, much stronger competitor to -- it's Equifax as the leading player in that market. That's just -- that's a pretty logical -- a lot of synergies. So that's one type. We then look at ClearSale, which is a fraud acquisition in Brazil. We've been building up our fraud business in Brazil for quite some time. ClearSale is focused on transaction fraud. That's not an area we're in. We see synergy between transaction and origination fraud. That now makes us the leading player in fraud in Brazil, so very logical add-on. Also, that comes not just with a fraud capability, but also with a quite significant data asset. And that also strengthens our data business and back into credit, and we think that we can improve our scores in there. That's another example of a good product with -- it's got a product capability plus data, which is enhancing to an existing position. And then you slip down into something like NeuroID, which is behavioral biometrics for originations fraud. It's kind of a fundamental kind of add-on feature that we need to add to our solution set. We have a product in the U.S. called Precise ID, which is one of the leading authentication products for originations fraud. And increasingly, you need to add more capabilities to that to cover things like biometric behavior at the point of application and so on and so forth. So they all fit into that kind of narrative. Same. We've just done a small acquisition in the U.K. And again, another fraud business takes us into a slightly different vertical, quite synergistic from a product and from a cost perspective. That's kind of the things we're looking for. We also do it in the other verticals as well. You've seen us buying health. And you've seen us some time ago buying marketing services to really help build out things like our identity products and so on.

And so when you think about some of those acquisitions that you've done, you typically talk about them taking 3, 4 years to become margin accretive. So could you talk a little bit of what you do with those? And are there any examples of where actually been positively surprised by either the quality of the data, or the pace of integration, or anything like that?

Yes. So I think we -- what we usually find with acquisitions is that we build a case for the acquisition, and we typically always outperform on the EBIT side. So we build a margin assumption in and the synergies tend to come probably a bit quicker than we had modeled. I think we try and sort of sell them as part of our overall solution. So NeuroID will be on the Ascend platform. It will be matched up to Precise ID, is an integral part of that. We look for the product synergies, and we look for the cost synergies. ClearSale is an example of a business that we bought. ClearSale was a public company. Normally, we're able to do extensive diligence on the data assets because we tend to buy private companies. Obviously, public company a little more difficult to do that. Actually, post acquisition, we have found that the data asset for ClearSale is even stronger than we anticipated. So that's really a pleasant surprise. So I think getting up to sort of group average margin can take that amount of time, can be quicker. It just depends on the business. Group average margins are pretty high. We tend to buy businesses that have not got the same level of scale as us, so therefore, don't have the same margins, but we think that's the opportunity where we add value.

Yes, absolutely. And when you think about timing, I'm sure there's assets that you have your eye on, but sometimes the timing or the valuation doesn't quite stack up. So when you think about capital allocation over the coming months, what for you are the priorities either in terms of -- if you can't do the M&A, is it returning cash to shareholders? Is it investing more in the business based on the opportunities that you see?

Yes. So we've got a very kind of clear capital allocation framework. We believe that we can do all of them. And if you look at where we are from a debt-to-EBITDA ratio, we're pretty low. So one is not sort of exclusive to the other. We always set out our capital allocation framework at our annual results. So we'll give an update on that then.

Perfect. We have 5 or so minutes to go. Are there any questions from the audience? Nope.

I have time for another question.

Yes, absolutely. So I wanted to also ask, I think it's a really interesting question around consumer services because, again, this is something that has also come up quite a bit. So how you position the business against sort of fintechs, challenger apps? And are you seeing improved usage metrics, specifically in consumer, for consumers using some of those Gen AI features?

Yes. So I'll unpick that a little bit. So all the metrics in our business are trending positively. That's traffic. particularly organic traffic, traffic from LLMs, that's increasing quite rapidly. So, so far, that's really proven to be quite positive. We have a very strong brand. I'd kind of characterize and say, even if we decided not to have a consumer business, we would have one anyway because we can't stop people from coming to Experian to either dispute, or ask questions, or find out about their credit reports and scores. That is millions of people every year. Pretty much everybody in the U.S. knows who Experian is and what it means to them. So we have a tremendously powerful brand and a lot of organic traffic. I think our brand resonance is right up there with any financial asset you can. In fact, in Brazil, the Serasa brand, which is the brand we operate on the Serasa, is actually only second to Itau, and maybe Nubank, in terms of customer acquisition. It outranks pretty much every other financial property. So that's very powerful. People understand what role we play in their lives, and I don't think they think that's going to stop. We have really successfully built on that, not just through a membership product, but also through building out a marketplace. And that's all about making that as convenient as possible for consumers to gain products that are relevant to them. What do you need for that? You need data. You need our data. You need the analytics, you need the scores. You need to understand how financial institutions will score them. And we actually have all the financial institution scorecards because we built many of them, and they're all on our Activate platform, our consumers. So we can match up consumer need, demand to the right product. We help the institutions actually start to recalibrate their scores to attract different audiences. It's just tremendously powerful network, and it's not just about sort of putting a few offers in front of them and telling them to get a credit card. It's deeply personal to them, and it's deeply relevant to our clients. So that's a very powerful thing, and I think it's going to continue to be very powerful in the future. I think there's a real opportunity actually as we think about the strength of our brand and how that could work maybe in partnership with some of those properties, lots of discussions going on. So we see that as a way of enhancing our distribution actually. But the most important thing here is the brand is really what drives this. And you're seeing a lot of properties that have weak brands suffer, but you're seeing properties that have strong brands actually get stronger.

And another great example, I think, is marketplace and what you're doing there. So you're adding home from auto and insurance and so on. So what's the progress there? I think that's been...

Really good. I think you've been covering us for a while, and you've actually seen some of these things from inception through to now getting to scale, particularly insurance, I think, started that a few years ago, now pretty big. Insurance is sort of a bit lumpier than we thought because the insurance, they tend to sort of turn on and turn off at different times of the year. But if you look at the trajectory over a multiyear period, you can see it's very strong growth. We still have loads of verticals that we're not in, or not in a very big way. A great example is mortgage. We will be launching a mortgage proposition on the platform pretty soon. We're very excited about that. We are in a brilliant position to bring relevant mortgage offers to customers. By the way, this is a perfect Experian opportunity because mortgages are really messy and complicated. It's not like credit cards where you can have 5 different -- you don't have 5 different mortgage offers. So complexity of process, deep information on the consumer, also information on the asset, which we have, housing information, we're able to actually create these propositions that are going to be, I think, really a really great experience for consumer and a really great experience for the FI customers.

Absolutely. So we've covered an awful lot of ground today across all the different businesses. What for you and for the audience, do you think is the one key takeaway that you'd like people to take away from about Experian today?

Well, I think the strength of the business, the track record of growth that we have, I think the number of areas that we believe that are white space to us that we can continue to expand to defensibility of our data assets, actually, it can't be replicated, the ability and strength of strategic position that gives us. And I think the opportunity that we see from advances in technology, particularly AI, in opening up more of those opportunities to us. We see all of those as really white space for us and actually more accessible now than ever.

Yes. Perfect. A very strong note to end on. So thank you very much, Brian.

Thank you. Thank you.