All right. Good morning, everyone. Welcome to day four of the Morgan Stanley 2025 TMT Conference. We are thrilled today to have Bill Ready, the CEO of Pinterest, with us. Thanks for having me. How are you? Good morning. Good to see you again. Welcome back. Good to see you as well. Before we get started, let me do the disclosures. For important disclosures, please see the Morgan Stanley research disclosure website at www.morganstanley.com/researchdisclosures. If you have any questions, please feel free to reach out to your Morgan Stanley sales representative. All disclosures are also available at the registration desk. 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 Form 10-K for a discussion of the risk factors that may impact actual results. All right, check. A lot to talk about. Yeah, absolutely. Both from a macro and micro perspective on the platform. So I always sort of like to start with the big picture of you've laid out expectations over the last couple of years a couple of times with your Analyst Day and sort of how you think about growth and profitability. Maybe sort of now stepping back from the key KPIs and areas you've executed on following the Analyst Day, are there two or three areas where you'd say, "Here is where we're really executing ahead of where we thought we would be when we laid out the forecast"? Yeah, I mean, a couple of things I'd say, like, you know, coming up on three years into the business, 10 quarters, not that I'm counting. But the things that we feel really great about are that we set out to make Pinterest a shopping destination and really solve actionability on the platform, stand up performance advertising, and really drive the unique curation behavior in Pinterest to do some things that I felt could be really differentiated in the shopping space. And if you look at where we are, take what we did in 2024, quite a transformative year for us where we more than doubled our revenue growth rate, we're at record highs on MAUs, record highs on weekly active to monthly active. So even as we're bringing record new users onto the platform, we're seeing deepening engagement per user. Oftentimes, as you add record levels of users, that could be dilutive to engagement per user, but we're actually seeing our best ever engagement per user because we've got our best product-market fit ever. And at the core of that is actionability, it's curation, and it's the fact that we have made AI deeply integrated into every part of the platform, and we have completely unique signal to go do really unique things with that AI. That's one of the things that I saw coming into the business, that this curation behavior where people were working out what they wanted to buy days, weeks, months before they bought it, that was truly unique in the ecosystem. It didn't exist anyplace else, but it was lacking actionability a few years ago. As we brought that actionability in, we've seen that that actionability has been at the core of the increase in our engagement. We talked about 90%+ growth in clicks to advertisers in Q4, even as we lapped 100% growth in clicks to advertisers the Q4 before that. So the flywheel is really spinning on that. And the performance ad side of it, you're seeing that not only is that working really well for users, that's working really well for advertisers as well. That, as we've stood up the basic components of a performance ad platform, that's letting advertisers tap into that really unique shopping behavior, and we're seeing that flywheel just really, really spin. And importantly, I would say each of these things that we've been launching have compounding effects, right? They're not one-off discrete launches that just are sort of like one and done. It's really compounding effects of, like, as we've made the ads really great content, as we've brought shoppability into the platform, that's driving deeper and deeper shopping behavior, brings more advertisers, makes the shopping even better, brings more users, and you're seeing that flywheel spin, and that's what we laid out at our Investor Day, and we're seeing that go really well, so we feel great about that and just the fact that the AI-enabled benefits of that, both in the quality of our recommendations, the AI-enabled performance ad platform, Performance+ that we just recently launched, all those things are driving really great results for users and advertisers, so feeling very good about that, but obviously there's a lot more of that opportunity in front of us, so a lot more to do there. Yeah, you've done a lot on both the engagement side and the ad tool side over the last few years. It really is sort of showing through in the numbers. Maybe the other side of the coin, though, are there areas either from a micro execution perspective or from a macro perspective, whether it's trade, tariffs, name your latest macro source of uncertainty that are sort of behind where you thought they'd be or holding back growth? How do you sort of see any headwinds to growth you'd anticipate, and how do you sort of execute through those? What I would say is that, well, first of all, we don't have any better ability to predict what happens with tariffs or trade or things like that than anybody else. But if you just look at what we have done, we have turned Pinterest into a shopping destination, right? And in any environment, people are still going to shop, and we've created really differentiated ways to do that that advertisers are leaning into. But on the macro, I'd say, while we have no better ability to predict than others, at least in terms of what we currently see, we don't currently see advertisers overreacting or those kinds of things. It's obviously a fluid environment, and so I'm sure, like everybody else, everybody's watching every day to see what happens. But we're not seeing signs of overreaction at the moment. But we've created a unique shopping destination. We shared this on our last call that 66% of our weekly Gen Z users think of Pinterest as a first place to go shop. Also, in that notion of us as a destination, it's also the case that 85%+ of our users come to us via our mobile app, and 100% of our reported MAUs are logged-in users. 85% of our users come to us via our mobile app. That's up 10 percentage points since I joined the business 10 quarters ago. More than 90% of our revenue comes to us via our mobile app. We are a destination, and at the core of that is great shopping behavior and really unique shopping behavior. Even as you see shifting consumer behaviors, we're finding really unique ways to go to satisfy that. For example, Deal Ads was great for us in Q4, delivered great value for advertisers and for shoppers, and last thing I'd say on this is that, again, no better ability than anybody else to predict the future, but we played through an ad winter, and we were the only major ad platform that grew consistently every quarter through that ad downturn, so we've been growing faster than the market, taking share, especially in the places that we have invested in shopping, lower funnel, and so I think we have a real secular growth story here where the fundamentals of the business have just never been better. We're a unique destination for shoppers. Gen Z is our largest, fastest-growing demographic, 40%+ of the platform. Pinterest is where Gen Z goes to shop, and we think that gives us lots of ways to go execute regardless of environment. I want to actually ask you about sort of your last point. Pinterest is a place people go to shop. You have invested quite a bit the last few years in basically GPU-enabled machine learning and some early changes because of Gen AI and GPUs to make it more shoppable, more personalized. Can you just sort of walk us through some of the more important products or the more important changes you've made in how you're analyzing your data so we can sort of understand what's happened? And any way you can sort of give us more quantification of the benefits you've gotten from GPU-enabled changes? Sure, yeah. So first thing I'd say is AI is core to everything that we do, and we, I think, have clearly demonstrated that we are a winner in the movement toward AI. We're clearly demonstrating that we're a winner in that. And the way that we're doing that is really aligning AI to unique use cases that really happen on our platform in a different way than any place else, the curation behavior, the shopping behavior. But to make that even more tangible, every recommendation we serve, both organic and ads, comes from our AI-enabled recommendation systems with GPU serving behind that. I shared a little over a year ago, about a year and a half ago, that when we moved to GPU serving for our ad stack and for our organic stack, that we saw more than a 10 percentage point lift in the relevancy of our recommendations. And then we've had lifts further since then as we've done things like this last earnings call. I talked about how we increased the context window by 30-fold. And in doing that, that gives us many more signals, much more of this unique curation behavior, the clicks that are happening on our platform. Increasing that 30-fold gives us even better insight into users' tastes and preferences, and that led to a 250 basis point plus lift in saves, 150 basis point plus lift in clicks, even after having had a 10 percentage point lift in the relevancy of our recommendations a year and a half ago. I also talked about the relevancy of our ads. The relevancy of our ads in search has more than doubled over the last two years. That really is about the AI-powered recommendation systems and our unique signal to train that AI, to bring that point home a little more even. We are doing things that are unique to our platform around computer vision, but we're also leveraging things off the shelf. We find that even when we take large language models off the shelf and train them against our unique signal, we're able to see several hundred basis points lift beyond what those models could have done on their own, which just gets to the value of our unique signal. Those were some examples of where we're seeing real benefit, but we think there's a lot more of that to go. Our advantage here is that I've talked about this before. I think the basic building blocks of AI, like cloud compute, are going to be available to everybody. And then the real question is, who has unique feedback loops in their business to go train the AI to do something unique and different for them, and who has a direct relationship with the consumer to go do something there? So I think that's where our unique curation signal, the fact that we are a destination, 100% of our users logged in, 85%+ of our users coming to us via our mobile app, 90%+ of our revenue via our mobile app. So we are a destination. We have a unique experience. And if you boil all that down to, like, what's it mean for a user? How has Pinterest become the place that Gen Z goes to shop? Go ask a Gen Z user, and there's two things they'll say to you. First, Pinterest just gets me, right? We understand their taste. We understand their preferences. Our Taste Graph has grown by 75% over the last two years. That's what our AI is building upon. So first, they'll say, "Pinterest just gets me." Second, they see Pinterest as an oasis away from the toxicity they're experiencing elsewhere. So it's a place where they can go to figure themselves out, and it's just unique and different than anything else in the social media space. And so we feel really good about what that lets us go do with AI as that continues to be part of our secular growth story. It's interesting. You've been very consistent in this because I remember when we did the TMT conference right after GPT launched, and you and I were talking with your background at Alphabet about how these capabilities are going to enable us to make the platform more personalized, more shoppable, more visual, more user. It really is. You've been very consistent through that. That's exactly right. That's exactly right. Try to keep a say-do ratio of one. That's right. That's right. So let me ask you, so from an advertiser or spend-per-advertiser perspective, you're rolling out these new tools. You're better analyzing your data. You're seeing good incremental signal. Where have you made more progress on getting more advertisers on the platform from a vertical perspective? And what are sort of some of the areas where you think you can have a lot of greenfield to add even more advertisers in verticals? Yeah. So I've talked about this a bit, but step all the way back from it. Over the last two and a half years, we started first with getting the user behavior, right? A few years ago, Pinterest had solved for sort of digital window shopping, but all the stores were closed. First thing we had to do was open the stores, make it so people could take action. So we had to get the user behavior first. Then as we had the user behavior and the clicks going to the advertisers, we then, you saw us at the start of 2024, start to really push on privacy-conscious measurement so that the advertisers could measure those clicks and conversions. Then following that, we launched Performance+, which is our AI automation suite for advertisers that just came out of beta in October, so right at the start of Q4. But that makes it easier for the advertiser to take action on it. So it's like, get the user behavior, drive them to the advertisers, make it measurable, then make it easy for the advertiser to take action on it. And if you look at where we are in that journey with advertisers, we've talked about how initially we saw the best strength from the largest, most sophisticated advertisers. Now that's really promising because they're the most discerning. In past life, I'd always say if you can make them happy, you can make anybody happy. And we've seen them really lean in, and they've been a consistent source of strength for us. Then the next step beyond that is they sort of scoop up the arbitrage first, right? They're the most sort of quick to pounce on sort of an arbitrage-type situation. We're like, "Okay, there's a new shopping opportunity here." But then as we've made the measurement more accessible, meeting advertisers where they are integrating with 30+ different measurement sources of truth, we saw consistently as people could measure it, they'd shift share of wallet to us. And then now with Performance+, we're early on in that, but it cut campaign creation time by 50%, right? And so it's making it easier to take action. And it's that next group down. Think the sort of mid-tier enterprise retailers, large direct-to-consumer in the $1 billion-$30 billion kind of annual sales range. We're seeing that as like the next cohort down starting to lean in. As Performance+ gains adoption, not only will that make it easier for a broader swath of advertisers to easily get benefit from our platform, it's also the case that it's letting us get to a larger share of wallet, even for the very largest, because for the very largest, they weren't necessarily advertising their whole catalog with us. For example, we just launched item-level bidding, and ROAS bidding is coming out before the end of this quarter, has been in beta. We've seen that there's been really nice performance from that. On ROAS bidding, for example, we saw a large retailer that was testing that through Q4, and as they were testing that, they saw performance improvements and share of wallet shift. And so all these things sort of stack together and have compounding effects, but there's still a lot more to go, both in terms of getting more share of wallet from the largest and getting to that $1 billion-$30 billion group and just working further and further down the market. And we're still a long way from having monetized the full shopping behavior on our platform. We grew 90% in clicks in Q4, and our revenue growth rate doubled, more than doubled last year, but it didn't grow 90%. So there's still a lot of those clicks and saves and conversions that are yet to be monetized, even as we get more users coming, more clicks happening. So there's a lot of that sort of pent-up shopping behavior that the flywheel is spinning, but we are a long way from having fully monetized that. 90 would be something. We'll take 20. You talked about the Performance+ product and some of your larger advertisers have adopted, and now we're sort of looking to go down the longer tail and sort of smaller advertisers. Are there any learnings you have on the type of lift you see in spend per advertiser from those larger advertisers? They adopt those tools that perhaps we can think through as you go down the tail? Yeah, I mean, the main thing I'd say is that we have pretty consistently seen that as we drive shopping behavior and as advertisers implement privacy-safe measurement, we talked about how very consistently we would see that lead to share of wallet shift. And Performance+, we're early with that. Much more of that opportunity is in front of us than behind us. And again, all these things, think of these as multi-quarter, multi-year cycles. But you cut campaign creation time by 50%, make it easier to take action. We've seen 20% improvement in CPAs and beta testing on those things. And so those are things that we think will have durable benefit, especially as it's plugging into a totally unique shopping behavior on our platform where you're meeting the user in this magic moment where they have clear intent but haven't yet decided what to buy. That's very different than the rest of social media that may be more serendipitous kind of discovery, but here there's clear intent, but they haven't decided which brand they're going to buy, exactly which product they're going to buy. It's a great magic moment to meet the user, and the more tools we give the advertiser to tap into that, the more they do it, and it makes the product better for the user as well, and so again, we see just a long-term compounding effect to that, and think of that as a multi-quarter, multi-year cycle, and you've seen that be true for Performance+ with our AI automation suite. As you look at our larger peers, take like a PMax at Google or an Advantage+ at Meta, those things are many years in and still launching new improvements and those kinds of things. So think of that similarly for us that we'll be improving that in perpetuity. But the consistent phenomenon is every time we give the user, or excuse me, the advertiser, a better ability to tap into the unique user shopping behavior on the platform, we see share of wallet shift, and we feel really good about that. It could be helpful to sort of talk about the rollout of Performance+ operationally to some of those smaller advertisers because I think there is sometimes this external perception of, "Oh, you just flip the switch," and advertisers, they use it. But maybe to sort of walk us through some of the key execution hurdles you have to clear over this year just to get those next 10,000, 20,000 small, medium-sized businesses using Performance+ and really spending more. What I would say is that these are frequently, most oftentimes, multi-quarter adoption cycles, especially once you get beyond the largest, most sophisticated, again, like the largest, most sophisticated sort of spot and act on arbitrage much faster. The rest tend to see they will want to see behavior be consistent over multiple quarters. They will lean in gradually on those things. And so I've said this very consistently. Don't look for hockey sticks in one quarter. Don't think of each of these as like a discrete launch. Think of these as we've basically stood up the major components of a Performance+ ad platform over the last few years, Performance+ being sort of the third leg of the stool there where you get the user behavior, the measurement, then the automation suite. And now those will continue to improve, and there's a flywheel effect, a compounding effect to that. Think of that going over multiple quarters. We do see consistently that as advertisers are able to see and measure the clicks, it's easy to take action. They shift share of wallet to us, and we think there's a lot more of that to go because again, when I just look at the monetization per unit of intent on our platform, sort of thinking of analogs from past life, we are still a long way from having fully monetized the shopping behavior that we have now. Add to that, we're just getting started with that internationally as well, right? It's rough numbers here, but roughly 80% of our users are outside of UCAN, but only 20% of our revenue, right, ish. You're seeing that start to accelerate as we're taking our shopping story international. I think there's more of that pickup to go as well. But think of those as just steady, long-term compounding improvements in the platform and getting more and more of those advertisers taking advantage of both in our largest market here in UCAN as well as international. Let me ask you about the deepening engagement you mentioned. The platform is much more shoppable. Can you give us examples of how consumer behavior on the platform has changed as you sort of better analyze your data? Are they searching more? Are they coming back and searching more boards? What are they doing more of? Yeah. So at the very core of how we've gotten to record MAUs, record depth of engagement via that WAU-to-MAU ratio, the actionability and the curation is at the very core of that. And so we shared at our investor day how our most recent user cohorts are approximately twice as engaged as prior user cohorts. And at the core of that engagement is it's shopping, it's actionability, it's curation. And the really powerful thing is that curation behavior is what gives us a unique ability to serve recommendations. So when a user on Pinterest says, "Oh yeah, I go to Pinterest as the first place to shop because Pinterest just gets me." Well, how does that happen? Hundreds of millions of pinners are on our platform, and they're curating how they're going to style an outfit. There's other places via cookies across the web. Other places can say, "Oh, they know every piece of clothing that you bought." They've got no idea how you put that together in an outfit. And that's what users do on our platform every day. And as we've made that more granular with things like collages, we're getting even more granular signal that the flywheel effect of that is that users then curate more. That better curation or that increasing curation gives us better signal to make better recommendations. And so then the next time the user comes into the platform, it's like, "Oh my gosh, yes, that's exactly what I want." Or if they go search for something, it's not just going to be a search result for how to pair sneakers with a suit. It's going to be nicely done. It's going to be, well, how to pair sneakers with a suit for your style and taste, right? Or how to pair sneakers with a skirt in your style and taste. "Oh, do you want to pair it with a pair of Golden Goose, or do you want to pair it with some Loewes or whatever is your price point or your taste?" Those are things that we can uniquely do. There's a compounding effect of that, so that curation behavior is from the very onset what I thought was really unique about Pinterest. As we've leaned into that, we're seeing it really take hold. Back to like that 75% increase in our Taste Graph, that just tells you how much more of that signal we're getting from the users from that curation and shopping behavior. But then the compounding effect is like as we do more with the AI, more with our GPU-enabled serving recommendations, the relevance of the recommendations keeps getting better. Double the relevancy of the search ads, 10 percentage point increase in the relevancy of our recommendations that we've then made multiple hundred basis point plus improvements on top of that, and we'll continue to do so. Those are the things that just make us a truly unique shopping destination for our users. But it is that actionable and curation at the very center of it. Said simply, Pinterest is where Gen Z goes to shop, and that's beyond Gen Z as well. I want to ask you about the Gen Z, non-Gen Z. One of the investor perceptions that some investors are concerned about is the non-Gen Z user behavior, user growth, given this perception that it's the older people have more money. They're the ones that are going to spend more. And so maybe talk to us about sort of non-Gen Z behavior. Is that a concern of yours? Is it Is it being monitored? What's going on with non-Gen Z? Yeah. So Gen Z is now 40%+ of our platform, largest fastest growing demographic, and our most deeply engaged. At the same time, we've seen broad-based growth across demographics. So there's plenty of millennials shopping on Pinterest as well. We've seen broad-based growth across geographies, growing across all of our reported geographies. And with that, it is pretty consistently the actionability of the shopping and the curation that comes with that. And within Gen Z, I remember when I was building Venmo, the team and I would get the question all the time, like, "Are these millennials ever going to shop and buy things?" And millennials do a lot of shopping and buying, including on Pinterest. But within Gen Z, Gen Z spans from high school through college through young professionals. Just as I saw with what happened with millennials in my past life with Venmo and PayPal, like, yeah, those Gen Z users, like again, 66% of them say they see Pinterest as a first place to go shop. That's up from 54% a year ago. So that's increasing. And they go to college and get a credit card and start making more of their own purchasing decisions. And then they graduate college and get a first job, and then they're buying more things. And so we see that maturation. And interestingly, unlike other places where users may age out of a platform because, "Oh yeah, this was a cool thing to do in high school. Yeah, I'm growing up now. I don't need it." As the user ages up, we just become better and better for them, right? Because as they age up, if shopping is the core use case, which it is, as they age, they do more and more shopping, and we see exactly that phenomenon on the platform, that as they age, we get better and better for them as they age. Okay. I want to talk with you about sort of investment philosophy in Gen AI and sort of how to think about margin pressure, profitability profile of the company as you're sort of balancing how much should we be investing in next-gen capabilities as opposed to allowing flow through to free cash flow. Yeah. I mean, I think this is one of the strongest parts of our story and what we have done is that it's not just that we've hit record levels of revenue, users, engagement. We've had significant margin expansion as we've done that, right? We're at our best product-market fit ever. AI is core to what we do. We're deeply invested in AI, and we're expanding margins and showing that we can deliver profitable growth. And the way that we have done that is that we have, first of all, I said this from the day I came into the business, that I'm a huge believer in Theory of Constraints and that in accepting that and embracing that to say, "No, no, we're going to find ways to do this that are monetizable and that we can do so cost-effectively." That's been a goal from the beginning, and we've done that. And you've seen examples out in the market now. Like, "Hey, if you focus on that, you can do it," right? Like DeepSeek's a great example of necessity is the mother of invention. Similarly for us, like, "Yeah, we're not a trillion-dollar market cap company. We can't go throwing tons of money at things, but we can drive creativity." And we've got a really unique use case with really unique signal to train that AI tied to highly monetizable events. And so as we have consistently pointed our AI investments toward highly monetizable use cases, they're unique and differentiated. That's allowing us to go invest significantly in AI, but also do so in a way that is high ROI and where the payouts are near term. And I think if you step all the way back across the AI landscape, I said, again, very clearly, I think we are a significant beneficiary of the move to AI. We're a winner in the AI space. But broadly speaking, in AI, you hear people talking about either, "Hey, we're driving all this cost out of our business through productivity improvements," which, by the way, we're doing a bunch of that as well. But you hear people either talking about, "We're driving a bunch of cost reduction through productivity improvements," or, "We're driving a bunch of usage that someday is going to lead to revenue." And we are using AI to make significant improvements in our product that are leading to revenue now, leading to good profitable growth now. And yes, we're driving great productivity. I talked about 15% of our code bases now generated through AI. We have strong adoption of that with our engineers. So we're doing things that do drive greater productivity. But on our core product, we're using AI for every pin that you see on the platform, every ad you see on the platform. And it's highly monetizable, highly differentiated, and so therefore delivering great profitable growth. $1 billion+ in adjusted EBITDA last year. Importantly, with a 90%+ free cash flow on that, which just gets to the asset-light nature of our business and the fact that we do have really great sort of free cash flow in the business and an asset-light model. It's interesting. The improvements you've made to the platform, they're so much farther along now. I would argue than even a year ago. A year ago, I remember we talked when we were 30 minutes in. We probably talked about partnerships last year, seven minutes in. Here we are 30 minutes in. So let's talk about partnerships. So Amazon and Google, we've had those partnerships now for a little while longer. Can you just sort of help us understand areas that may have gone better than expected on Amazon and Google partnerships and sort of anything that's been more challenging than you thought? What I would say is just from when I first talked about bringing in third-party demand, first, it was that we wanted to solve gaps in our auction, right? And we needed to bring in relevant ads to solve shoppability. And stepping back from it, the platform is more shoppable than ever. The ads are twice as relevant as they were a year ago in our search ads. And so you've seen us do exactly what we said we would do in terms of using it to round out gaps in the auction. But also what I said then is like it was going to be complementary to it. It was coming to the same auction. It would be complementary in that first-party is still the primary way that we're going to drive that. What I talked about on this last earnings call was that we're actually seeing our best strength in first-party, which is as advertisers are seeing that we are a unique shopping destination. We're delivering great performance ads results. They're coming to us directly. And all else being equal, of course, we prefer our advertisers coming to us directly. And that is our greatest source of strength. And so we continue to see partnerships helping to round out those gaps. But first-party, which is what we want, is our greatest source of strength. And we'll continue to leverage not only third-party partnerships, but resellers for international, things like that that sort of keep us cost-effective as we go into longer-tail markets and those kinds of things. So we've built a capability that is additive to our platform. And we'll continue to leverage that to round out gaps or help solve for markets where we may not have first-party sales or things like that. So we feel really good about the capability that we created, but we feel even better about the fact that we're delivering great value for advertisers. And as we do that, they prefer to come to us directly. Yep. Externally, when we build out our multi-year models and we sort of have growth rates and assumptions about how quickly Gen AI could move throughout a company to have benefits like you're seeing now, we're external. You see a lot more than we do. Are there any sort of challenges or constraints that sort of dictate how quickly you can keep rolling out more generative AI products or move more of your data toward these GPU-enabled machine learning algorithms that you're not asked about, that you just think this is something that investors should be mindful of how quickly these benefits can accrue? I'd say for us, the main thing I call out for us is don't think about the things that we've done as sort of one and done. These do have compounding effect, right? That as the AI, as we have that unique curation behavior, that unique signal that we understand people's preferences, the AI then helps us make better and better recommendations. That brings more shopping behavior. The AI will continue to get better. As the AI continues to get better, our unique signal will make us more and more differentiated. Think of that as a compounding effect and a flywheel effect. There's a lot more of that to do for us. But I do think that's the kind of thing that there's just a natural benefit that if you have unique signal, as the AI gets better and better, I think you become more and more differentiated because of that unique signal. Like the example I gave of, "Oh, we expanded our context window by 30-fold at a fraction of the cost increase of that." In fact, like sort of Theory of Constraints and being really efficient with it, but driving significant improvement and saves and clicks and those things from that. Well, the next iteration of the model that's even better than that, and we get even more context, the recommendations will just get even better. That then brings more shopping behavior, brings more advertisers, more signal, and there's a real flywheel effect that compounds over time. And so I would think about it that way. Got it. All right. Well, Bill, well, thank you very much for joining us. We look forward to seeing the compounding benefits this year, and we'll see you next year. Yeah. Thanks for having me. Thank you.