$YOU.L: is YouGov really an AI loser? | Jonathan Cohen, Zipperline Capital

You're about to listen to yet another value podcast with your host, me, Andrew Walker. Today we have my friend John Conan on from Zephyr Line Capital. You're going to enjoy this podcast. I we started off 30 minutes just riffing on the UK market, which I've talked about quite a bit over the past 18 months on the podcast just jokingly referring them to an emerging market. It's a market that I I wish I could crack the code on, but there are a lot of cheap stocks there. There have been a lot of takeouts. It's a really interesting market. So we talk all sorts stuff about market there and John spends a lot of time there, which is why I ask him. And then we dive deep into YouGov, which is a stock that John owns, so you should see the the disclaimer at the end of the podcast in the show notes, but we dive deep into that. It is a really interesting company because it trades quite cheap. They say they're going to start buying back stock and they canceled the dividend to do a stock buyback. And the the most interesting thing about it, and we'll discuss it in the podcast, is the market thinks these guys are an AI loser and as I say, the first slide in their earnings deck that they just sent says AI is going to be great basically says, I don't have the quote right by me. Says AI is going to be great for them. And John thinks these guys are an AI winner. They think they're an AI winner. And when you've got a company that trades for probably six to seven times EBITDA and it's really EBITDA because there's no as we joke, there's no stock comp in the UK. So it trades for a very cheap multiple. When you've got a company that's trading for a very cheap multiple, the market thinks they're an AI loser and they might be an AI winner. Well, that's got the potential to be a really interesting stock. So John's going to get into all of that. We're going to get there in 1 second, but first a word from our sponsor. This podcast is sponsored by AlphaSense and more specifically my upcoming AI webinar with AlphaSense. Look, the AI landscape is crazy if you're an investor. It's crowded, it's confusing, everyone's selling you to adopt AI, but nobody is telling you what tools to use. How do you adopt AI? Should you be focused on using it as a superpower Google? Should you be building your own tools? How do you get used to it? All this sort of stuff. I personally find it's a lot of experimentation. It's a lot of fun, but it's really confusing and it's really scary. So, anyway, I I told AlphaSense about my problems and they organized a webinar to try to help out. I'll be sitting down with Dave Wang of Wall Street Prompts and Ben Collins of AlphaSense to break down the modern AI stack for investors. What horizontal platforms like OpenAI and Claude and agent workflows and finance-specific intelligence tools, where each one can actually fit and help in a real research process. So, if you're trying to get better at AI, improve, develop AI-enabled workflows, uh you're not going to want to miss this webinar. Join us on We're going to record it next week, middle of like June 18th, and it'll be going live June 25th. So, there'll be a link to register in the show notes and, you know, please feel free to lob in any questions you have on using AI, whether they're general tools or AI-specific tools like AlphaSense. So, thanks, AlphaSense, and I'll see you for the webinar soon. All right, hello and welcome to yet another Value Podcast. I'm your host, Andrew Walker.

With me today, I'm happy to have on for the first time my friend, Jonathan Cohen. Jonathan, how's it going?

Doing well, man. Thanks for having me on. I'd like to say, long-time listener, first-time caller. Huge fan of the podcast.

Well, look, anybody who goes to Sweetgreen as much as I do always has an invite on the podcast.

We should have done a live event.

Well, we'd be waiting in line for our salad all day. That would be the issue.

I should mention Jonathan Cohen from Zipline Capital—that's his firm. We'll get into all of that in a second, but before we get started, a disclaimer: Nothing on this podcast is investment advice. That's always true, but I think Jonathan and I are going to talk a lot about UK stocks, which tend to be on the smaller side. As we all like to joke, the UK is an emerging market these days. So, heightened risk and all that. See the disclaimer at the end of the podcast in the show notes.

Jonathan, we've got a stock in particular we want to talk about, and we can dive right into that if you want. But I very rarely get someone who specializes and focuses on a particular market, and I know you would do anything, but we talked before—the vast majority of your portfolio is in the UK.

I've talked a lot on and off about the UK over the past 18 months on this podcast. I'd love to get your overall thoughts on the UK market in general: Why would somebody who sits 4 blocks from me right now on the Upper East Side, and can invest anywhere in the world, hone in on the UK? Let's swap some thoughts on that, and then we'll dive into the specific stock, if that makes sense.

Yeah, absolutely. I think of myself as Ted Lasso, but Ted Lasso season 3, not season 1, where I'm out of my depth.

I've been investing in UK and European public equities for about a decade. Zipline is specifically set up to pursue absolute-return opportunities, long or short, specifically within UK and European small- and mid-caps. So, the entire portfolio is focused on that.

In terms of why the UK and Europe, what I'd say is that I have no historical connection. Out of business school, I spent some time in the Tiger Cub network, focused largely on U.S. names, and just found the game to be incredibly competitive. The areas of competitive advantage—where “edge” is a dirty word—were sparse and difficult to find.

For me, honestly, it's simply game selection. As I know you're aware, there's less competition and greater inefficiencies in these markets. They're far less liquid, and there's far less analyst coverage.

A sell-side analyst job in the UK is largely a job where you get 6 to 8 weeks of vacation. It's not paid that well. They have a corporate-broking system in the UK, so a lot of it is not independent research. There are far fewer dedicated long-short investors, so it makes for a lot of opportunities on both sides.

In particular, increasingly so for me, one area that's a particular focus with YouGov, the name we're going to talk about, is shareholder engagement. There's a lot of low-hanging fruit around capital allocation, market communication, and corporate governance.

These are, frankly, things that, if you think about what Dan Loeb was doing back in the '90s in the U.S., these UK PLCs haven't had to think about before and are really coming into focus.

That's a great overview. Let me tell you, as someone who's dipped my toe in there, a lot of what I've come to see in the UK is that you find a stock that's obviously too cheap. Whether it's quantitative—you're looking at it and saying, “Hey, this trades at 6 and the U.S. peers trade at 12”—or you do more fundamental work and say, “Yes, this is too cheap,” I could point to several examples, whether it's YouGov or other companies we're talking about.

Then what you come to find is that it might not be too cheap because the market is wrong. It might be too cheap because the market's slapping a huge corporate-governance discount on it.

We can talk about all the reasons for that: small insider ownership, the fact that CEOs generally aren't paid based on stock price, and all this sort of stuff.

I've come to see the UK market—I don't know if this is right or wrong—as a market where the only way you get paid is when a private-equity firm comes and puts you out of your misery. If you can time it correctly, you can and will make a lot of money in the UK market.

Timing it correctly means buying the stock before the private-equity buyout. But if you don't, you're sitting here pulling your hair out. Every year, they come out and their earnings are a little worse than you thought they should be. It's still really quantitatively too cheap. The cash is building up on the stock. They probably refuse to buy back shares, or maybe they do a token share buyback.

Unless you get that buyout, you get nothing. Maybe the answer should be that we should all just sit around and wait, because when the buyout comes, the premiums can often be huge.

But I thought I was going to be buying these things at 6 times earnings. If earnings don't fall off a cliff, you should make around 18% annualized without multiple expansion. I kept seeing them go from 6 to 4 times earnings. Maybe I was too impatient, but I would only make money if the buyout came.

What would you say to that? It seems like it's more of a corporate-governance question than anything.

Yeah, I think there are a few things in there, and the pulling your hair out definitely resonates. If we had done this 5 or 10 years ago, I'd have a nice head of hair.

As poofy as this guy right here.

So, the first—I don't want to say mistake, but the first thing that I want to pick on—is the biggest lesson I learned when I started investing in the UK and Europe: You never, ever compare them to U.S. or even, frankly, European multiples. And you're laughing, but I'm totally serious.

I will never pitch anything, and I will immediately dismiss a pitch that says, “You have a UK business trading at this and a US business trading at this.” There are a lot of reasons, including many of the things you discussed: higher growth rates and different views on regulation. There are just a lot of different things, so I will never do that.

Can I ask for one nuance in that? I’m just going to use homebuilders because the homebuilders have been in the spotlight and stuff. Would you never compare a UK homebuilder to a US homebuilder?

That completely makes sense. But if you had a business—and I can think of a few in the UK—that has a lot of international components, would you say the same thing about multiples there, or would you be more flexible? Because I could see it going either way.

No, I would say the same thing. If you think about the multiples that you need to adjust, first, as you pointed out, capital allocation isn’t as optimal for a lot of different reasons. That’s changed a bit now; we can talk about that. Buybacks have become a lot more frequent. Tax rates are higher in the UK.

So even if you are a global business, if you are domiciled in the UK, leverage is very different, right? Leverage may be too high in the US, but frankly, leverage in the UK is suboptimal. Our mutual friend Ben always laughs, but a very wise man told me that there are only 3 natural buyers of UK stocks: share buybacks, short covering, and takeovers, right? There is just a thing about being on the UK stock market where there isn’t a natural flow to those businesses, so the multiples should definitely be lower.

There are arguments for those businesses to be relisted in the US. There’s a mixed track record on that. It’s really just dependent on the business. Something like Ferguson, I believe, has done well; CRH has done well. So it really just depends on the business.

I would say YouGov is a global business. I own a number of different global businesses in the UK; they just always trade at lower multiples, for the most part. There are some businesses that trade higher. I mean, YouGov is really interesting.

What I’ll tell you, just to riff on what you talked about, is: how do you make money other than through a takeover? I think there are a few key things I’ve learned over time. I spend a lot of time doing diagnostic analyses of what’s worked, what hasn’t worked, and where I tend to focus today.

First of all, you mentioned incentives. UK management teams are not incentivized, let alone with incentive compensation; even just on base salary, they’re not paid enough to really think about shareholder equity over the long term. One thing I definitely do now is avoid owning businesses where there isn’t more than $1 million of insider ownership, so I screen out all businesses that don’t meet that.

The other nuanced thing is that I tend to look for businesses that already have a ton of analyst coverage. Where you can get in trouble is when you own a business that has always traded at 3 times EBITDA, which you think will go to 6 times, and has 1 analyst covering it, and you expect that once it gets figured out, 5 analysts will cover it.

The names I tend to look at—I call it betting on things that have happened already happening again. Businesses like YouGov have a lot of analyst coverage because they’ve been at a far higher market cap in the past. They have traded at far higher, even normal and high, multiples in the past, so we know that they can get there. What I’m solving for is trying to bet on things that have happened before happening again, rather than trying to bet on something that hasn’t happened before.

You know, just to riff off what you said, it’s funny you said, “Happen before, happen again.” Obviously, we are in June 2026, and I don’t think this is the dot-com bubble, though there are some—especially more of the quantum-computing, pie-in-the-sky things, maybe some of the much lower-quality things.

I don’t think this is the dot-com bubble, but there are elements of a dot-com bubble. I remember, throughout my career, you would hear, “We’ll never see something as crazy as the dot-com bubble again,” and then you get the SPAC bubble and the meme squeezes of 2021, which probably weren’t as crazy as the dot-com bubble, and then you get this. You’re like, “Man, everything that happens before happens again.” I know that’s not quite what you’re going for.

No, I’m not smart enough to bet on things that haven’t happened before. I like to use history as a guide. It’s just my philosophy.

On analyst coverage, it’s interesting you mentioned that because you front-ran one of my questions on YouGov. They’ve got pretty good analyst coverage. Just to spoil the plot, this is a $500 million EV company, if that, and it’s covered by multiple large banks. I think UBS covers them; JPMorgan covers them.

I tend not to do a lot of sell-side research, but when I’m looking at it, I’m like, “Oh, this $500 million market-cap company has more big-bank coverage than a lot of the $2 billion market-cap companies I look at over in the US.”

Yeah.

Not you guys specifically, but how much do you think about analyst coverage when you’re looking at the UK? If it’s a company other than YouGov, do you think, “Oh, this only has 1 analyst covering it; as soon as 2 more banks pick it up, that could be the thing”?

I mentioned jokingly that the UK is an emerging market. One thing that emerging markets have, which I have seen in the UK, is that when an analyst changes their recommendation, it can be a big catalyst for the stock. You’ll see a lot of stocks up 15% on an analyst change.

How do you think about that interplay? Again, with pure value investing, 10 years ago I would have said, “Nobody cares,” but especially in a market that’s maybe a little more inefficient, thinking about those things can help you with timing. How do you think about that?

I think about it a few ways. The important point to mention, which I don’t think people think about enough, specifically in names that are down a lot and are relatively low-volume, less-liquid stocks, is this: you used to have, let’s call it, a £1 billion business; it’s now, on a market-cap basis, close to $250 million to $300 million, right?

If the banks are making money via trading, they’re not really making money via research anymore. When it was a £1 billion business and you had a lot of interest in it, banks were making money trading the stock. What I’ve seen a number of times—and another business, not to go off on a tangent, is Greencore, which is a sandwich maker in the UK—and this was kind of my playbook.

On my last podcast with Alex Roper, we mentioned Nomad. Greencore—it’s funny you mentioned them—is a loose peer that I had in my database. Their margins are around 10%. Can I trust Nomad’s at 15% when it is branded? I literally had their PDF up 10 minutes ago, so I’m just laughing.

It’s a brilliant business, and it went through what I hope will happen to YouGov over time. What you realize is that when it goes to a $200 million market cap, the analyst coverage stays, in theory, but no one is interested in it because it becomes too illiquid for people.

The bank itself isn’t making enough trading it because, even if they’re trading the same volume, it’s 20% of what they were making before on a dollar- or pound-basis. On the analyst coverage, I find it to be basically tracking the share price, right? They wait for ultimate certainty on the business, to have ultimate visibility, likely when I’m selling, and then they upgrade it.

What I will say, and I find helpful, is that the UK is a bit strange, where you have a corporate-broking relationship. There will be a NOMAD, a nominated adviser in the UK, which you’re required to have. A lot of times, you see UK stocks where they may only be covered by 3 analysts, and all 3 will be the UK brokers, right?

Usually, those brokers have buy ratings because they’re doing business with the company. So I’d say it’s probably more inefficient on the short side than it is on the long side.

On the long side, what I use them for—and I think sell-side analysts have a real role to play—is industry knowledge. A lot of these UK analysts have been covering the same industry for a very, very long period of time. I get up to speed quickly because I have a concentrated portfolio and spend a lot of time, but I’m an industry generalist for the most part, so having a lot of industry knowledge is wonderful. They have access to management teams.

The difference in what I do is that I spend a lot of time with the corporate brokers, and this gets to what we talk about with shareholder stewardship. I call it a nice way for activism light, and those guys are a conduit into the management teams and boards, helping them think about market communication or capital allocation.

I would say it’s just different from the US. I think that you use them for different reasons, but I never use them for share recommendations.

Last question. You mentioned talking to the brokers. One thing that I've certainly torn my hair out over, and I know you have as well. Look, you mentioned capital allocation a few times.

I think there have been UK firms—and maybe I'm just looking at more, and a couple have popped up, so it's starting—but I think there has been a little bit of a shift toward these firms looking to do share buybacks. I wouldn't say we've gone the full Japan, where Japan was leaning on companies: You cannot trade below book value; you have to return capital. I think they've actually backed off a little bit.

I don't think we've had that yet, but I wouldn't be surprised. I think we're starting to see some reception there, and I do think the UK is looking around and saying, “Hey, if Andrew and Jack are out here saying the only way for UK stocks to work is for private equity to come put them out of their misery, and that's like the only buyer, as you alluded to the matter of share buybacks…”

I think they're going to have to come to grips with that because they don't want no publicly listed companies. Have you started seeing or feeling the UK—whether it's companies, regulators, whatever—become just a little bit more receptive to, “Hey, maybe we do need to buy stuff back. Maybe everything trading with net leverage under 1 is inefficient”? Have you started to feel that a little? I have, but I'm biasing the witness.

Yeah, no, 100%. When I first started investing in the UK and Europe, not a single business I owned was doing a share buyback.

Now, I spend a lot of time—75% of my day is focused on getting companies to do share buybacks and think about capital allocation. I'd say basically 80% of my portfolio today is doing a share buyback. So it has definitely changed.

It's definitely changed from the corporate-broking perspective, particularly because they're incentivized to do it because they make money doing it. So they're the ones who are assigned to do the buyback on behalf of the company. It's another way for a bank to make money in a difficult market.

That's one. I'd say the biggest thing is educating management teams and brokers as to why it makes sense to do it. Literally, the academic theory and the math behind a buyback—it takes a long time, but you kind of see the light bulb go off, and it makes a ton of sense. So from an academic perspective, 100%.

The one thing that I would say—and I'm being a little cute here, thinking about stocks versus businesses—is that, beyond the math of it making sense when it trades at an attractive valuation, there has always been, probably in the U.S., an idea that doing buybacks for an illiquid stock makes it more illiquid, so it just creates a further overhang.

I actually think it's the exact opposite. You may know this, but in the UK, you can only buy back—I believe it's a maximum of 25% of your daily volume per day.

Tighter rules, yeah. I think the U.S. actually has that rule, too, but you can't do something like that.

So what happens is, if you have a relatively illiquid stock—let's say it's a £300 million market cap—and you go to do a £20 million buyback, it could take months, if not a year, to do that.

As I mentioned before, there are no natural buyers of stock in the UK. So what it does, just again to be cute from a stock perspective, is add a natural buyer of your stock in the market every single day, which really helps. Again, beyond the math of it just being attractive, that has been a harder thing for brokers to understand because a lot of these brokers have been doing it for a long time.

But I think the buyback thing, I'd say, is one of the most encouraging things in the market. The leverage has not changed whatsoever. I think where I'd say people probably get frustrated with me—I mean, even frankly at YouGov, when I speak to people who don't focus on the UK in particular but move to the UK and want to do an activist thing—is that I've kind of come to grips with the idea that the market doesn't like leverage.

Frankly, if that's the worst thing in the world, that's kind of okay. If they'll put on 1x leverage versus 2x or 3x, again, it kind of limits the downside risk a bit to the equity. So I'm kind of okay with that. That has not changed.

On the leverage, I kind of agree with you. It's just maybe because I'm getting more impatient in my old age. It is still really quick. We go from 0x leverage to 2x leverage in this business that probably could support 3x leverage. You can return a lot of capital quickly.

On the liquidity front, it's funny: our mutual friend and I were out to lunch the other day, and we were talking about this. We've had the same pushback from companies, right? “Oh, our stock is illiquid. We couldn't…” It's like, okay, I do understand if you're trying to buy back £100 million of stock and your stock trades £100,000 a day, it's going to take a long time. But still, on the margins, it helps.

All these companies say, “Oh, it's illiquid. If we buy back, we'll get more liquid.” I would like a company to show me the proof point of that because, A, if people know you're going to be in the market for 25% of the volume every day, I think a lot more of the trading firms will just have algos that are spinning up. I think liquidity becomes—

And B, if you're—it's not like you have to buy, right? But if you're going to be in there and buy 25% of the volume every day, and it's going to create this liquidity problem, your stock is going to go up, right? So you're going to get your stock up, and then when it goes to a point where it doesn't make sense, you can always stop.

But all of it just feels like a false dichotomy to me. I would love someone to show me the proof point of, “Oh, we've bought till our stock's gotten so illiquid that we've got some weird rump.” And I also like to say, hey, if you do that, the stock's probably screaming higher because you're buying cheap. So I don't see the problem here.

Yeah, I mean, look, I know plenty of illiquid stocks that are expensive and plenty of liquid stocks that are cheap. So I don't think the argument holds at all.

I think the other thing that I would mention that's really changed—and just because I have a little history of it—is that back in the day, Neil Woodford and Mark Barnett were like the kings of the UK market. They were dividend-income investors. The easiest thing a UK business had to do to get a natural flow of capital was to start a dividend.

So that is why all of these UK businesses have dividends. There's a dividend culture. And, look, Neil and Mark are not in the market anymore, and dividend-income investors have massively dwindled as part of the register. So it's just been a real education.

One of the things that made me the most happy, and was a direct result of hundreds of emails and calls with the management team and board of YouGov, is that I was able to get them—they were paying a £10 million dividend, and when you're a £1 billion company, that's nothing. When you're £200 million, that's something.

We've now gotten them to agree to turn that into a buyback at a minimum. One of the things that I've worked on, to your point about turning off the buyback, is this idea of a very flexible capital-allocation plan.

It's not that one commits to doing a dividend or commits to doing a buyback at all times. It's having the flexibility to say, “Our stock is cheap. Let's buy back stock,” or eventually turn it into a dividend, and so having that flexibility. That will be a longer evolution, but we're getting them to do buybacks.

And that's a great transition. Let's talk YouGov. I think people will probably say, “Hey, that was really fun, but we've talked for almost 30 minutes about the UK, and you mentioned YouGov multiple times,” so people can probably see here how excited you are about YouGov.

Let's turn to that, and I'll just start with the question I like to start every podcast: What is YouGov, and why are they so interesting? Before you get there, I have to note their CEO's name is Stephan Shakespeare, and that is quite the name, man.

It's even better when you pronounce it that way.

Stephan. Quite the name. But what is YouGov, with their CEO Stephan Shakespeare, and why are they so interesting?

Yeah, just interrupt me. I love this business, so I'll go on a bit of a tangent. But basically, in terms of talking about the business, I would say YouGov probably screens at first as a market-research firm at first glance. In reality, it's a proprietary data business.

Everything they do and sell—products and services—is derived from their proprietary database, a proprietary data set of attitudinal data. This is from 30 million people across 60-plus countries over the past 20 years.

Your next question is kind of, “What does that actually mean?” Which is a very good question. Businesses basically use YouGov to track things like awareness, perception, and where people are in their purchase funnel for various brands and products.

Just to give you a more layman's example, this can be businesses tracking their own brands. BYD, which I'm sure you're familiar with, is a recent customer. They're trying to expand into Europe. They are using YouGov to better understand their customer awareness and perception in Europe.

Things like Marks & Spencer, which is a big UK retailer, use YouGov rather than having their own in-house data collection. The best example, I think—just because I spoke to one of them as a customer—is a book publisher who has a new book and is trying to figure out what the right demographic, or best demographic, for that book is. They can take that demographic and market it to the right retailer, right? So you can use all of YouGov’s services there.

It can also be advertising agencies and marketing firms helping their own clients. Interestingly, it can even be people like us—investment firms. A great example is when the tariffs came out in April 2005: YouGov actually had a hedge fund client come to them to track consumer sentiment toward tariffs and pricing sensitivity in all geographies.

That’s a broad way of thinking about what it is and what it’s used for. In terms of how it’s used, it can be consumed in a few different ways. These are largely syndicated tracking products, so you go on and see a syndicated product where they have Marks & Spencer, your suppliers, and your peers, and you can check how you’re doing. This is basically refreshed daily.

You can have custom tracking or surveys if you want a very specific audience that is targeted, or if you want to do it just for a specific brand or product. Increasingly, they also have a self-service platform, which is enabled by AI as well.

The business is recurring or repeat—they don’t really report it, but it’s about 60% of revenue and 70% of operating profit. Those are the rough estimates. So again, it’s much more repeat, recurring revenue than your big-project, large consulting or market research firm.

The way I think about the world is that I divide value investors into different camps. Broadly, this is a massive generalization, so no one get angry, but you have value investors who focus on situations—turnarounds, regulatory events, that kind of thing—and then you have value investors who focus on business quality and long-term compounders.

I’ve done both, and I combine the two to do what I call a business-plus-situation framework. It’s not that creative. When I think about YouGov, I want to own a high-quality business that is growing, with a large addressable market, high barriers to entry, barriers to being good, and aligned management teams. But I also want to understand why the opportunity exists today and why I’m competitively advantaged to own it.

When I think about YouGov in terms of business quality, I start with barriers to entry. YouGov, as I mentioned, has one of the few truly proprietary, high-quality, trusted data sets of attitudinal data. Kantar and Nielsen are other large-scale players that have something similar. The important difference is that they don’t have the same magnitude of longitudinal data, which is a fancy way of saying they don’t have all the historical data that YouGov does.

YouGov has been doing this for 20 years, and they don’t have the same magnitude of historical data. YouGov has by far the longest-standing data set, and there’s real value in that because you’re looking for changes versus history. Importantly, their data is more frequent. Kantar and Nielsen might be weekly, monthly, or quarterly. YouGov’s panel data is refreshed daily.

The best example, I think, when I think about YouGov and want to understand the value proposition, is: Do you remember when Dieselgate happened to Volkswagen?

Of course.

Volkswagen reached out that day to YouGov. Why YouGov? YouGov had been tracking consumer perception and awareness about Volkswagen for 20 years, even without Volkswagen being a customer or client. Volkswagen called them because they wanted YouGov’s panel to track, in real time, what was happening to their brand perception since the scandal and whether the things they were doing were working.

Another solution would have taken months to set up, and you would be flying blind in the heat of the scandal. That’s really where YouGov is wildly valuable and helpful.

You mentioned being Ted Lasso. I try to be like a golden retriever and ask eager questions on this podcast, so I’m going to ask question 1. You say YouGov gets the panel, right? I understand the panel is made up of humans, but what is this panel of people who, in real time, are responding to Volkswagen’s brand?

I’ve seen the thing that pops into my head: during a presidential debate, everyone has a dial, and if somebody says something good, they shift it. They can tell you what happens during the presidential debate for 2 hours. But with Volkswagen—yes, Volkswagen Dieselgate—if I’m a casual news consumer, I see it and think, “They cheated on their diesel emissions. That’s terrible.” But in real time, how are they even tracking that across the panel?

What’s interesting about YouGov is that they have uniquely high brand awareness, particularly in the UK. You may have even heard of YouGov. YouGov is one of the 1 or 2 most-cited research sources in the world, right?

I mentioned elections. The Economist—you like it if once a week there’s a new poll, and a lot of them are the Ipsos poll or whatever. One of them is the Economist/YouGov poll, if I’m remembering correctly. That’s one of the big polls that comes out. So, yes.

Particularly in the UK, YouGov is known as a brand. You and I could go to the YouGov website today and ask, “Where does Greencore’s brand rank in terms of competitiveness?”

It’s interesting because that’s a sector where it’s not typically known for that. You and I know Kantar and Nielsen, but your average person on the street wouldn’t. Whereas if you go out on the street in the UK, everyone will know YouGov.

Why does that matter? That matters for 2 reasons. First, how do they get these people? You have a loyal fan base of people who love doing this. They like being part of the YouGov brand and part of the YouGov fan base and loyal base. They get a gift card.

Why does it matter that YouGov has high brand awareness? As I mentioned, there are 2 reasons. First, it strengthens the quality of the data set. To your point, who are the people doing this? One of the reasons why the data set is so valuable is that, just as when we’re looking at a business, you want apples-to-apples or like-for-like comparisons.

It becomes incredibly valuable if you have high retention and lower churn among your panelists. YouGov has by far the highest retention of panelists—people logging on every day and getting prompts. They love doing this. They love being part of it. Secondly, it reduces costs because you have a greater incentive. Again, people like being part of the brand. They think about it as a second job or a side hustle.

I’ve seen LinkedIn profiles that mention the person is a YouGov panelist.

How much do they pay the panelists?

I don’t know off the top of my head. It’s not as high as others. It’s usually in things like gift cards: every 10 responses, you get a gift card, something like that.

Panel costs are reasonably high, but the real cost of the panel is maintaining it technologically rather than the cost of acquiring people. Again, that’s because if you’ve been doing this for 20 years and doing it for a long period of time, you have this embedded group of people who just enjoy doing it.

We’re getting a little off tangent, but it’s probably hard to contextualize for someone like you or me, who would probably never do this as a panelist—or at least I wouldn’t. When we talk about AI, YouGov is very good at showing when things change.

For example, Volkswagen has the scandal, and you will clearly see a drop-off in perception. If you didn’t know the scandal was happening, you would go back to Volkswagen and say, “Hey, I’m just showing you that this is happening.” Volkswagen would then have to say, “Why is this happening? We have to figure this out.”

There would be 2 ways to do it. Volkswagen could do it themselves, or you could hire YouGov to do a custom survey. You would hire a bunch of people, have to go out, take a bunch of time, and spend a lot of money to answer the “why” question.

What’s interesting, relative to why people do this, is that YouGov recently launched something called BrandIndex Voices. At the end of every panel, they now have an AI bot that pops up and says, “I see you changed your answer on this. Why?”

The panelists will talk to the bot. About a quarter of panelists today are engaging with it. What’s interesting about it, to get back to the original point, is that there are panelists who stay on with the bot for an hour and talk about things that are totally unrelated to the question they’re asking because they like having someone to talk to.

It’s amazing for YouGov because they continue to collect increasing amounts of data on each person. They showed it at a demo at a recent investor day, where they talked about what kind of music and movies people like. They get more and more demographic data.

So I think it’s hard to appreciate why someone would do this, but there is a loyal group of people. All the calls you do with competitors and customers talk about the quality of the data—again, in particular, the high retention and the lower cost—which really creates a very, very high-quality data set.

Let me use that to ask what I think is the most frequent question I ask: What is the company, and then what is the market missing? I think you’ve started to hit that. But let’s talk about, I think, the elephant in the room.

You know, the stock, 2 years ago—2 and a half years ago—was trading at 1,000. I think they had a bad earnings report, but over the past year, let’s just round up and say it’s been cut in half. And I think a big reason is—

80%.

That’s if you go back 2 years. I think last year it was—either one, whatever. It’s down a lot. I think a big reason for the drawdown in the past year has been AI fears.

The company will come and tell you—I was reading their H1 earnings report. God, I hate the UK H1s and everything—but the first slide they have is, quote, “AI is creating a powerful structural advantage for YouGov,” right? And this is really interesting to me because you led off when we were talking about buybacks. The company is canceling its dividends to buy back shares, right?

So that’s a sign. The company is saying AI is creating a powerful advantage for YouGov, and we’ve started to talk about that. We can talk about that more, right? But the market is saying AI is a disruptor. These guys are part of the SaaS apocalypse. They are down 50% over the past year. Why is the market scared of AI for this company?

Yeah, sure. We can also at some point go back and talk about some of the idiosyncratic things that caused the drop.

Hit him, hit him. Yeah, yeah.

Let’s do AI first, because as you pointed out, it’s the elephant in the room. Let me talk about this in 2 ways. First, I’ll share how it relates to YouGov specifically, and then I’ll riff on how I think about this in the broader context.

With regard to YouGov specifically, I think the market’s knee-jerk reaction, as you appropriately pointed out, is that they do surveys or something like surveys, so they must be an AI user. What’s interesting to me is that this is not a business like Qualtrics, right? Qualtrics, as far as I understand it, is an online form distributing surveys that used to be done by hand.

In YouGov’s case, one literally can’t get the data from a YouGov survey without access to the proprietary data set and the proprietary panel of 30 million people, with 20-plus years of data. So if anything, applying AI technology to YouGov’s proprietary data set only further enhances the value of it; it doesn’t detract from it.

I gave you the example of BrandIndex Voices, right? That’s amazing. I think that’s a game changer, because it now allows YouGov to answer the “why” question. It creates a new service for them. They can do tens of thousands of these every single day, at minimal to no cost. This is something that would take weeks or months and be super expensive for the client.

Can I pause you on that one?

Yeah, please.

They’ve got 20 years of data, and that’s something you and I cannot go recreate, right? That’s in the past. Unless we can discover time travel à la Harry Potter and turn the clock back—in which case, we’d probably do other stuff than this panel, right? We’d probably go video call. So we can’t do that.

There is a little bit of moat there, but in the present, you’re basically saying, “Hey, they’ve got the 30 million people and they can do these brand surveys.” I guess my question is: Why couldn’t you and I—we aren’t going to do this, right?—but why couldn’t a lot of people spin this up?

I was talking to a friend and had him over, and he works at a consulting firm. He said, “Oh, we just did this big project where we created an AI customer audience for a company, right?” The company could ask this AI audience things. I knew we were doing this podcast, so I was thinking, “Oh, that’s not quite YouGov, but that’s getting pretty close to YouGov.”

So why isn’t this—I hear you that AI could improve this—but why doesn’t this also create a lot of competitors that can say, “Hey, we can go create a lot of people for them”?

Yeah, so you actually just jumped to the point I was just talking about anyway. This is good. I think the market’s knee-jerk reaction is relatively unsophisticated, as we talked about: that it’s just a survey and you have the data.

I think the more sophisticated take, which of course you have because you’re very sophisticated, is around using synthetic data. This is effectively using AI bots instead of humans to generate human-like responses as data points, right? And so there are a number of ways to think about this.

First, I’d say, not to be too academic about it, but there are a number of studies that have recently been done which basically show synthetic data to be highly problematic. There was a UK organization, I think it’s called Strat7. They basically did an entire study which found the data lacked logical consistency.

The more interesting thing is that there’s a great New York Times article, which you may have seen. I think it was called “The Death of Polling,” or something like that. They talk about a Cornell study which showed that synthetic-data results were basically wholly unreliable because they were full of extreme bias.

The article mentioned a company—I forget the name of it—that ran a simulation using its AI synthetic data of the 2024 presidential election to see what it would come out as. It inaccurately reported that Kamala Harris would narrowly win. At the bottom of the article, The New York Times just calls all of these results pure fiction.

The other thing I’d mention is: When you think about what YouGov does, we talk about YouGov trying to find and measure change, like abnormalities in the trends. That’s what you’re using it for, right? Using one of the examples from before, like Dieselgate, Volkswagen wanted to see how its brand perception deviated from the trend once the scandal broke.

By definition, synthetic panel data is simply extrapolating from past data, right? It’s not a human changing, so you’ll never really be able to use it for what YouGov is providing.

I think the last and perhaps most important differentiation is that there’s a difference between synthetic panels and synthetic data. YouGov actually can, and has started to, use synthetic data based on real human panelists and responses.

So if you want something done at greater scale or, more importantly, greater speed and potentially at lower cost, YouGov can effectively extrapolate or interpret from its real-human panel a synthetic data set. The difference is that this can be audited and actually verified by putting the question to the human panel if the data needs to be verified, right?

If a customer comes back and says, “Look, this answer seems weird,” you can go back and say, “Well, let’s just ask a couple of panelists to verify if this is really how they would answer based on their demographics and this.”

What’s interesting for YouGov is that it tends to be at the higher end of the market, right? These are guys who will pay $50,000 to $100,000 for a service like YouGov. This is actually a product which I think expands YouGov’s use case, where it can be used for lower-end people and lower-end businesses, because it could be done more quickly, at greater speed, and at lower cost.

The other point I’d mention—and it’s a little tangential—is that part of what happened during COVID, and why they saw a little bit of increased competition, is that you had increased demand in 2021 and 2022 for these types of services. YouGov benefited from that.

But you also had a really low cost-of-capital environment, right? So you had businesses like GWI in the UK, Morning Consult, a Swedish business called Cint—which is a fun stock chart to look up—and another one called Dynata.

What all of these businesses tried to do, in different versions, was take a bunch of small third-party data sets, combine them into one data set, and claim it was one large proprietary data set. Basically, what happened is they all raised capital at unicorn-like valuations, and so they had no regard for price and went really aggressively after revenue growth.

Basically, all of those businesses are struggling now. Cint is an amazing example: It’s down 95% since its Swedish IPO. They basically found they had to report fraud within the data set. They didn’t have people accurately responding, and they couldn’t verify panelists.

Morning Consult has done a number of layoffs. With GWI, you can look at the UK Companies House accounts: Its revenue growth has slowed, and it’s now in a debt position. Dynata has filed for bankruptcy.

So what I would say is that, if anything has happened over the past few years, it’s actually been that data quality has become more important. When you speak to people in the industry, competitors, customers, and clients, data quality and accuracy are by far the most important things.

And there are just fewer and fewer people who can do that, right? You need to have all the systems in place and all the quality behind it.

It's one thing I've thought about in the AI world. You mentioned synthetic data. There was actually an article, I think in the FT recently, about a deck KPMG released on using AI; there were all these AI hallucinations in it. It was a funny article, but the one thing that jumped out to me is that there's a line about two-thirds of the way through that says, “Hey, people like KPMG—known brands—in a world where there's a lot of AI slop coming out, the known brands actually cut through because people will say, ‘Oh, KPMG put this out. I can trust them.’”

I'm sure in sports, if ESPN put it out, I feel pretty good that it's something verified, versus if it's on NBA Centel, which I think gets a lot of people with its jokes. There is something to, “Hey, YouGov put this data set out.”

Now, I don't know—there is the question of whether the brand matters for what is being put out or something—but there is something to, “YouGov put this data set out. I trust them,” versus Andrew and John spinning up a ChatGPT and putting the data set out. Our data might be right, but people don't trust it.

I've been wondering if that brand is something that matters. I think this might be what they're saying in the AI world: it's easier for them to do it. They can sell bolt-on services. But I am wondering if that brand is just, “Hey, we can trust this YouGov data,” versus Jotform or whoever else. It's something I've been wondering.

So, again, I think the brand matters not just from the consumer perception, right? But also, as I mentioned, it sounds soft and like a throwaway term, but YouGov—despite most people on this podcast probably never having heard of it or knowing that it's publicly traded—is consistently the number 1 or number 2 most-cited research source in the world for 20 years. Even someone like Nate Silver put it in his top 5 most accurate pollsters.

One thing that's interesting is everyone knows YouGov for polling; they basically make no money on that. The idea there is to create a brand of accuracy, right? They're by far one of the most accurate pollsters around politics and different things like that. That is to create a brand that is trusted and verified and everything like that.

And so, what I've seen when you speak to clients—and again, you can see it in the market—is that the value of a proprietary data set, which can be audited and verified and is proprietary to you, has just gone up increasingly. I think, again, I enjoy being a contrarian: I think that is even more valuable in the world of AI.

One thing specific to YouGov, I think, is a great irony that I like mentioning. Not to go off on a tangent about expert networks—we can come back to that later—but I've started to use LinkedIn a lot more for primary research. I'm constantly doing searches about once a week for every company I own. It's amazing. People overshare like crazy.

You can learn so much about who they're hiring and firing, if they're going on sales trips because they beat a budget, and different sorts of things. It's free and unbiased. One of the interesting things is that over the past few months, YouGov's share price goes down basically anytime Anthropic launches something, right?

A few months ago, or a month or two ago, I noticed that someone from Anthropic called out YouGov, thanking them for helping with their Super Bowl campaign. More recently, Anthropic launched something called a public records series, which is basically surveying Americans about how they feel about AI. Again, they used YouGov. It's pretty entertaining to think about the fact that one of the largest AI companies in the world is reliant on YouGov for a lot of these things.

To do it a little more broadly—and you can tell me if this goes on a bit of a tangent—I think, to your point about the market environment today, one thing that's unique to my experience is that I was at a fund called Coltrane. We were largely net short during the 2020, 2021, and 2022 bubble, whatever you want to call it: SPACs, meme stocks, all this stuff.

That was an incredible experience for someone who didn't live directly through the tech bubble, and I think it's relevant to a lot of stuff today. I think the few lessons I learned and have been thinking about a lot are that experts are cyclical, right?

Whether it's COVID, tariffs, or AI, everyone on social media has a strong, quite often hyperbolic view. There's very little context: are they qualified? They might be. Are they qualified to have a view? Maybe they're not. What are their incentives and biases?

What I notice, and what I would guess will happen here, is that they typically exit stage left as quickly as they appeared once this happens. As a result of that, and I think it's directly relevant to what we're talking about with AI today, my experience is that the market is a particularly poor predictor of medium-term to long-term winners and losers during periods of change.

Not to go off on a tangent, but all you have to do is look at this. I was short Peloton and HelloFresh, right? These were touted as structural winners during COVID, which I was short. I was long Greencore, and I was long a business called Marston's. Greencore makes sandwiches and relies on people going places. Marston's is a pub and relies on people going to pubs and socializing.

These were structural losers to the market, and I was long them, right? All you have to do is pull up the charts of those businesses from April 2020 to today. You can look at the fact that Peloton and HelloFresh are down 80%, and, if you look as of January of this year, Marston's and Greene King were up 75% to 100%, right?

I think what's interesting during this period is that there's a lot of correlation-of-one moves: AI losers, AI winners. During COVID, we had pandemic losers, pandemic winners, and these different baskets. I just think it takes time for the market to flush out what are real winners and losers.

Can I push back slightly on that? I don't disagree and obviously I don't even know if Peloton is the same because there isn't much crossover, but let me just push back a little bit. AI is evolving so fast, and one of the things I've worried about is Chegg.

Chegg, for those who don't know, was like Google for college textbooks, basically, right? I went and looked, and the day ChatGPT came out, the stock was down, I don't know, 30%. I went and looked at its history; it's really interesting history to look at.

Management came on as soon as it came out and said AI was going to be great for us, right? We were going to be able to catalog these books better and give better answers, and the stock went down more. They bought back shares, insiders bought, and the stock—I mean, this was a compounder—has gone from $150 to $100, I can't remember. As you and I were talking, it's $1.12.

Okay.

And there were no physical assets here, and AI replicated it and destroyed it, right? I worry about a lot of these SaaS companies, and I've come around to the view that I can build a tool that's fine for me: it's got rough edges, and it doesn't scale. If you vibe-code your organization's CRM or mission-critical stuff, you're going to die.

But with Wix.com, I wish I had had the balls to short it when I was writing this. It was really obvious to me: a lot of websites are going to get made with ChatGPT, not Wix, right? The whole thing with Wix was that it was easy to build, and then ChatGPT comes out and they say, “We're going to be a winner. We bought Base44; we're going to be a winner there.”

They do a huge tender, lever up, buy shares, and the stock goes down, down, down, down, down. They're cutting guidance—everything. That's what a lot of these SaaS companies are worried about.

So I guess what I'm saying here is I definitely hear you that the market overemphasizes this, but when I look at a lot of the companies today, it seems like maybe it's too much. There are companies—I kind of agree with this—like Oracle or Salesforce, or even Microsoft. These are mission-critical businesses, which are going to be hard to replace.

But at the same time, when I see a Wix or some of these lower-quality SaaS names, I'm like, “Yeah, it's pretty scary there.” There's a lot of terminal value priced in. So that's not specific to YouGov, but I just wanted to ask you—

So, I think about this a lot. Before I get into a broader riff on this, what I would say is that one thing I've been careful about with a number of these businesses that I own is that I don't know Chegg at all. So I may be totally wrong, but if some element of their market is the academic market or education, one thing about owning B2B businesses—which, importantly, are not in academics—is that ultimately, with academics, your end consumer is a student, right?

And so there's inevitably going to be pricing pressure. Even if there's not pricing pressure from the university, there's inevitably always going to be pricing pressure because the consumer is always a student and there's always a bunch of open-source alternatives.

For me, when I think about YouGov or a number of other businesses that I own that are on a similar theme, it's always been, to your point about Salesforce or Oracle, B2B-focused and mission-critical, right? YouGov only sells to businesses.

You know, if anything, the idea is that YouGov hasn’t pushed prices as much as they should, and there’s probably more value that could be attracted from clients for YouGov. So I’m very focused on that.

The other thing I’d mention is—and the nice and the bad thing, I guess, about the UK—is that you never have to worry about share-based compensation. So the multiples are the multiples in YouGov, and at least you don’t have to worry about that. It’s definitely not expensive, which is also somewhat of an issue with some of those SaaS businesses.

Share-based compensation. It’s funny you say that because, again, I’ve written a ton about this before. You get these people and they’re like, “Hey, the company’s trading at 10 times free cash flow.” I’m like, “Yeah, but their adjusted EBITDA is 200 and stock comp is 300. So if you take the stock comp, it’s negative.” And you’re saying, “Oh, it doesn’t matter if the stock comp was issued at a higher price.”

I know one company I went and looked at. They just filed their 10-Q, and they had 40 million shares outstanding before the quarter. The stock was down 90%, and they issued 4 million shares intra-quarter in RSUs, right? So you’re saying, “Oh, the stock comp doesn’t matter.” Well, you own 10% less of the company than you did 90 days ago. And stock comp is going to start to matter real fast if something doesn’t change there.

Yeah.

Go, go, go, go, go.

Sorry, not to interrupt, but you mentioned something really important that I wanted to get to, and it’s a broader framework I’ve been thinking about—not because of AI, but I think, somewhat luckily, it actually applies to AI as well. You and I have kind of talked about this before, but increasingly, I think about the world as divided between content businesses and distribution businesses. Again, I’m a very simple man and not that smart, but I think about the world in those terms.

So when I think about content, I mean literal content. Something like YouGov, right? You own literal content; its data. Think about brands and IP. I even think, honestly, about brick-and-mortar experiential real estate. Something like a pub—that’s an experience, that’s kind of content you’re going for.

A distribution business is something where you’re the middleman, right? So you’re a department store, one of the payment processors, or a third-party marketplace. One of the things I’ve come to realize over time is that I have a greater appreciation for investing in content businesses because I think it makes long-term compounding easier.

I’ll explain why. I think the barriers to entry in a distribution business, for the most part, are scale. That’s generally how they work. It could be network effects to some extent, or complexity, but it’s generally scale, right? For content businesses, the moats—the barriers to entry—are more sustainable. Scale is somewhat easier to replicate, but the moats are deeper.

These businesses are less susceptible to disruption because, inevitably, you can move the content to a different form of distribution. You have more optionality, right? No one thought Disney would be a theme park business, but it’s a theme park business. There are a lot of things to do when you own content. There’s pricing power, and there’s simply greater runway for growth.

That’s the way I think about businesses today. A lot of the businesses you’re mentioning, I don’t know them, so I don’t want to speak out of turn. But the way I think about the world today is that I don’t want to own a business that may not be disrupted by AI or might be minimally disrupted by AI, because I feel like you’ll be fighting the terminal multiple.

Oh, I heard that.

It could be cheap, and there could be periods where it works out, but you’d just be fighting on this hamster wheel. I want to own businesses where the market thinks they are disrupted by AI, and I think they’re literally going to be worth more because of AI.

One great example is music, right? When we were teenagers, we were buying CDs. We were buying them from Sam Goody. In the UK, you’d be buying them from Virgin Records. The top music labels were Warner Music, Sony, and UMG, right?

Warner Music, I think, IPO’d at around $2 billion in 2005. It got taken out for $3 billion in 2011. It’s worth $15 billion today, even with some of the pullback. All 3 labels are the same. There’s been some consolidation, but we’re now consuming that music through Spotify, right? We’re not doing it through the record stores. Those have all gone by the wayside.

That’s your distribution versus content. It’s kind of the same with movies. The pushback would be Netflix. What’s interesting is that Netflix only became as valuable as it did when it started investing in its own content, anyway.

No, look, as you were saying that, I was thinking—as soon as you said it—because, again, I’ve written this before: in media, in the past 10 years, if you bought anything in media, you’re dead. You’re out of business. I know this because I had a big media and cable exposure. It’s a big drag on performance.

The same with the UK, by the way. Don’t worry.

The only things that have worked in media are Spotify and Netflix. Those are the only things. So you wonder: Did they get lucky? Yes, probably.

But with Netflix, I did have a little bit of a revelation. I’ve always been with you on the content, because who would have thought that, with AI coming along, all of a sudden everybody’s making Star Wars? Disney’s going to get paid on that in some way, shape, or form. Who would have thought that Star Wars would get turned into a theme park?

All this sort of stuff just keeps flowing. I think sports teams are great. Every year, the Knicks have a new sponsor on their patch. Who would have thought 4 years ago that there’d be an official prediction market of the New York Knicks, an official crypto partnership, and all of this sort of stuff? It just keeps expanding.

But in the past 10 years, every media stock has been slaughtered except for Netflix and Spotify. So you do wonder, as the world fractures and distribution gets easier. It is interesting to think about that. I mean, you don’t have any answers.

I think the music businesses have done very well up until the last year or so. What I think about is that the better example is Walmart versus retailers versus brands.

If you and I had simply owned Walmart, we would have been fine, right? Target would have been fine over very long periods of time. But there were a lot more retailers than there were brands, and there’s only 1 Walmart today after the internet, and there’s only 1 Target today. But there are a lot of brands, and there are a lot of very, very long-standing brands.

So, yeah, maybe it would have been better to simply buy Spotify rather than Warner Music, Sony, or UMG, but you had more options than Spotify. There were a bunch of different distribution businesses—Napster, for example. There were a ton of them. I used to download music off LimeWire. There were a bunch of different options where you could have gone wrong.

They have.

Yeah. What I think about is that the better example is Walmart versus retailers versus brands.

Pandora didn’t work. The other thing with UMG that’s interesting—and this actually might bring it back a little bit to YouGov—is: What’s the most valuable business of UMG?

One of the reasons I think UMG has done so well over the past 10 to 15 years—this is my supposition; I can’t claim I’m a huge expert—is that the back catalog has gotten increasingly valuable, right? The new music is so difficult, but the back catalog, with Spotify and all these things, has just absolutely exploded.

I think part of that is that distribution has gotten easier. They owned it, and people are listening to it a lot more. But I think part of that is also because the world has gotten so fractured. New media and new TV shows aren’t hitting as hard. Friends and The Office have much longer lives than they ever would have.

So I think that back catalog—and for YouGov, maybe there are brands out there for which that 20-year data becomes really valuable. I don’t know, though, because the other piece is that SpaceX is now the 5th most valuable company in the world, maybe the 3rd most valuable company as we’re speaking. SpaceX doesn’t have a 20-year history. So maybe—I don’t know. It’s just, yeah.

Yeah, look, and by the way, to be clear, YouGov is working with BYD. They have no data on them, right? BYD is moving into Europe. They want to better understand their perception. That’s something where they want to track and build up data over time.

The point is, you don’t need 20 years of data. It’s just that the more time you have it, the more valuable it is. The one thing that we kind of didn’t talk about was that if a customer were to turn it off, YouGov owns the data, right? So the data is only as valuable as how long you’ve been tracking it. The longer you have it, the better.

This is a long-winded way of saying that the way I think about AI, specifically with respect to YouGov but more broadly, is that I think YouGov will ultimately be more valuable. When we think about what we were talking about in terms of music, I think about music. I think I must spend more on music today.

And Warner Music or UMG must make more from me per piece of music than they did selling CDs, right? So, I worked in investor relations as an intern at Warner Music because I'm a big music fan, in 2005 and 2006. I can tell you that definitely wasn't the consensus view, right?

The biggest pushback would be that those businesses went through very difficult periods of time before they figured it out, right? As we talked about, Warner Music IPO'd at $2 billion. They were taken out in 2011 for $3 billion. It took a while for them to get there.

What I'd say is interesting and different about today is, to the point you made earlier, AI is moving really fast. We didn't have streaming that was really popular 1 minute after it came out, right? It took a while for it to come out. It took a while for the Apple Music store to come out and for everyone to figure all this out. YouGov already has commercialized AI products that they're working on, right? These things are just moving faster. So, yeah, there's been a hiccup, and I appreciate it's a super anti-consensus view. But I think, yeah, I'd stick by that.

All right. Well, I actually had a lot to talk about, but we're running long and I'm going to have to hop in a bit. So, let's wrap it up there. We'll have to have you back on because I think there's 7 UK stocks in your portfolio. We've only talked about 1. We've got 6 more to go. So, Jonathan from Zipline, this has been awesome, and we will grab a Sweetgreen soon.

Sounds great. Thanks so much, Andrew.

A quick disclaimer: Nothing on this podcast should be considered investment advice. Guests or the hosts may have positions in any of the stocks mentioned during this podcast. Please do your own work and consult a financial advisor. Thanks.