Pinterest, Inc. (PINS) Earnings Call Transcript & Summary

Okay. I think with that, we're going to get going on the next one. Okay. So it's my pleasure to have the team from Pinterest here for our second fireside chat of the day. With me is Bill Ready, CEO. I'm going to read a quick safe harbor, and then Bill and I are going to get into the back and forth into the dialogue. So some of the statements that Pinterest will make today may be considered forward-looking. These statements involve a number of risks and uncertainties that could cause actual results to differ materially. Any forward-looking statements that Pinterest makes are based on assumptions as of today, and Pinterest undertakes no obligation to update them. Please refer to Pinterest's latest Form 10-Q and Form 10-K for a discussion of the risk factors that may affect its results. Okay. So we've put the safe harbor behind us.

Bill, thanks, as always, for being part of the conference. You've always been great about making yourself available. So I really appreciate it. To level set, here we are. We're in the second half of 2025. I wanted to talk a little bit about your key strategic priorities for the business, how they've been progressing as you move through '25 and we will use that as a jumping off point for the conversation.

Yes. Excellent. Well, we did our Investor Day almost 2 years ago now and strategy has been consistent with that. And I think if you look back at that our execution has really aligned very directly with what we laid out there, both in terms of what we're doing and the results that it's delivering. So on what we're doing, our strategy to make Pinterest a shopping destination with visual search curation and AI fueled off that unique curation behavior has really played out. And then leveraging that to turn Pinterest into a true performance ad platform. And so the things that I'd point to with that 8 straight quarters of record high users. Gen Z is now more than 50% of the platform. So 3 years ago, when I came in as CEO, users were declining, the narratives of the Pinterest was aging up and aging out. This strategy has been so successful that 8 straight quarters of record high users. Gen Z is now the largest, fastest-growing demographic, over 50% of our users. We're growing across all geographies, all generations that we track. And at the core of that is the shopability of the platform. And the shopability of the platform is being driven by using AI, tuned on a unique curation behavior within Pinterest. It's truly unique in the Western world. in terms of users making -- we have nearly 600 million users on the platform and what they're doing is associating products all the time and associating products as to like what products go together, what fits their taste and style and that curation signal, this is the thing I saw from my prior seat at Google is like in this AI-driven world, if you have unique signal, you're going to able to do really unique things with the AI and that's really at the heart of why it's not just that we're winning with Gen Z. Again, we're growing across all the generations that we track. Shopping is at the core of that, driving that actionability, the performance ad platform also using AI to drive results for advertisers all that has hung together really well. And you see that delivering the solid mid- to high teens growth that we talked about at our Investor Day as well as helping us drive margin expansion cash flow generation while also being more relevant to our users than we have ever been with our best product market fit ever.

Got it. And in terms of just on the go forward, how should we think about how some of those priorities might change looking out over the next 12 to 18 months, if at all?

Well, a couple of things I'd say. So we see that strategy working exceptionally well. A good strategy has multiple phases to it. And I'd say, over the last 3 years, we've laid a lot of foundation on these things. But we've created a flywheel effect on multiple of these. We also talked about this at our Investor Day as well, that the ads can be great content when we have shopping behavior when the users in a commercial context, they don't care whether it's organic versus ads as long as you show them the right product. We're making that flywheel spin, but we're also using the AI to drive better and better visual search capability. So the way this manifests for users, if you ask Gen Z why they're coming to Pinterest, One of the first things they'll say is, well, Pinterest just gets me. Well, what's behind Pinterest just gets me. It's using the AI to make really, really relevant shopping recommendations. And now we're bringing that -- we started laying foundation with AI in the background around recommendations. Some of the things you're starting to see us doing this gets more to your question of like what you can expect going forward. We're bringing more and more of that to the foreground now, right? So with our visual search that we talked about. Visual search. Visual search has sort of always been at the core of Pinterest, but the visual search keeps getting better and better. I shared that our latest multimodal visual search models that are proprietary and in-house trained on our unique signal are outperforming off-the-shelf models by 34 percentage points on the relevancy of their shopping recommendations and so that's the tech behind it, but that's letting us do great new experiences directly to the user around how they can search individual elements of an image, how they can curate more and more of the individual elements of an image. And starting to bring things to the user that also give them more language when they didn't have it. What we're really competing for is visual search. And in that world, what we're solving is the -- I'll know it when I see a problem. It's very different than the way sort of chat bots and general purpose search pursues that. We're solving the I'll know when I see a problem, and that is working quite well. In fact, so well. Adobe put out a study, it's not something we sponsored and independent of us. But that's 39% of Gen Z starts their searches on Pinterest. So how well is that shift to visual search and using the AI going, 39% of Gen Z is saying they start their searches on Pinterest. So expect more and more of us bringing purely visual experiences, AI forward, helping people solve more and more of their commercial journeys with a great visual search experience with AI at the foreground.

So maybe just 1 quick follow-up on that. So with visual search at the core, when you think about -- and you actually have a background in this from your time at Google, how do you think about the shifting competitive landscape for search and positioning Pinterest visual search against your world view for that competitive landscape.

It's a great question. So step back a little bit, even before you got to the sort of latest round of AI, search has been fragmenting for years, right? Google continues to be fantastic for general purpose search. But over the last 15 years, you've had lots of even as they were growing very nicely, search fragmenting and competition increasing. How many products searches start on Amazon versus Google? How many travel searches start on a Booking or an Expedia rather than Google. And so there's been this sort of fragmentation and verticalization of search, it's been a decade-plus long trend. My view with Pinterest was not that we would go solve general purpose search and compete for anything that you might search for, but that around purely visual search, particularly shopping at the center that was an ownable space. And particularly, one of the things I've talked about before that the first 25 years of e-commerce sort of solved buying, but killed shopping. The distinction being that the utilitarian part of the journey had been solved. If I knew what I want -- if I knew what I wanted, there were tools, Google, Amazon or the things that would help me get it the cheapest and the fastest. But if I didn't yet know what I wanted, if I didn't have the words to express what I was looking for, there weren't great tools for that. And if you look at the 75% or so of commerce that still exists outside the digital world, that's a lot of what people are doing is well, "Hey, I need to go look, I need to go sort of walk through a store, I need to go walk to bazaar, how do you solve that?" Well, that's an entirely new space in my mind. And I think you're seeing that that's -- we're being quite effective at solving for that. And it's a rapidly growing pie. Even as search fragments, the pie is expanding because there's a lot of the experience that hadn't yet digitized, so we're going after new experiences that hadn't digitized, I'll know it when I see a problem and bringing that into the digital world. And that's -- and I think that's a unique ownable space and you're seeing that play out for us that even as you have lots of others that have introduced new experiences, chat bots and things like that are -- were 8 straight quarters of record high users and Pinterest is a platform being Gen-Z, they see us as a -- 39% see us as the first place to search. That's a great demonstration of the fact that they see us, our users see us as a unique experience separate from what's happening elsewhere and both can grow simultaneously in our.

Got it. So building on that Gen Z point, just in terms of the engagement levels you're seeing from Gen Z, how do you think about sort of building towards sustaining and building further momentum on top of that engagement on top of that user growth, what do you think about your product road map and sort of sustaining that in the years ahead?

Yes. So when I think -- I think about share of wallet on 2 perspectives, share of wallet with our user, our consumer and then share of wallet with our advertisers. And I'd say on both, where we are 3 years in, we have made tremendous progress improving Pinterest as a shopping destination, right? And we are a destination. 85% of our users come to our mobile app directly, 100% of our users are logged in. We are a destination. We are not nearly SEO dependent as others. We are a destination. But how do we continue to differentiate that? We see that we are still, even as we have made tremendous progress, a relatively small portion of the overall share of wallet. And we've talked about on some of our earnings calls how shopping and retail has been a strength, but we see financial services that is adjacent to that becoming a strength. We talked about other emerging categories like autos and things like that, where -- or travel where our visual first experiences apply to a lot of other commercial journey. So I think that's opportunity for us to continue expanding share of wallet with our consumer as well as share of wallet with our advertiser where we've talked about how we've started to break in more and more to the always-on performance budgets of the advertisers, particularly with the largest, most sophisticated. But even there, we continue to gain more of the catalog, deepen share of wallet there. Even as we go into more and more advertisers, we talked about that sort of $1 billion to $30 billion GMV group is our next segment. How we're starting to pick up traction there and then SMB after that. Well, those are all new buying experiences to come on to the platform that deepen the share of wallet with the advertiser but also give us more great products to show users that make it so that users see us as a great place to go for more and more of the things they're shopping for.

Got it. And in terms of continuing to build momentum there, how do you think about what's in your control in terms of product road map on the shopping side in the years ahead versus some of the threats or challenges people talk about when they talk about maybe a shift towards Agentic shopping over time. How do you think about balancing opportunities and the challenges.

Well, what I'd say the things that are directly in our control, we feel really great about the relevancy of our recommendations. We talked about we've more than doubled the relevancy of our search results over the last couple of years, our taste graph growing 75% so all the things within our control, we feel really great about. We also feel really great about the way that we're managing the ecosystem. So when you talk about new experiences like an Agentic, you really need to have thoughtful management of the ecosystem. And if you step back from it, we don't call it Agentic because that's not how our users think about it. But when you just say things like Pinterest just gets me or I started a bit of my journey, but then I came back 2 days later and Pinterest had all these great recommendations for other things I needed to buy to complete or help me finish the journey. What is Agentic? It's helping users complete journeys. We're taking users in a completely automated way through more and more of the journey using AI. And so we feel like we are very well positioned around sort of the broader notion of Agentic experiences where people will look for the AI to take them further and deeper through journeys. At the same time, I think we're striking a good balance in the ecosystem management where -- and I've got past experience on this, where a lot of the sort of what the tech platforms would like to do around completing these experiences. Like Agentic buying is not really a new thing. There's a thing called Google Duplex. If you remember that product, it was out for many years. And with a lot of those retailer participation is going to really matter. Understanding the user and what the user really wants matters. Retailers are not going to be happy about being relegated to dumb pipes and being disintermediated from the consumer. So when you look at the experiences that we're creating that create really seamless buying, we've done things like with Amazon, where you can do a one-tap buy right inside our platform but it's clearly branded as Amazon. You don't even have to use our -- leave our platform to do it. You can just tap buy and the purchase just happens, but it's clearly Amazon-branded, they're clearly still getting a customer not just a transaction. So I think we've been laying a lot of the right foundation for the retailers and the ecosystem to feel like we are not making this intermediation play. We're actually still bringing them a customer. And from a user perspective, keeping them in the loop, keeping them in control, but then taking them much deeper in these journeys and then they can be in the loop when they're ready to complete. And I think that's actually what we're going to see a lot more of for the next couple of years versus telling the agent just go buy everything for me, let me know how it works out. And that sort of litmus test I always give people to sort of conceptualize this. Think about the person that knows you best in your life, whoever that is, spouse or whoever, would you let that person pack your suitcase for a 2-week trip without you looking at anything that's in the suitcase? I have yet to meet a person that said yes to that question. And so the bar for what it takes to allow an agent to like fully complete your shopping journey, that bar isn't just be better than a human. It's be better than the human that knows you best in your life. So I think people are going to want to be in the loop for a while. But I do think and we're seeing it on our platform, back like we basically created an AI-enabled shopping assistant already without calling it that is that people are very happy when you take them right up to the end of that journey and say, yes, that's the thing. That's it. And so if you lay all the stuff out and said, "Hey, I think I have got your suitcase, Is this right?" Yes, yes, yes. No. Yes, that was very helpful. Just back the whole thing, and I don't need to look and it just shows up on my doorstep. I don't think people are there yet someday, but I think that's not what's next. But if we look at what I think people are ready for in the here and now, we feel really great about how we're progressing that and how we're balancing the -- making sure it's great value for the retailers as well.

Understood. Okay. I like that test. I'm not going to ask my wife that. Switching to the advertising environment. What are the key messages you guys are receiving as a company broadly from advertisers about the operating environment today. And how is Pinterest roll in that environment, whether measured by brand spend or performance marketing spend continuing to evolve?

Well, we have really proven out Pinterest as a performance ads platform. I think when you look at not just where we are but the progress we've made in 3 years' time, Three years ago, Pinterest was almost entirely upper funnel brand ads, right? There were very few clicks and conversions happening on the platform. Now we are primarily a performance ad platform. So we've talked about strength with the largest, most sophisticated retailers out there. For those advertisers, we see 90% plus of their spend with us now is performance driven, right? And so we have a very -- and I think it's a pretty unique thing. You haven't had somebody really break into performance budgets, particularly search performance budgets in a long time and we're carving out a space for ourself there. Now we're still as much progress as we made, I'd say still pretty early innings, which means there's a lot of runway ahead but I think we have very much proven out a role for ourselves because of the unique nature of our consumer shopping experience, right? So we've built up the basic components of a performance ad platform certainly not to the degree that the very largest have done because they've been at it for 20-plus years. But in the last few years, we've stood up the basic components of a performance ad platform and then use that to tap into this completely unique shopping experience that happens on Pinterest, where users curate and then now increasingly take action, that's advertisers, particularly those are engaging in shopping and things that are shopping adjacent are seeing that as a really unique opportunity. We're able to drive performance increasingly through AI-enabled tools like Pinterest Performance Plus that's making it so that we cut their campaign creation time in half make it easier for them to come on board, those things are working well. And then I think some of what was embedded in your question also was just you talked about the operating environment, sort of the macro and those things. And my comments here will be very consistent with what Julia and I talked about on our most recent earnings call, which is, I think, the macro, particularly for advertising is more constructive than it was at sort of peak tariff uncertainty. At the same time, there are puts and takes in the market, right? And I think even since our earnings call, you've seen other earnings calls happen since then where you see some of these puts and takes playing out? Do you see some of the very largest retailers talking about margin pressure from cost of tariffs and things like that. But then you also hear those that cover small businesses, like the Shopifys of the world, talking about how well the small businesses are able to be dynamic and ship supply chains or raise prices to cover these things. And so you see how different parts of the market are handling that differently. So we played through those puts and takes previously. And I think 1 of the things that we've continued to demonstrate back to that consistent mid- to high teens growth is that we laid out at our Investor Day multiple ways to win. And so when we did face a bit of tariff uncertainty last quarter, you saw us play through that, I think, quite well coming in ahead of top of our guidance ahead of expectations or at least the analyst expectations anyway, it doesn't mean everybody's expectations. And a big part of that was because we have multiple levers there, where even with a little bit of pressure in the U.S. from tariffs, we saw advertisers redistributing spend to international and international really picking up steam for us. And so I'd say the macro, again, very consistent with what we talked about at our last earnings, but you've now since seen others report earnings, and you see some of those puts and takes around there are places where there are pockets of strength like small business, which is a newer but growing exposure for us in a very positive way. But also we have some retailers feeling margin pressure. And so that means we're going to demand more performance than ever. And good news, we've been very focused on delivering performance for them.

Got it. And that's the externality that's the stuff that's sort of out of your control. When you pivot back the conversation to the stuff in your control. Talk a little bit about some of these mechanisms for monetization you're building for the long term. You talked there about Performance Plus. We've written positively about that from our own advertising industry conversations. But talk a little bit about the building blocks you're putting in place to continue to drive Pinterest as a platform towards more always on, more performance marketing dollars, more lower funnel conversion over time.

Yes. Well, maybe I'll start before I come to the ad side of it. The hardest thing in these is always getting the consumer behavior. That's always the hardest getting consumers to shop and engage and purchase that's the hardest thing to do. If you've got that, then it's sort of basic mechanics like how do you help advertisers tap into that. So I talked about the 8 straight quarters of record high users. You've also seen from us, I think this was maybe overlooked a bit in the last quarter, like in U-CAN, our largest revenue market we saw the fastest user growth, year-on-year user growth in U-CAN that we've seen since Q1 of 2021. And we have seasonality in our business, which is why sometimes people miss that because of sequential, but you've got to look at the year-on-year so the seasonality doesn't mislead you. And year-on-year growth for U-CAN users was our fastest growth rate since 2021. So acceleration there even after 8 straight quarters of record high users overall. So we're getting really great user engagement. And then further to that, how are we getting the engagement. We've talked about even as we're hitting record high users, we are deepening engagement per user. The way that we're doing that is with search and actionability at the core. So if you think about like query growth on our platform, even traditional queries, query growth is growing faster than user growth. Then within that, we're getting more and more people to visual search type experiences. So our visual search engagement, particularly when denominated by like the amount of clicks that we drive, right, because you want to drive actionability while you've got search queries with traditional search queries will be growing faster than users. The visual search engagement driven by total clicks growing even faster than that. And then ad clicks growing even faster than that. And so we've talked about that dynamic before, like the deep engagement and the acceleration or the higher growth rates of the types of engagement that we want and those things being very commercial. That all is playing out very, very nicely, including our newest experiences like visual search and specifically my comments there around like related items, right, which is a purely visual search type experience on our platform that we continue to advance. Then on the advertiser side, what we've been doing is make it easier and easier for them to tap into that, right? So I sort of think of this is like 3 legs of the stool that, first, we needed to go drive clicks and conversions to the advertisers. So a couple of years -- actually, roughly 2 years ago, we launched mobile deep links and direct links. That sent much more traffic to the advertisers. Then we came behind that and provided measurement solutions that made it so the advertisers could measure that, and that was really sort of start of last year. We drove a lot of adoption through last year. And then at the end of last year, we launched Performance Plus, which was a third leg of the stool, which is the sort of automation suite to go make campaign creation instead of really seamless and easy cut campaign creation time in half. So now we're seeing all that really, really working and continuing to drive forward. And so we're taking that from the largest advertisers into the mid-market and small business, mid-market, we define is more like $1 billion to $30 billion in GMV. That's been a multi-quarter phenomenon of deepening engagement there. SMBs, we're starting to see really good early signs with SMBs now the Performance Plus is out there. And international, we were taking that shopping playbook to the international markets and you've seen the denominators there are still small relative to U-CAN for us. But it's a huge opportunity. We have 80% plus of our users outside the U.S., but only roughly 20% of our revenue addressing that imbalance is a huge opportunity, and this is sort of the sequence of events we expected that we needed to solve shopping in our largest home market first. But now we're deploying that to international, and you see that in the very nice growth rates we're putting up on international. So those are ways that just broad strokes, how I think about that continuing to play forward and the AI at the center of that. I think we're just -- all these things, I've said repeatedly like no hockey sticks, these are multi-quarter, multiyear. The shift to these AI-driven experiences in the kind of efficiencies that can be driven on this. This is a -- for the industry. This is a multiyear phenomenon and there's still a huge amount of sort of pie expansion to happen. When you think about just how many ad dollars are still spent in inefficient places where you don't get great performance. All advertising dollars are eventually going to be performance advertising dollars. And I think the unique shopping behavior that we have our ability to help advertisers tap into that in a seamless way that's driving performance for them, I think, is carving out a nice place for ourselves in that, and there's multiple years of that runway ahead.

Maybe following up on that and maybe first starting with just partnerships, but it dovetails into international. How do you think about some of these efforts scaling and how much of it is elements of partnerships you've struck as much as its elements of owned and operated and building and scaling yourself. I think international is a good example, maybe to double-click on that.

Absolutely, yes. And I think it's also -- as we've done all these things, as we've built this performance ad platform doubled down on AI, really completely revitalized the platform. We've also delivered great margin expansion. So we're making this profitable growth, great cash flow generation and this gets to your question around how we use partnerships. We're being really thoughtful about how we make investments that are high ROI. On the AI side, that is like we build in-house for things that are truly unique to us but then we can leverage things off the shelf and tune them on our AI to get even better results. That makes us really efficient with how we deploy AI dollars. On sales, in our largest markets, we -- first party sales is absolutely a focus for us. When we talked about expanding, we said, okay, we can work with partners to go round out gaps in the auction in smaller international markets where it may be less efficient for us to go fully do first-party sales, we can use partnerships to enter those, whether those be third-party ads, whether those are resellers, whether those are agencies and we use a composite of those things and sort of the balance between those shifts where larger markets, we tend to be much more first-party. Other places where it would be less efficient for us to do fully first party. We'll mix in more partnerships to make us very sort of capital efficient on those things. And we see that working out really well. And I think when you survey those partners, particularly around agencies that given Pinterest didn't have much in the way of performance advertising 3 years ago, our agency partnerships are still new and evolving, but we see really, really good strength there, and to be very clear, unlike some others in the space who think that AI is just going to replace the agencies like we think the agencies are going to have a real meaningful role to play for some time, and we're very focused on how we partner closely with them, bring them into the equation. And I think there's channel checks and things out there that talk about that where you hear from those agencies, I think the ones that are working with us are seeing more and more that we're delivering great performance and want to lean in there. So the partnership's aspect, partnerships with large, whether that's agencies, resellers, third-party, we think those are doing exactly what we hope to do, which is helping us be really efficient going into new markets or new areas of the auction where we needed to round out gaps. And we're demonstrating performance for them, making them part of the equation in a way that they feel good about as well.

Maybe just 1 quick follow-up. I think investors generally understand the opportunity when they see the base of users internationally versus the current level of monetization. Talk a little bit about what are the unique challenges internationally that maybe you don't face in North America that just sort of make that something that could play out over years and that people just need to maybe have some framing around like duration of execution.

I mean, I think all these things play out over years, even North America is we're 3 years in, and we have many years to go, and we continue to gain share of wallet with our users back to like deepening engagement per user, even as we grow users. But international, there are 2 things I'd say. One, the AI-driven experiences, the shopping experiences that we're driving, those are working at a global level. You see that in our growth rates. If you look at our -- I talked about how we're quite happy about how our U-CAN user growth rate was the highest it's been since Q1 of 2021. Well, our international growth rates on users is even higher. So the things we're doing around visual search, AI-driven recommendations, that is working at a global level. But then to your question, like the multiyear trend or the multiyear duration on these things, there are some things that you need to do to solve 4 specific markets where shopping may happen slightly differently in some markets versus the next. I've solved that multiple times in past lives. These are very solvable. I just think about it is like you want to be sort of 95% global and then solve that 5% last mile or last kilometer in those international markets. And so that takes time, but we're actually seeing the generalizable part of that play out really well and it's reflected in our growth rates. And again, that's a general phenomenon I've talked about of query growth being faster than user growth and then the actionability and the clicks being faster than that, particularly around the monetizable clicks. That general phenomenon is playing out globally for us. So I think the generalizable part of that is holding globally. We don't have to do a totally different product market by market, but there are sort of last kilometer of things that you need to do in some of these markets. And so that's where it played out over multiple years. But you can see reflected in our revenue growth rates in those international markets that we are picking up steam there.

Okay. I know we only have a few minutes left. So I wanted to end on sort of a bigger picture question that maybe ties some of this together. So we've talked a lot about AI development. Talk to us a little bit about how your priorities around AI development, most anchor around where you want to take visual search for the medium to long term. And where you want to take visual search, how does that feed into dynamics around competitive advantages for the platform? How does the data that Pinterest have feed into that broader dynamic when you think out over the next sort of 3 to 5 years?

Yes. This is the most differentiated thing, this is -- I've talked about from the day I joined. The curation that happens on Pinterest is truly a unique signal to feed the AI. It's something in past life is like you say, "Hey, we know everything that you buy, right?" But while there are other players that may know everything you bought, everything in your closet, they generally have no idea how you style it into an outfit. How do you pair those products together, right? And I'm not talking like the easy product associations like, "Oh, you just bought your first bag of dog food like you're going to need a whole bunch of other dog accessories like that's super easy." Which handbag looks good with which outfits and which one is actually aligned to you. And when we talk about our taste graph having grown 75% over the last 2 years, that is hundreds of millions of people on our platform every day, making these product associations where that lets them -- it's not just that we can personalize better for that user. The next user can come in with just a very small starting point of maybe like I like that handbag. And we will know how many, many other people styled an outfit around that handbag. But even better than that, we know the intersection between how other people style that handbag and those that have tastes similar to your own style that handbag and then we can make a recommendation, then lets the users back to so many of our users say, "Well, Pinterest just gets me, right?" We're not ready to replace the most trusted person in your life yet for what you're doing with shopping, but we're going after that kind of joyful experience, right? Not utilitarian part of the people that want to get away from shopping. It's like, send a bot off to go do this for me, so I don't have to worry about it, the utilitarian journeys, make sure I don't run out of milk and eggs and detergent, bots may be good for that. But the stuff that is -- there's a joyful aspect of shopping for so much of what is shopping where people want to be involved in what they're choosing, they just want help and assistance. That is going really well. And it is that unique curation signal that lets us tune the AI in ways others can't. So again, like our latest multimodal visual search model, which is proprietary and built in-house on our unique signal outperforms the most popular off-the-shelf models by over 34 percentage points on the relevancy of the shopping recommendations. And that is, one, we have a very focused effort. We're not trying to win generalizable sort of anything you can ask the model, but focus on shopping recommendations, we're able to do that when we have some great AI engineers, but also we have really unique signal around those product associations. And I've shared this in other examples where I've said even when we take the off-the-shelf models and retrain them when we're using them for things that we don't want to build ourselves, even in cases where we say, we're going to take something off the shelf to be cost effective, but we're going to retrain it on our signal. I've shared that we'd see 300 basis points of lift just from training of our unique signal, which gets to that curation behavior, which, again, at least from my vantage point, is totally unique in the Western world, and I think is at the core of our advantage here and our ability to do things that are truly unique. And so again, that flywheel spinning, users are curating more and more, gives us more signal, lets us make better recommendations, brings more users to the platform, gains more share of wallet, spins the flywheel but even more unique signals to train the AI, even more ability to bring advertisers in that, we see that flywheel spinning. And that was very much the theory of the case 3 years ago. A lot needed to go right for that to work. When I look back over the last 3 years, like it has gone exceptionally well. There's always use I'd like to grow this part a little faster, that part a little faster. But overall, that has gone exceptionally well and that unique curation signal, I think, is something that people don't still -- I think a lot of them still don't fully appreciate just how important that is. It's -- in the Western world, I don't have another experience out there that has that signal. And that's why you see us really gaining share in search back to 39% of Gen Zs as Pinterest is a first place to go search. And now they're more than half the platform, that's the best evidence I can give you of just how effective that has been even as we still have a ton more to build.

Okay. Well, Bill, I always appreciate the opportunity to have a conversation. Thanks so much for the conference. Please join me and thank you Pinterest for being part of the conference this year.

Thank you, Eric.