Pinterest, Inc. (NYSE:PINS) Goldman Sachs Communacopia + Technology Conference September 11, 2024 11:10 AM ET

Company Participants

Bill Ready - CEO

Conference Call Participants

Eric Sheridan - Goldman Sachs

Eric Sheridan

All right. I think we're going to get started. Let me just make sure everything's working technology-wise, yes, okay. The mic is working. So it's my pleasure for our first fireside chat for me for the day to have the team from Pinterest here as part of the conference. We've got Bill Ready, CEO.

I'm going to read a safe harbor and then Bill and I are going to have a conversation. 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 most recent Form 10-Q and 10-K for a discussion of the risk factors that may impact actual results. Pinterest may also discuss non-GAAP financial measures today.

Refer to Pinterest's earnings press release dated July 30, 2024, which you can find on their Investor Relations website, for important information about Pinterest's non-GAAP measures, including a reconciliation of non-GAAP measures to the most directly comparable GAAP measures as required.

Question-and-Answer Session

Q - Eric Sheridan

Okay. So the safe harbor is behind us. July 30 does seem like a long time ago, but that was earnings, I guess. It's been a very long summer at least in my world. Bill, thank you for doing this. I really appreciate the opportunity to have a conversation. You had an investor event last year. We've been going through 2024. To level set for the audience, why don't you walk us through your key strategic priorities for Pinterest in the coming years?

Bill Ready

Yes. Excellent. Yes, it's been almost coming upon right on one year since we did our first Investor Day. And maybe just to lead it off, I'd say, the initiatives we lined out there, the long-term plans that we lined out there, we see each of those key initiatives, tracking as well or better than expected from what our expectations were at that time and how that stacked up to those three to five-year plans that we laid out. So we feel really great about that.

And if you sort of dive into that, I'll start with Users and Engagement and things that we're doing there, where you've seen now, eight straight quarters of user growth from us. We're putting up record levels of new users and deepening engagement per user. So our efforts to go drive engagement with users, which have really been centered on making the platform shoppable, driving actionability. Pinterest has been a place for window shopping previously. We've now made the place where not only new people window shop.

We've opened the stores. People are clicking, converting, buying. And that's not just good for the revenue side, which I'll come back and talk about, it's great for the engagement side. We see that one of the key drivers of engagement has been around the actionability that we brought. But then as we drive more actionability and drive more curation behaviors.

That's the second thing. So the actionability then there's a curation that is completely unique to our platform, where we really leaned into helping users put together outfits, put together rooms, getting more granular that they can grab each individual item, that just gives us a really unique signal that doesn't exist anywhere else in the ecosystem. In my past life, I'd love to have that signal. It's one of the things that from the outside in, I found really compelling about Pinterest. As we're leaning into that curation, that's driving engagement.

But then that extra actionability, that shopping behavior and that curation behavior is completely unique to Pinterest, that's giving us a really unique signal to feed our AI. And then we're seeing huge improvements in the relevancy of our recommendations. And so on the engagement side, those have been the key points around make it shoppable and actionable, use curation to get unique signal, feed that to the AI for relevant recommendations, that flywheel of spinning.

And the more we drive that actionability, the stronger our signal gets, the better our recommendations get and there's a great virtuous cycle happening there with our users, and it's very much at the core of why we're seeing that deep engagement per user that's been a central part of the story. On the monetization side, we talked about how we wanted to demonstrate that because most of our users are on the platform to shop, when the user is in a commercial context, ads can be great content, right?

So if the user's shopping, as long as you show them that right pair of shoes or that right outfit, ad can be great content. And so that's been a central focus for us. We've really proven that out, driving up ad load while simultaneously increasing engagement per user, really hard thing to do, but that's proving that we're making the ads great content. So we see that tracking well. And then expanding that monetization into new markets with resellers, third-party bringing on more demand, all those things we see tracking really well.

And you already mentioned, all that comes together with our lower funnel efforts, where we've been citing that we have real strength in our lower funnel offerings, which again tie back to that shopability for these users.

So all those things that we laid out there continue to track as well as expected or better. And those are the key things. And again, all those things are -- they come together really well because they complement each other exactly as we laid out then, and we see that continuing to not just take hold, but to that flywheel is really spinning.

Eric Sheridan

Okay. All right. Really interesting stuff, and there's a lot that we're going to follow up on that and maybe mine from that as a jumping off point. But first, I think away from AI, the number two theme at the debate so far has been the broader macro environment. So what are you hearing from advertisers in terms of recent performance and the outlook for the second half of '24? And maybe how do you view the competitive landscape in terms of advertiser looking across places where they can allocate budget?

Bill Ready

Yes. So we view the macro consistently with our comments on our last earnings call. And we said there that we see a relatively stable macro. We see some puts and takes in that, but a relatively stable macro. So consistent with our prior comments on the last call.

If you decompose that a bit, we talked about several calls in a row now that our strength in retail -- our strength in retail is really driven by the significant improvement in our lower funnel offerings and the performance that we're driving there.

So we've seen that our lowest funnel conversion objectives, those things continue to perform exceptionally well for our advertisers as a source of strength for us. So that's partly macro in terms of where retailers want to spend. They want to spend in that lower funnel. Great news is we've got a great product there that's performing well.

And so that's partly about us and how we are moving into that space. We talked about how for some of those largest, most sophisticated advertisers, we're getting to 5%, 10% of spend, getting into performance budgets instead of the sort of experimental budgets. So all that sort of addresses how we're competing, to the second part of your question.

And what we also talked about is how, as we have -- really over the last two years or so, we have really transformed Pinterest from what was previously mostly an upper funnel platform to a true full funnel with actually a really great strength in the lower funnel. More than two-thirds of our revenue now comes from lower funnel.

And we've talked about how we started with the largest, most sophisticated advertisers, and we've been working our way down. And we're seeing that progress as well, that is largest, most sophisticated, I think of those as $30 billion to $100 billion in revenue or more, performing really well there. And then we've talked about how we're driving down that adoption curve as they get to the next group. And we mentioned on the last call, we're starting to get into that next group, that these aren't small businesses. These are sort of $1 billion to $30 billion in revenue.

We're seeing those groups really start to take advantage of our -- those retailers start to take advantage of our lower funnel offerings. And so again, that have us feeling good about performance on retail and what we can drive there with the focus on lower funnel. And then on the flip side of that, we called out were there some headwinds around food and beverage, for example, which we have a little bit more exposure to that than some other platforms just given the use cases on Pinterest.

But I'd also call out very clearly, and we did this on a call, but just to make sure it's clear, that wasn't a commentary on broader CPG. It's really about food and beverage because even within CPG, there's other places where we see bright spots, around beauty and household goods and things like that, that continue to be areas of strength.

And so those are sort of some of the broad contours. But again, very consistent with our comments on the last call that we see the macro is relatively stable with some of those puts and takes and again, the strength in retail, the move to lower funnel, that's been at the core of our strategy, and we see that performing really well.

Eric Sheridan

Okay. Very clear. You talked earlier in your answer on strategic priorities about engagement. Maybe talk a little bit deeper about where you've been taking engagement on the platform? At what point does increasing ad load have a negative effect on users? You talked about ads as content in your first answer. And this is a question we get a lot from investors. So talk a little bit about your efforts around engagement and then also tying content back to engagement as content.

Bill Ready

Yes. Yes. So this has been -- I think one of the most exciting things about the business over the last two years is that not only have we demonstrated that users will shop and buy and click and convert. It's actually not just good for advertisers. It is good for engagement.

So we've talked about even as we're putting on record numbers of new users, hitting all-time highs on users, our engagement per user continues to deepen. And that's demonstrating that we are making the ads great content. We think we have a lot of runway to go in terms of where ad load can go.

It's really driven by the commerciality of the user's intent and the relevancy of our recommendations. But the real proof in the pudding as to how well we're doing with that is the fact that we're able to drive deeper engagement per user, even as we take ad load up and where you can see that, and we called out on the last call, is our weekly active user ratio to monthly active user ratio.

And that continues to improve, where we're seeing more weekly actives relative to monthly actives, which means as we drive more of that shopability and actionability and that curation behavior, the relevant recommendations, all of that is making the platform more compelling for our users. And so I said when I first came in, that Pinterest wasn't just about -- particularly in our mature markets, about how many new users we could see.

It was about how do you take those users, where we have seen most of the shoppers as say, in the U.S., but how do you move them from episodic usage to regular usage? Well, in that WAU to MAU ratio, the weekly active to monthly active ratio, as more of those monthly actives are becoming weekly actives, that really says that we're making great progress there. And at the core of that deepening engagement is the actionability and the relevancy of the recommendations, and that ties through to the ads as a great content.

So we've seen that be quite complementary, and we think there's a lot of runway to go. And again, it's really driven by the fact that more than half of our users are on the platform to shop, they have commercial intent. And the ceiling is really driven by the relevancy of our recommendations, and we're seeing it more and more. Our users see Pinterest as a place to shop and to take action, including with GenZ, where Pinterest where GenZ goes to shop now, right? And then so we see that working really well.

Eric Sheridan

Well, following up on that last point and bringing it back to your first answer. I want to talk a little bit about lower funnel drivers in the business. Can you talk about the success you've had with some of the lower funnel advertising tools? So I'm talking there about mobile deep linking, direct links, you've recently announced Performance Plus. And how should investors think about the timeline for when positive results on clicks and conversions start to translate an increased budget allocation from those efforts?

Bill Ready

Yes, yes, great question. So the first thing I'd say is like we're really pleased with how those efforts have gone. Again, two years ago, Pinterest was primarily an upper funnel platform. And we've -- really from almost a standing start a couple of years ago, stood up a great lower funnel for our advertisers. I think we share the urgency of how fast we want to move on those things.

When you sort of step all the way back. Mobile Deep Link, which was our first product that really was driving click and conversion for advertisers, that only went GA just over a year ago. That was like middle of '23 that, that went GA. And that was really just for the very largest advertisers because it was a deep link into native mobile apps. Well, it's really just the largest retailers that have big penetration on their native mobile apps.

We saw that working really well, sort of middle of last year into Q3. Then we said, okay, how do we take it to the next group of advertisers. And that was direct links where we said, "Oh, well, now this can work for even those that don't have a mobile app. In parallel to that, we launched Conversion API, where if you're an upper funnel business, you don't need Conversion API because you're not driving conversion. So -- and you don't need lower funnel measurement.

So we launched measurement and we started driving real adoption around measurement. And direct links, to keep in mind like direct links didn't get the full rollout until Q1 of this year. So we're still quite early on, even though we've seen tremendous momentum. And we're driving through that adoption curve. And each one of those things sort of adds to our penetration there, both the penetration of the relevant inventory we can bring on to the platform, so that the user is seeing the right shoppable inventory from all the retailers they want to buy from.

And so of course, we started with the largest ones that could bring the most inventory, but we're expanding that now from those 30 billion-plus retailers to the 1 billion -- to 30 billion retailers, seeing traction there. And that really is driven by the compounding effect of being able to drive clicks and conversions. We've talked about three consecutive quarters in a row, more than doubling the number of clicks we sent to advertisers year-on-year, really driven by mobile deep links and direct linking.

We've talked about driving adoption of our measurement tools -- our privacy safe measurement tools, that lets the advertiser measure it. And then on this most recent call, we talked about the launch of Performance Plus, which is our AI-driven suite that makes it so that, that next group of advertisers down that has sophistication, but they need these things to be simple.

It makes it so that -- we've driven the performance, the clicks, the conversions, the measurability. And now we're driving easy campaign creation setup, right? So with Performance Plus things like automated bidding, budget allocation, targeting, the dynamic creative optimization where we will help optimize their creative. It's still in beta. These things have a multi-core adoption cycle, but we're seeing really great results from that.

And on the last call, I called out advertisers like Poshmark and Timberland, which would be in that sort of next group down, right, as we're driving further down in the adoption cycle and how they were seeing great results in our beta. We've also had that broadening continue with retailers like Tractor Supply Company that, from our dynamic creative optimization, has seen their return on ad spend, double for the ads using our dynamic creative optimization, where we can take our unique signal of what's relevant to the user and help optimize the creative to get better performance for the advertiser, but also it makes a more visually compelling ad for the users. So again, the ads are great content. So we see all that coming together. But if you just step all the way back from it.

We are well along our way in terms of proving out the theory of the case, right, that we have broken into performance budgets, we're getting 5%, 10% of total spending over 1% of them across the market.

But for some of those largest, most sophisticated, we're getting 5%, 10% of budget from them as performance budgets inside of experimental budgets. Now we're taking that to the next group of retailers down. And all these things, the mobile deep linking, direct links, the measurement and now Performance Plus, these all have a compounding effect where they drive us deeper and deeper into that adoption. And while we got very clear proof points, really great results, we're still early innings in terms of the yield that we can get on that.

And even for the very largest, we're still getting more yield. We're opening up new categories and those kinds of things, where they're bringing more inventory on. So we're still seeing great yield from the largest ones even as we're now getting to those next tiers of retailers. And so we see a lot of adoption cycle in front of us even as we have really great proof points to say, we're a real player in performance now.

Eric Sheridan

Okay. You've announced a number of third-party partnerships over the last 12, 18 months. Maybe pivoting to that as a topic, what are your key learnings across those partnerships? What inning are you in, in terms of scaling and revenue contribution? And how do you see those partnerships evolving over time?

Bill Ready

Yes. So we announced our first third-party partnership with Amazon, a little over a year ago sort of early 2023. So a little over a year into that. And it's performing exactly as we expected. We said when we announced it that we wanted third-party demand to round out gaps in our auction and bring more shoppable relevant inventory for our users, and we see it doing exactly that.

And we've seen nice steady progression of that. We announced our first international partnership with Google earlier this year. And so we see nice steady build of those things. It's still relatively early innings with a lot of opportunity left to go, but we are continuing to see that stair step up as we talked about on our last call. And we're expanding further, expanding into other categories in our mature markets and seeing continuing yield from that.

But we're also launching new markets. So we are starting our international expansion now of the Amazon 3P partnership starting with Canada and Mexico. And so we see good opportunity on the international front.

But again, we're still getting a lot of yield out of our largest, most mature markets. But across each of those dimensions, we see our third-party partnerships overall, doing exactly what they were intended, which is rounding out gaps in our auction and helping us do more in unmonetized or undermonetized markets, as we talked about before.

Eric Sheridan

Okay. Understood. I want to turn to artificial intelligence. You guys have been at the forefront of a lot of AI and machine learning. In my opinion, things like visual search and computer vision jump out. Can you walk us through how you're leveraging AI from a user and the monetization perspective and touch more on the broader long-term strategy around AI for Pinterest as platform?

Bill Ready

Sure, yes. AI is the core competency for us, and we've seen really tremendous yield from it. Just to put this in perspective, just from things we talked about before. We talked about when we moved to GPU serving large models, that was a little over a year ago that we started that. We saw that we got 10 percentage points lift and the relevance of our recommendations from that.

And part of that is about the power of large models and GPU serving. Another part of that, though, is about the really unique signal that we have. And so that's part of our AI strategy as well, which is in addition to using large language models, tuning them to our unique signal, we're leaning into the things that gain more of that unique signal for us, like the curation behaviors and things like that, where when people come on to Pinterest, we don't just see what they clicked and bought.

We see them days, weeks, months before they buy, where they're thinking about how to update their wardrobes for fall, we see them well ahead of fall shopping, starting to put outfits together and say, okay, well, yes, I'd put this dress in this handbag and these shoes together. I put this jacket with these pants.

And we see hundreds of millions of users that are making all these product associations, that computer vision alone wouldn't get you because it's not just pattern matching, right? You don't want your dress and your handbag and your shoes to be the same pattern necessarily. That doesn't make a great outfit. But what does make a great outfit is being able to see, oh, here's how millions and millions of users are thinking about pairing this kind of dress. What kind of handbag and shoes would they put with this dress?

And then how does that line up with this user and the recommendations we make. So those things are about us really capturing more and more totally unique signal. And so I think as we project forward with AI, there's -- I think you guys sort of break AI into sort of two components.

There's one part of this that is analogous to what happened with the cloud compute providers, where it's sort of a Clash of the Titans that in terms of how many cloud providers where they're going to be, you're going to have a few that would be the major providers of cloud infrastructure. But then you also have lots of people who would benefit from consuming that cloud infrastructure in those building blocks.

And we think about AI similarly, where if you think about the building general-purpose large language models, that's massively expensive. It's sort of a Clash of the Titans and sort of $1 trillion market cap companies or those that are very closely aligned to $1 trillion market cap companies because it's a lot of spend.

But then what we're seeing is that we can go take those large language models and tune them to our unique signal and see significant increase in yield from that signal. So to make that really tangible, when we test this across most of the major model providers, we're always testing those. And we see that when we take those large language models, the general purpose large language models bring them in-house, and then tune them to our unique signal, we see 300 basis points lift from tuning to our unique signals.

And to put that in context, this is a world where in past life, if you got 10 basis points improvement, 15 basis points improvement, like, oh my gosh, you're doing cartwheels. It's fantastic. How powerful is our unique signal, 300 basis points improvement over what we would get from those large language models off the shelf from those providers when we tune them to our unique signals.

And then that unique signal is getting higher and higher fidelity as we get more engagement per user, more shopability, more curation and action on the platform, we're just getting totally unique signal that others wouldn't have, and that's evidenced by that 300 basis points lift that we get when we tune to our signals.

And so in that, I think you've seen us really, really demonstrate AI as a core competency where we have some of our own deep computer vision, stuffs like really narrowly fit for purpose, where we do some of our own models, just where these things are really fit for purpose, but then we're taking the off-the-shelf large language models, but getting totally unique results from them because we have totally unique signal to feed to it.

Stated very simply, the AI is only as good as the signal upon which is acting. And that's why we're really curating and harnessing this totally unique signal that we get from the human curation on our platform, which I've talked about a lot. But that's just another example of quantifying that a 300 basis point difference in the relevancy of the recommendations based on how that model would have performed off the shelf versus how it performs once it's tuned for our unique signal. So we're continuing to lean into it that way.

Eric Sheridan

Okay. So that's an example really interesting about how you can take AI and drive a lot of improvements in your own business. A pivot on that, that's probably one of the biggest debates I get or questions I get from investors is in a world of digital advertising, you are competing with other companies that are either hyperscalers or spending money like they're hyperscalers. They're spending tens and tens of billions of dollars on AI. Maybe bring it back to the competitive landscape. So how do you see AI the propensity to spend versus bringing it back to competition and where Pinterest fits into that world view?

Bill Ready

Yes. I think on that, a couple of things there. One, I think we've proven out that we can compete really well. Like we are taking share in performance budgets. And I think that really gets to just how much the uniqueness of your signal matters, right?

And the inherent intent of users on our platform, right? You can't sort of -- it's very hard to change what the user wants to do on your app. And this is one of the things I saw from the outside looking in before I came to Pinterest a little over two years ago, was that the users were here to shop ,but they just didn't have the ability to take action. They were window shopping, but all the stores were closed. While as we've opened those stores, they're taking action, but then this curation signal is a totally unique signal.

And so when you ask like, Bill, how are you breaking into these performance budgets when you were sort of outstanding start two years ago? It's because we have this total unique signal around the curation behavior that lets us make these really great recommendations. And so if you go talk to users on Pinterest and ask them, well, why are you going to Pinterest, right? One of the things they will say is that, well, Pinterest just gets me. And I talked about that 10 percentage point improvement in the relevance to our recommendations when we first went to large language models roughly a little over a year ago, that is very user perceptible.

After we did that, that's when we started getting feedback from users that were like Pinterest just gets me like, wow, like it's given me great recommendations. It knows my style. It's helping me discover new things.

And that's very different than what has existed elsewhere where there's lots of other places that when the user knows what they want, that they'll go search for Nike Air Max 270, size 11, in stock now, well, okay, those are things like the users decided what they want. And I think the last 25-plus years of e-commerce were really about, when the user knows what they want, help them get the cheapest and the fastest.

I think this next decade is really going to be about the rest of the shopping journey, which is people start with a much more general sense of, help me find some cool running shoes. And from that, because we have our taste graph and that unique curation behavior where we don't just know that user's taste, we know how other users styled those sneakers. What did cool running shoes mean to them? What were the different brands available? That's why we just got a really unique signal that's letting us give great shopability to the user in a very different way than they experience any place else because the user has intent, but it's a loosely defined intent, which is a magic mode for the advertiser to meet the user when they have clear intent, but haven't yet decided what to buy.

That's a unique moment in the user journey that we have, but then unique signal we get from that, that lets us drive really great performance. So it's not a bet about can we beat trillion dollar-plus companies on the raw horsepower of the AI. My bet in coming to Pinterest was that, that raw horsepower, just like cloud compute, was going to be broadly available just like how compute -- those building blocks will be available.

And then who has unique signal to go do something, who can assemble those building blocks in a unique way. And we have a user behavior, user intent and really unique signal to let us just go do something different with that AI, and we're proving that out quarter after quarter after quarter as we gain share in those performance budgets, capture share there.

And I think it's also -- it's not even that we're -- I think when you look at our user behaviors, so talk about the average [indiscernible], when you look at to the user lens, I don't think it's even that we're competing for existing behavior. I think it's more blue ocean, where the first 25 years of e-commerce solved for buying, but not really for shopping, with shopping being the hey, think about how you shop in the physical world.

You want to refresh your wardrobe for fall, whether you walk into a mall or a series of boutiques, you have sort of a general idea and you sort of walked the bizarre and discover and proves, that's what Pinterest does and it's purely visual search kind of way and its unique behavior and it's I think letting us compete really effectively.

Eric Sheridan

Okay. I know we're going to have a few minutes left. So I'm going to try to squeeze one more question in. You talked earlier about Gen Z as a percentage of your base. I think 40% now of your base, fastest-growing demographic. Why do you think the platform is resonating with Gen Z?

Bill Ready

Yes. So this has been one of the brightest of bright spots for us. I think two years ago, when I was coming in, I was getting questions like is Pinterest going to exist? Sort of the narrative was aging up and aging out. And our app is actually aging down now, which is really unique for mobile apps.

That over time, you tend to sort of age up, and we're aging down. Gen Z is now more than 40% of the users on our platform. They're the largest, fastest-growing demographic on the platform. We are growing across all demographics that we track and all geographies that we track, but Gen Z is our largest, fastest-growing. And at the core of why we're winning with Gen Z, there's the shopability and actionability that I've talked about, that is part of it.

But Gen Z loves the curation behaviors, right? They really -- Gen Z, they're at a time in their life where they're still figuring themselves out, right? And so investing in that time to go, what's their style, what do they want that to be? They're getting their first apartment well, how do they want to design that first apartment where they want. They spend a lot of time curating and they love that.

And then in addition to that, we've talked a lot about investing in Pinterest as a more positive alternative to social media. And if you ask Gen Z, why they go to Pinterest, two things they will pretty reliably say. One, they'll say that they see it as an oasis away from the toxicity they experience elsewhere on social media. Stated simply, social media is where there's lots of engagement over their selfie. Pinterest is where they go to invest in themselves.

Social media is about their selfie, Pinterest is about investing in and figuring out themselves, for themselves. And sometimes with a small group of close collaborators, close friends or people of like interest, was a very different place than the rest of social media. So they really love that positivity, but then the curation behavior and the shopability is quite central to that as well. And so I think that's worth noting that we've got these things are working for Gen Z. But they have broad applicability, because the curation is broadly applicable beyond Gen Z.

The shopability is broadly applicable beyond Gen Z. The positivity is broadly applicable. It's not just good for the user, it's great for advertisers. Brands want to be in a positive environment. And so we see that really cutting through and against why we're -- Gen Zs our largest, fastest-growing demographic, but we're growing across every demographic, every geography that we track.

And I gave that stat on weekly active to monthly active and that deep engagement, we shared at our Investor Day a year ago, our most recent -- well its also -- well, it's exceptionally rare for an app to age down. Our most recent user cohorts that we're bringing on are approximately twice as engaged as cohorts from years prior.

So we truly have our best product market fit ever on the platform. And it is all these things together, working really well. And so that weekly active to monthly active, you're seeing more of the monthly become weekly actives, that's broadly applicable too.

And really important to call out. It's the most true in our most mature markets. So we see the best weekly active to monthly active ratio in the U.S. in UCAN, and we see that across our mature markets as well. So all these things line up quite well.

And I'd say, just as we come up on time here, one of the things I didn't mention in the AI comments, I think part of us having AI as a core competency, you've seen us really, simultaneous to transforming the product for shopability, standing up a performance ad platform from scratch, we've also driven great margin expansion. We've put great discipline in the business. And all those things where we're lining up, really efficient use of AI aligned with things that drive clear user engagement, clear monetization.

We're getting good unit economics on what we do with AI. So as we're building the business, we're opening up new demographics, driving deeper engagement, getting great use of AI, but also making Pinterest as a much more sustainable, durable business, great margin expansion and finding ways to get great yield out of our investments in the product.

Eric Sheridan

Okay. We are going to have to leave it there. Bill, thanks so much for being part of the conference.

Bill Ready

Thanks for having me.

Eric Sheridan

Please join me in thanking Pinterest for being part of the conference this year.