Ipsos SA (IPS) Earnings Call Transcript & Summary

Good evening, everyone, and welcome to Ipsos 2023 First Quarter Results. This conference call will be hosted by Mr. Ben Page, CEO of the Group; and Mr. Dan Levy, CFO of the Group. As a reminder, this call being recorded. [Operator Instructions] I would now like to turn the call over to Mr. Ben Page. Please go ahead, sir.

Thank you very much, Ben. So I'm Ben Page, the CEO, and it's great to be here to talk you through our Q1 results. I think before we get into the detail, I think it's just very important to remember what I said back on the 15th of February when we released the full year 2022 results, and that is that the path of revenue recognition during 2023 at Ipsos will be completely different to 2022, and that is, of course, what we are now seeing. So let's just look at the numbers. We have positive momentum but negative revenue at this point. So first of all, we have EUR 532 million of revenue. That is the second highest figure that Ipsos has ever achieved. But as my colleague, Dan, our CFO, will explain, we are facing a cliff-edge comparator. And as a result of that, the revenue is down 2.8% because of this cliff-edge effect of the major COVID contracts, which were present in 2022, but of course, ended at the end of March 2022. So there's a very, very tough comparison in this quarter, which then gets much easier for the rest of the year. If you put aside the COVID contracts, the organic growth net of those is 0.6. But the order book, which is basically the revenue, which is visible and confirmed, but not yet booked in the system, and Dan can explain the technicalities on that, but it's visible for this year, net of the COVID contracts is currently up 3.3% and has been building during this quarter. So March was up 6.4%. So we saw a deceleration at the end of last year with uncertainty. We are seeing an acceleration during Q1. But it is a very volatile situation out there. I thought the fact that the economist talk about the Monalisa economy or however you look at it, it looks different. It was quite interesting. We'll come back to that during this presentation. But overall, we have good momentum. And I'd like Dan now to take us through just some of the underlying factors with the adverse effects on the comparison and some of the other things that are going on. Dan?

Thank you very much, Ben. So obviously, as Ben just explained, the decline in Q1 revenue is, first of all, due to very strong adverse base effects, which were actually expected as we announced in February for our full year results. The first adverse effect is obviously the fact that we had a very strong Q1 last year with an organic growth above 12%. We had more than 20% growth organic in the Americas in Q1. And we had double-digit growth in China before the lockdown that hit China in 2022. And obviously, this is a huge base effect, which hits now our revenue in Q1. And the second base effect is actually the cliff-edge effect of the large COVID contract which, as you will remember, ended at the end of Q1 last year, which were in the revenue of Q1 2022, and which are not anymore in the revenue of Q1 2023. If you strip out the effect of this COVID contract, organic growth would be in Q1 of plus 0.6%. In addition to these 2 base effects, which were expected, we are facing in Q1 2 other headwinds. The first one is a kind of wait-and-see attitude from our major tech clients in the U.S. These clients are reorganizing, which means that some of the clients Ipsos of the contact that Ipsos has with them has changed. They are changing their strategy. And all of this means that some of the projects, which they forecast were actually delayed. And as you see on this chart, the revenue on the big tech clients in Q1 is down by EUR 7 million. Of course, there are a lot of needs that remain from these clients because we do a lot of things with them. We do trackers. We do mystery shopping. We do corporate reputation, and we are going to keep on doing that. And also, we have new leads from them, particularly on AI because we are currently working with some of the Big Tech clients on their private version of generative AI models. So there are a lot of opportunity for growth going forward. But currently, in the Q1, we have this cliff-edge linked to the reorganization. The second headwind we have seen in Q1 is actually linked to the rebound in China. In China, the rebound doctor of activity is clearly seen in our numbers in the order book, in the order book, as you can see, the other book, as you can see is growing organically since January by more than 13%, but it is not seen yet in the revenue. The revenue are down by 3.9% because the Q1 revenues are still impacted by the lockdowns. It is important to see that when you have a recovery in the way we recognize the revenue, it -- the recovery is always seen first in the order book and then is transformed into revenue because the recognition of revenue at Ipsos is basically from the start of the work till the end of the project. So there is always a lag between order book and revenue when there is a rebound in activity. And this is actually exactly what we see this quarter in China with revenue down and a very strong order book. And maybe more generally speaking, I think it's important to note that the Q1 at Ipsos is the smallest quarter in revenue usually as opposed to the other quarters. And as a consequence, Q1 revenue is not a very -- is not very predictive of the full year performance, whereas order book is a far better indicator because usually, at the end of March, we have more than half of the order book of the year, which is already booked. Despite the adverse base effect that we described before, as Ben said just before, the first quarter of 2023 is the second best ever performance. As you can see on this chart, we have grown by 9% in Q1 as compared to 2021. So to an extent, if you do the bridge above the cliff-edge effect of 2022, and we have grown by 25% organically as compared to 2019. I think it is important to show that because it shows the resilience of Ipsos business model, and it is always important to look at medium and long-term performance rather than to focus only on the quarter-by-quarter results, which can be volatile as we see in our Q1 results. If we now turn to the breakdown -- revenue breakdown by region. Obviously, we see a decline in EMEA by 6% organically, which is basically linked partly to the war in Ukraine, but more importantly, by the end of the COVID contract. If you strip out the effect of this covered contract, we will have an organic growth by -- of more than 1% in EMEA. Organic growth in America was overall 1% with quite different situations between Latin America, which grew organically by 9% in Q1 and North America, which was overall stable, impacted, as we said before, by the reorganization in some tech big tech clients. But again, this kind of stability in revenue in North America has to be put in perspective with a very strong Q1 we had last year because globally, Americas grew by 22%, as you can see on the slide, in Q1 2022. And it's important to note as well that the order book in Q1 in the U.S. is positive with a 3.3% organic growth. Asia Pacific, we have a very good momentum in India, growing by more than 20%. But obviously, Asia Pacific suffered from the situation in China at the end of the zero-COVID policy. China is the biggest country in Asia Pacific. But I had explained before, we are seeing a recovery in China. In terms of audience now, our performance in the clients and employees and consumers research activities remained broadly stable in Q1, again, after a very strong performance in Q1 last year. You can see on this chart that it was respectively 20% organic growth and 7% organic growth. Again, the reorganization in Big Tech clients weighted on the performance, but we have, on the other hand, very good performance on our brand health tracking activities, consumer experience activities and data analytics and Ipsos.Digital as well is doing very well, and Ben will come to that in a few minutes. Citizens is decreasing by 30%. That's also expected because of the end of the last COVID contract. If you take out the last COVID contract, then the public sector would grow organically by 8%, which I think showed a gain the need for the government in a multi-crisis world to inform public policy decisions with reliable data. And finally, our doctors and patient activity is down by 5%, again, with a strong comparison to last year, which was 11% and also due to delays in decision-making in some pharmaceutical clients. And now over to Ben for some highlights on the business and for the outlook for the year.

Thanks, Dan. So the United States, as you've heard, the order book is up 3.3% over the quarter. And just to give you a bit more detail on that because the United States is our largest market, and indeed, it's the largest market for the research industry generally. So the tech sector is very mixed. So you've got retrenchment that is well publicized in several of the largest players, and that's causing people to reappraise programs, et cetera. At the same time, it's worth saying that one of the largest tech companies in the world has massively increased its spending with us. And I think it's actually close to our annual budget for that client. So it really does vary. But net-net, spending is down in Q1, we can see visibility on Big Tech company spending. We have several contracts that have been confirmed yet to be signed coming through in the next quarter. So we're expecting an acceleration there. We've also, of course, got double-digit growth in some of our offerings that we're prioritizing. So things like market mix modeling, that's up 25% in our marketing management and analytics business. Our work on ad testing in the United States, again, driven heavily by Ipsos.Digital, where we've automated a lot of that work. That's up 13%. And Ipsos.Digital itself in the United States is up 35%. And finally, the other thing that we've seen, which also explains this slower growth this quarter is the tension in the American government between the Senate and Congress, which is making government federal spending slow down and decision-making slowdown. So again, these contracts, we have some very major contracts that are definitely confirmed but are being commissioned on a quarterly rather than annual basis because of potential cuts in government or freezes in government spending while the Republicans and the Democrats agree budget. So that has slowed down the confirmation of contracts, but we have no concern about those in the long -- in the medium term for the rest of the year. So that's the United States catching up after delays. China, as we've discussed, is up 13.5% in terms of its order book as it bounces back, although a lot of that has come through in the last month and therefore, isn't recognized in terms of revenue. And that's important. We can see growth across our business in China, both in terms of food and beverage, in FMCG, in telecoms and also interestingly in automotive. You all know, you may have read about the very strong competition in the EV market in China, which again is driving demand. We've also interestingly got rising demand from Chinese companies wanting to go global. And although the depth of globalization is often talked about in the media, actually, China is fully integrated into the global economy in most aspects. And those businesses still want to grow outside China, and we can help them. And finally, Ipsos.Digital itself, our automated solution now available in Mandarin and again, growing rapidly. And so that's the situation in China. So again, we are expecting the recovery, perhaps coming a little bit slower in revenue terms than we might hope but certainly there very visibly in terms of confirmed orders, which will be booked as revenue in 2023. Ipsos.Digital itself, we are aiming, as we've said, for a revenue of EUR 100 million at the end of -- by the end of this year. Currently, the revenue is up 36% and is again, like the rest of our business has accelerated during Q1. So things that are helping drive that growth, multi-language availability now for Switzerland and Belgium, very important. The platform being live now in Mandarin, but also coming live in Japan in April. And after that, we will move to Spanish for the rest of Latin America. And then alongside expanded geographies that we're able to offer the solution in. we're also, of course, adding new products all the time. The next 2 that plan to go live in Q2, pack testing, very important for our innovation business, and that will go live in Q2 and creative early-stage testing and that will also go live in Q2. And as one of our biggest health care clients puts a great option for us. We can write around survey. We're also supported by the account team when needed. The language and the results are understood by marketing. We get benchmarks that we need, the reporting is easy just to click of a button. So again, we're comfortable with the development of Ipsos.Digital. It's an important part of what we do, and we will continue and have a clear road map for the next few years of putting more and more of our products onto that platform, again, offering us faster results, more client control and generally something that we enjoy working with our clients on. So overall, we can see in our order book a year that is in line with our guidance. And as Dan has already commented, the order book at this point in the year is a better predictor of outcome looking back historically than is our revenue in the quarter. So on average, for the last 5 years, we've had 52.5% of our full year we've actually achieved revenue visible in the order book at the end of Q1. This year, it's 53% based on organic growth of 5%. And so because of that, I think all other things being equal, this is what we are seeing, and that means that it sustains our 5% growth guidance at this point of the year. Having said that, it's a volatile world out there. The work that we do, looking at consumer confidence, both for the world's largest banks but also for media organizations around the world, you can see here, you're just looking at the global aggregate data from our monthly tracker. You can see the COVID effect and the V-shaped recovery. But then, of course -- and I think this is the series of things that have surprised central bankers, the war in Ukraine, nobody much could have predicted that maybe with the knock-on energy effects, all of that, of course, amounting to a cost-of-living crisis in many countries. You can see Latin America recovering, and that's partly explaining our very rapid growth, continued rapid growth in Latin America, but North America feeling less confident than it did before the pandemic, Europe, Asia Pacific, still not quite recovered from before the pandemic. And this general volatility, which you see in the stock markets with reactions to things like the Silicon Valley Bank, I think it just means that it is still uncertain world. As I've said many times in these meetings, uncertainty is the only certainty in many ways. But having said all that, in terms of our order book and what we can see booked, the trajectory that we can see taking place in our major countries, we're confirming at this stage, organic growth of 5% and an operating margin of 13%, around 30%. So I know you will have some questions, which we'll come to. We have an AGM on the 15th of May, our Annual General Meeting. There will be some more details there. And on the 14th of June, Dan and I and some of my new colleagues who you have yet to meet, we'll take you through how we are going on our long-term plan for 2025 and some of the changes that are going on inside the company to try and improve productivity, make things faster, et cetera, et cetera. But with that, Dan and I are very happy to take questions. Thank you.

[Operator Instructions] The first question comes from the line of Emmanuel Matot, calling from ODDO.

3 questions from me. First, if I remember well, Ben, you are expecting in February, a very low organic sales growth in Q1, but not a negative figure. So why Q1 was below your initial expectations and which sectors should recover the fastest in the coming quarters? Second, in case the organic sales growth will not recover as expected, would you be ready to take quick action to protect the EBITDA margin? And my last question is about the deal between NielsenIQ and GfK. The tie-up is likely to face a full-scale EU antitrust investigation with regulator demanding asset sales in return for clearing the deal. Could that be an opportunity for you to buy some of the assets in retail and consumer measurement? Or it is not strategic businesses at all for Ipsos?

Okay. Interesting questions. I think it's fair to say that, as you know, we were expecting revenue to be lower than the previous year, but not to be -- not to see a negative effect in that way. And I think the part of the reason is that when I think about the business, I tend to think about the AOT, as we call it, the order book, which is a much -- for the first part of the year really is the way in which to understand how the year is turning out because basically, the revenue, as I say, is not a very strong predictor. So nevertheless, yes, it is a surprise. And I think part of that is this combined effect, both of the slowdown in the COVID contract but also then the knock-on effects of the restructuring in the tech sector. We do have some very major orders in the tech sector that will come through in the next quarter. If we don't see the growth that we need in terms of revenue as the year pans out, we will be taking tough action on cost and payroll. We are actually employing fewer people at this point in the year already than previous years because by managing attrition, et cetera. But absolutely, and you saw our ability to do that in 2020. And absolutely, we will manage our costs to try and protect our margins as much as we can without damaging the business. In terms of GfK and Nielsen, it depends what they have to sell. We were very satisfied with our purchase of part of GfK back in, I think 2018, so never say never. Let's see what they're talking about. We have very little debt. And obviously, we are interested in acquisitions that fit with our longer-term strategy.

The next question comes from Conor O'Shea, calling from Kepler Cheuvreux.

So a couple of questions from my side as well. Just the first question just on the growth, understanding completely the multiyear comp in Q1 was on a different level to other quarters. But just wondering how we should look at the negative growth. I think you've said in the past Ben and Dan that 2/3 of your contracts are index linked to inflation or short-term contracts with current pricing. So are we seeing in Q1 at least price increases on most of your contracts, but much more volume? That's the first question. And second question, just on the TMT, the tech sector comments, which makes sense. But just wondering if how that ties in with EMEA being the weakest region, my understanding was most of your tech clients are in the U.S. I know there's COVID contracts, which are mostly in EMEA, but wondering if there are any other factors explaining that? And then just in terms of your activities, can you just explain -- I might have missed this at the start of the call, what's going on in your kind of health care unit down 5%. I think we've heard from some other areas in this group that growth has come off quite a lot post pandemic. So just wondering what you're seeing there.

Well, let me -- Dan will help me with anything -- any more details, but let's take those from the top. So we're continuing to put through price increases. In a sense, I mean, we just -- I just been -- it's funny, just, I was just looking at average price increases being put in by different markets. And that's -- we're not publishing that, but we are -- basically, we are putting through price. We continue to put through price increases. And so that isn't -- that in itself hasn't become a worry. I think overall volume is the thing that we need to look for. The TMT work in terms of Europe, actually, there is a lot of work for the tech players. So yes, yes, there's a lot -- it's a very large component of the United States. There's actually also a major component in across EMEA, particularly -- remember, the U.K. is our second largest market globally on it so. The other thing that's going on in, EMEA, is that we are in a rebound situation in Germany and it's obviously one of the largest markets in Europe. We've recently appointed a new country manager, Dr. Christoph Bruce. He is in the business -- in the process of restructuring, reorganizing and refocusing that business. And that, I think, explains part of it. So that is certainly a factor plus the Ukrainian situate -- ongoing war in Ukraine. On health care, the situation -- I mean, obviously, we had a lot of -- we had good growth last year. There is a slowdown with some of the major pharma companies post the pandemic, post-COVID, that some of them have been beneficiaries, others that are reorganizing. And that's, again, it's not as dramatic as in the tech sector, but it is leading to, again, some delays in decision-making. And we will be sort of in a sense responding to that. But it doesn't change my longer-term confidence in the health care sector at all, but we need to really make sure that we've got the right people in the right places. Dan, I don't know if you want to add to any of that.

Maybe just one point on prices. It is clear. I think that Ipsos has been successful in increasing prices in 2022 and facing the inflation wave. We are obviously continuing to do that every day. The point is that because of the inflation globally is decreasing a little bit, even though current inflation is not decreasing. It is true that it is a bit tougher after 2 years to keep on increasing prices. So we are doing that. But the discussion with the client can be a bit tougher than it used to be in the first place.

And we obviously -- yes. And pay rise is obviously under pressure. We're going to be pretty disciplined about that. But you've got the full year effects of May pay rises last year, now visible in our books this year. But we'll be very disciplined on that.

Okay. Understood. And just one follow-up on the comment you made there, Ben, on lower headcount in Q1. Is that versus December or versus March last year?

Versus December.

Versus December. Roughly how much in percentage?

I don't think we're publishing at the moment, but it's...

A little bit lower.

Let's put it that way. Yes.

[Operator Instructions] We have one question coming from Peter Testa calling from One Investments.

I was just trying to ask a bit about the link between pipeline or intake and revenue. Can you give a comment of accelerating in March on the signings after a period of less execution maybe of bookings into Q1. Would you expect -- what sort of time lag would you expect between the pickup in bookings and the pickup in revenue because of the slower Q1, now that'd be short? Do you expect it to be quicker? And then associated I have a follow-up for that, please.

I think it pretty much depends on the project. So typically, we have an average maturity of project, which is a few months, 4 to 5 months. And typically -- so if you have a project of 3 months, for instance, the recognition of revenue would be from the start of the project until the end of the project. So it would be over the 3 months. If it's a project of 1 year, the revenue recognition would be over the year. So typically, on average, because we have 3 to 5 months average maturity, you would expect the impact of the order book in revenue to lag by a few months. That's typically the average. But again, I mean, it's important to understand that if you do a project with Ipsos.Digital, that's pretty much a 24- or 48-hours project and for others, big projects like public affairs or tracking, it could be a year or many years project.

I mean, in one country, we won EUR 20 million of work. But of course, it's spread over the next 3 to 5 years. So, and that's just one team in one country. So yes, that's the challenge. But I think the next 3 to 5 months, it should -- what the bookings will be visible.

Yes. Okay. So the exit rate will be, say, visible as you -- in revenue terms as you exit Q2?

We should certainly see, and if we, we should certainly see a pickup.

Yes. Okay. And the other sort of associated question is that you've been good at trying to isolate COVID work, but there's been other elements which have been harder to define. I mean, obviously, in medical pharma clients, there's been a bit of blowback and the COVID and some of the government work. There's been budget that's been reallocated. So it's a bit harder to understand how that work -- that budget might get reallocated now. And I was wondering if you saw any sense on some of the customer segments, which were, say, spending a fair bit of COVID, who aren't to the extent to which they're now making progress and reallocating that budget into other Ipsos projects, are there other data projects, which are important to them, whether you've seen that now whether you expect to see that later?

Government is, as Dan has mentioned, is up 8%, putting aside the COVID work. So we can -- and we can see actually in some of the larger countries where we have our public affairs and government business concentrated, we can actually see a pickup in demand and particularly for very major contracts. So government, I think, is fairly clear. What probably has been more of a surprise in a way is the sort of the disruption in both the tech sector. And I mean I think the challenge for the tech players is that they saw accelerated growth during the pandemic and thought that there was some new paradigm that had arrived and now they've they returned to growth and they still employ many more people than they did in 2019, but it's not quite the growth that they were expecting. So they're reappraising their model. Generative AI is also a massive disruptor for them. It's a massive exciting opportunity for us in terms of applying it across our business. And then finally, health care, again, I think it's a very mixed bag depending on which clients you look at. And that certainly is slowing things down. But again, we will make some changes there, and we should see that pick up.

Okay. And the last question I had was just on the comments about employee, but on the other hand, on the margin, you talked about mix and productivity. I was wondering the extent to which the employee count is reflecting business mix changes as much as any particular leaning on the productivity at this stage. I mean, obviously, the less people clipboards more data and more strategic tool the very different headcount requirements, whether that's driving is mostly?

It's not -- it's so much that is trying to adjust our general desire to keep pushing on productivity as a business.

The next question comes from the line of Stephen Benhamou, calling from BNP Paribas Exane.

Just one, just quick one to you about your CMD in June. You're mentioning the fact that you will update your plan. Do you mean -- is it, I mean, a focus on a business perspective? Or would you update your guidance on a consecutive basis?

It's not. I mean, at the moment, we're not planning to change the guidance for the year or to do so on June 14. That day is really just about talking about some of the parts of the plan, like our plans for acquisitions, the progress on acquisitions, the progress on investments in technology and automation, further digitization and overall how the 3-year or 4-year strategy announced last year is playing out. I think there's a lot of work going on beneath the surface. We will perhaps detail some of that and some of the innovations there. So it's more of that than a business update, which we'll obviously will be reporting normally on Q2 in July.

All right. Very clear. And one last question, please. It's about M&A. So you're mentioning that you've identified a lot of targets in several sectors. Should we expect an acceleration of the M&A strategy in the coming quarter? Or is it more back-end loaded for the end of the year? How should we expect your strategy to be delivered?

Well, we are following the same strategy, but these deals, particularly where you're talking to owners who are selling you their baby can take a long time. I think as I've said before on these types of calls, when I sold my business to Ipsos in 2005, I think it was around 4 years from the first lunch to the final completion of the transaction. Now that may be a bit of a long court shift. But it makes -- it's just an example of why the predictions are making firm predictions I think, is probably not -- is not sensible. But yes, we have a lot of LOIs out at the moment, a letter of intent. So we will see what comes through, and we'll keep you updated absolutely as and when. We can talk a bit more about the detail on that perhaps on the 14th of June.

Another question coming from Conor O'Shea from Kepler Cheuvreux.

Just a couple of questions as there are a few on the line. Just you may have given this number at the start of the call, but on new services, we gave the growth on Ipsos.Digital 36%, but can you give us a sense of what the level for the overall group of new services was in the first quarter? And second question, just in terms of client group. I don't know if you mentioned CPG, how that's trending? Any change in sort of behavior or activity noticeable in that sector. For now, that would be very useful to know.

So on new services, we introduced the new services category in this reporting back in 2015. And I think one of the things and we have been faithfully updating new services in these presentations since 2015, we haven't done that this time because we are reviewing what is a new service. And this is -- it's pretty important because things that were new in 2015 really shouldn't -- probably shouldn't be defined as new in 2023. So we need to -- I think there's a whole range of things that are not in there that we need to put in there, and there's a range of things that we need to take out. So we will update on that on the Investor Day in June on how we're defining that. We probably give you a bit more detail of composition. So you can see exactly what we're talking about. But generally, we have a -- there's a lot of excitement that the knowledge panel work is going well. We've launched that in more countries. This is, of course, a gold standard digital product this allows you to represent your samples very, very quickly, but that includes people who are not even online, where we install the Internet in their house in order to get from the representative coverage of the country. And it's very successful in America and in the U.K. We launched in France, Germany coming up, et cetera. So we will -- I think the bottom line is we'll update that on June 14. And on CPG, basically steady as she goes. You've seen the profits. Most of those companies have made good profits. You've seen our beer clients in the news recently with improved margins, even where the volumes have gone down because they've been able to pass through inflation. So no, we're not -- I mean we -- I think we reported consumers were up 20% last year. It's only up plus 1% so far. But the order book, of course, is further ahead. And we've got some -- we will again have some interesting things to talk about on June 14 I suggests.

Okay. And I guess, at that stage, you'll go into a little bit more detail about the sort of practical impact of what you can see from AI's impact on the industry?

Yes. I mean, I think, yes, it's a really, really exciting time, and we are working at speed and pace with hundreds of people across the company using generative AI to make -- to see how it can be applied safely and securely respecting all privacy laws, et cetera. So yes, we'll update you on some of that, but it's moving pretty quickly, I think, is the bottom line. Actually, then sort of quantifying that in a definitive fashion for you may take a little bit longer, but it is one of the most exciting technologies that I've seen. I mean it's often generative AI is as good as a graduate trainee which from one of our Europe's finest universities with 12 months' experience, let's put it that way.

Which means that we can foresee a lot of things in terms of client-facing business, but also internally because generally I can be very useful in terms of being more efficient on the way we script the questionnaire and the way we…

Translations.

We report to the clients, exactly. So it's both external and internal.

And we need to make sure that it's applied consistently safely, et cetera. And obviously, you can't put any proprietary information into ChatGPT because it's in the public domain effectively. So -- but we have, as I say, sandbox versions of some of the beta tests from some of the major clients, major producers of or major tech companies that we're looking at the moment. So we'll -- we should have a more sort of structured update on that, I think, on June 14, but it is really -- I mean, I am -- every week, I am pleasantly surprised.

And just one last question. Could you give us a sense of what -- roughly what proportion of your revenues have come from -- still come from survey-based activity as opposed to observational?

Well, that say…

Or Other?

I like to say hopefully it's an interesting observational basis. I'm not sure we have a statistic on that. We think that about 68% perhaps might come from surveys as opposed to observational work.

68%?

Yes, 2/3.

And how has that trended for you the last 5 years?

Well, obviously, the observational work is rising. The survey portion is going down steady. But there's life in surveys yet. I think the key point, and this is a serious slide that in a sense, if you look at the media sector, there is still print. Print is suffering, but it has a place. And in our -- even in our world, and it's very -- part of our very deliberate strategy is to maintain the ability to speak to all of humanity. And as Google put it, 1/3 of the world is not even online yet. So we need to be able to reach everybody. And to do that, we need good off-line capabilities in some markets. So -- and that's important to us actually because of -- particularly because of the requirement often for governments to be able to reach property representative samples. So we, in the same way that the media space is fragmented, but nothing quite tight. Radio keeps going, it becomes digital, et cetera. In our business, we're obviously digitizing more and more, online data collection rises every year, which is great because it drives better gross margin. But actually, the ability to reach everybody and that was a major feature having that physical infrastructure to be able to do the COVID work, which was a huge opportunity for us. That's something that we won't abandon likely.

There are no further questions. [Operator Instructions] Well, there are no further questions, so I will hand you back to your host to conclude today's conference.