Good afternoon, everyone. We hope day two has been a productive day here at the 2023 Morgan Stanley TMT Conference. We are thrilled to wrap up day two with Bill Ready, the CEO of Pinterest. Bill, it's good to see you. Thanks for joining us.
Yeah, thanks for having me.
Gliding on the last thing between folks and drinks, apparently, so, hopefully we'll make it entertaining.
We have a lot to talk about.
Yeah. There's a lot to cover. First, the disclosures. For important disclosures, please see the Morgan Stanley Research Disclosure website at www.morganstanley.com/researchdisclosures. The disclosures are also available at the reception 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, and 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 Form 10-K for a discussion of the Risk Factors that may impact actual results. Okay. It has been about nine months since you joined Pinterest.
Let's sort of, start with a, with a high-level picture around you've had a lot of investor meetings, you've gotten your hands pretty dirty in sort of a lot of the ins and outs of the company. What is still sort of you think the most misunderstood opportunity for Pinterest as you sort of look ahead at the next two to three years?
I would say, by the way, I think there's emerging understanding of these things, but I'd say I think it's probably still underappreciated how fundamentally different Pinterest is than the rest of social media.
Mm-hmm
On some pretty important dimensions. I'd say first and foremost is that most of social media, the user's in sort of a lean back entertainment mode, sort of like passively consuming, whether it's pictures of friends or funny dance videos or whatever. They're in a sort of passive consumption mode. Pinterest, the user's in a lean forward mode. They come to Pinterest with intent and purpose, whether it's shopping or redesigning a room or, you know, finding a great recipe, they're in a lean forward mode, which is pretty different than the rest of social media. I'd say the second thing that's really different about Pinterest than the rest of social media is that, you know, Pinterest is a full funnel platform. Our revenue actually splits a third upper funnel, a third mid funnel, a third lower funnel.
There are other places, other ad platforms that you can find a, what I would call a synthetic full funnel.
Mm-hmm
Where the advertiser sees a full funnel, but it's actually distributed across multiple different consumer products. I'd say in the Western world, we're the only place that is a true full funnel from the consumer perspective, where the consumer goes from upper to mid to lower funnel in the same place. The last thing I'd say that's fundamentally different from the rest of social media is that, you know, Pinterest is, you know, it's quite different than the rest of social media, you know, in the ways that I already talked about, as well as from the user perspective. You know, it's avoided a lot of the toxicity of social media, and in fact, I've come out and talked about this a lot more publicly recently. I think that is a huge differentiator for our platform.
The fact that, you know, we can tune for positivity. Users, after they spend time on Pinterest, 80%+ of users, multiple studies on this, 80%+ of users will say they feel better after having spent time on Pinterest.
Mm-hmm.
The opposite is true of much of the rest of social media that users end up feeling worse after spending time. So I think that's a pretty fundamental differentiator and one that we're leaning a lot further into, and I think in the long term may end up being, our biggest, most important differentiator.
Okay. That's, that's a good place to start. Maybe, just sort of to level set before we get into a lot of the micro level drivers, can we just sort of, hear the state of the state from a macro perspective on the advertising side? Any changes or pockets of weakness across your different points in the funnel or verticals where you'd say, you know, here's an area that's particularly strong or particularly weak as you sort of start 2023?
You know, it's, a shifting sands across the ad market as I think most folks know. You know, I'd say, you know, we're seeing, you know, sort of puts and takes across the board and those shift a little bit, but, you know, there's been some surprising positive elements of that. We talked about how, you know, brand advertising has had some strength with us, especially as we're making it more performant, tying it into the full funnel. I think that's particularly true when we can do full funnel and say that, you know, users that do multiple objectives on our platform tend to see 50% better sales results. I think some of that is the full funnel nature of our platform, but we're seeing some pockets of positivity there.
I would say that generally advertisers are still adapting, you know, I think advertisers are making decisions closer to time, so there's a sort of shorter lead time.
Better way.
You know, while, you know, I don't know that I can predict any better than anybody else when sort of the ad market sort of, you know, fully normalizes, you know, our bet is not on just sort of riding the tide of the market. Our bet is that, you know, we can take market share.
Mm.
In that market. Even in a softening market, the last couple of quarters, you've seen us, you know, growing faster than most of our peers, if not all of our peers. I think that's because we're getting much better performance in our ad platform as we're investing in that. As we're, you know, solving for that full funnel, the lower funnel, the intent to action has historically been weaker for Pinterest, and as we're really leaning into that and really leveraging the fact that we have the user in the right mode, the user's here to shop. More than half our users say they come to Pinterest to shop. As we're getting better in intent to action, we're seeing that come through as well. That's helping us, you know, pace faster than the market.
That plus the investments in the ad platform I think are shining through. you know, we're 1% of the market.
Right.
Our plan is much more about how we grow that market share than it is about just riding the tide of the market.
Part of the one of the keys to driving the outsized advertising growth always starts with engagement with a lot of these platforms. There have been a lot of puts and takes around engagement over the last couple years. I guess as you sort of sit here now, what are sort of the key executional areas that you have to really hit right to grow, continue to grow users, sessions, time spent per user, and the overall minute pie on the platform?
Yeah. When I came in, at the start of Q3 last year, you know, the two biggest questions, and obviously I felt like you know, there was gonna be a good answer to these questions. I'd voted with my feet that that was the case.
Yeah.
The two biggest questions that I had and that I heard from others was, one, can you return to growth on users? You know, after multiple quarters of decline, after the sort of pandemic unwind. How do you get users and engagement growing again? I think we've answered that question pretty emphatically. You know, we've returned to user growth. Engagement growth is now double-digit plus. In fact, I think this is third-party data, it's not our data, so you all can choose how much weight you put in it, but most of the third-party data that we've seen would have us on a percentage basis growing engagement and time spent faster than pretty much all of our peers over the last two quarters.
I think this question of like, can we return to user growth, can we return to engagement, we've answered pretty emphatically. We've done that really by leaning into better personalization, core use cases on Pinterest that are differentiated from others, like really leaning into the fact that people come to Pinterest with intent and purpose and driving more of that intent to action. You know, we're seeing good traction there. Then the second biggest question, you know, when I came in was on this intent to action issue, you know, yes, users are discovering on the platform, will you get users to take action on the platform? You know, we're still early days in that, but we're seeing really good progress there.
We talked about, you know, shopping ads, you know, growing 50% year-over-year. New formats that drive seamless conversion like mobile deep linking.
Mm-hmm
Which is where, you know, you discover a product, you know, on, on our site or in our app, and we link you straight into the retailer's app to the place to go make that purchase. That was 40% of our shopping revenue over the Cyber- Five. That was a brand-new format that we released in Q3 and Q4, and that was 40% of our shopping revenue. That's really good indication that users will take action there. The last one I give is that as we're making shopping a core use case, so more than half the people on Pinterest say that they're there to shop, but if you looked at the product historically, it'd be really hard to find shoppable content.
You'd find a lot of products, but most of them you couldn't click to where to go buy those products.
Mm-hmm.
If you could, it was in a separate tab of the app, a separate corner of the app. It wasn't on the main surfaces. It wasn't in the home feed or search results. As we're bringing that shoppability onto our main surfaces, which we're early in that journey, but as we're doing that, we've seen really compelling results. One tangible example, as we updated our model, our algorithm for our home feed to consider shopping in our catalog, in the tests that we've done on that, we've seen 90% increase in the product saves from users-
Mm
When we update our algorithm on the home feed to consider shopping as a primary use case and our product catalog as a primary input to that, 90% lift in the product saves from users. Those are good indications.
Mm-hmm
On that second big question of, you know, you know, can you get users to take action?
Right.
I think those all, while none of that is fully solved, you know, I think good proof points. The, the theory of the case, that yes, that latent intent on the platform is very actionable and very monetizable.
Getting the users to take action, rolling out the shopping ads, getting deep linking, those are sort of, you know, important to driving higher monetization and ad revenue per user over time. What are some of the key underlying blocking and tackling steps you still have to sort of really get together on the back- end to scale those tools to have more shopping instances that are relevant for the users?
Well, I'd sa y, you know, a lot of the right raw materials sort of existed. Even when I joined, like they'd invested a bunch in the product catalog, 1 billion+ items in the product catalog, including, you know, most of the large retailers, partnerships with Shopify that would have most of the direct-to-consumer brands. Good product catalog, good foundation laid on that. We made a pretty meaningful update to the shopping strategy that shifted from, you know, Pinterest as the retailer, where we're serving a lot of the functions of the retailer, which would be a, you know, I'd say a slow and arduous implementation-
Mm
To one of Pinterest as a close partner to the retailer, where we're gonna help you discover and find the things you want, help you take action on those, but create a really seamless connection to the retailer. Not only is that better for the consumer, 'cause that's really what the consumer wants, you know, the vast majority of the time, better for the retailer as they partner with us, and you can move a lot faster. You know, we're, you know. The biggest thing for us has been changing the core user experience to have shopping as part of the core experience versus a sort of separate sideshow, or as a, you know, sort of, you know, implementing as a retailer, but implementing as a platform for retailers.
Those have been pretty big unlocks and we're, you know, like I said, early days, but good proof points and I think a lot more opportunity there.
How are you sort of prioritizing the U.S. Versus the international monetization story? I know for years, you know, the international piece has been a part of the story. Just walk us through sort of how the product roadmap is rolling out, you know, U.S. tools, then how long till they come international.
Yeah. For, for if you think about Pinterest as, I always have sort of in my mind a continuum from low intent to high intent for different platforms, then what's the monetization per unit of time spent based on that intent? I think at the low intent end you have, you know, platforms that tend to be more entertainment-based or you're, you know, looking at news or pictures of your friends. That's gonna be low intent. Then at the very high intent end, you'd have things like Google Search or Amazon that are, like, pure intent. You know, then in the middle you'd have, you know, things that have, you know, mostly entertainment but some shoppability, like an Instagram.
You know, I think Pinterest ought to be, you know, somewhere, you know, there's greater intent than you'd find on like a Facebook or an Instagram because the core reason people are here is to shop for more than half the users. It's not pure intent the way that Google or Amazon is. It ought to be somewhere in between there. Where are we now? We're at the very south end of that spectrum.
You know, as we think about where we ought to be and, you know, giving that user more actionability, more intent to action, not only does that, you know, solve a lot of leaked engagement for us, because the users are discovering things on the platform already, but then, you know, they'd have to go someplace else to take action. We bring that actionability in, that plugs a lot of leaked engagement for us. It'll also plug a lot of leaked monetization that should move us on that continuum to, you know, sort of the, you know, higher intent side of that. In, in the U.S. and in sort of the developed markets, I think that's more the story, like how do we capture that leaked engagement, capture that leaked monetization?
When you look outside the U.S., you know, we have a significant majority of our users outside the U.S.
Right
A tiny fraction of our revenue outside the U.S. Now, if you looked at other platforms, it would be pretty common that you would see the majority of users outside the U.S. because that's just where the people are, right?
Yep.
You'd see a better balance of oftentimes half or more of the revenue outside the U.S. as well. You'd have maybe a lower ARPU outside the U.S. than in the U.S.
Right.
Because there are many more users outside the U.S., even at a lower ARPU, you'd have more, you know, more than half the total revenue outside the U.S. For us, it's less than 20%, so, you know, it'll be a multi-year journey to go solve that.
Right.
We think that's a big opportunity in the long- term. You know, we've got good progress there as well. We've really increased our agency partnerships, particularly in Europe. You know, we're seeing that start to accelerate. Some of the things that we're solving, you know, are not distinct to the U.S., this intent to action.
Mm-hmm.
You know, users wanting to shop on a platform, these are core use cases outside the U.S. as well. You know, in the U.S., the intent to action, the better personalization, driving great results, better personalization works outside the U.S. as well. Outside the U.S., some of it is we're just very nascent in the ad stack outside the U.S., but as we're building more connections into more agencies, more tools for partners outside the U.S., we're seeing that start to pick up as well. We're not gonna get to 50% of the revenue outside the U.S., you know, in the next year or two.
Yeah.
Over the long- term, I think, you know, we'll start to gravitate more towards that kind of place in the long term, as you've seen from other, more mature platforms.
On the ad stack point, I wanna ask, well, I wanna ask a couple questions about increasing the number of advertisers on the platform. Let's start with the ad stack. Maybe just for everyone in the audience, walk us through the conversion API. Why is that important, and what has that done to your overall pace of advertiser growth and spend per advertiser?
You know, implementing privacy-safe measurement tools. I would say generally speaking, Pinterest ad platform has been nascent relative to others. There's a tailwind for us just in sort of catching up to where others have been, just because it's been a younger, you know, more nascent ad platform. As we're implementing privacy-safe measurement tools like conversion APIs, like clean rooms, we're seeing those be really effective. The adoption on those is accelerating. You know, when we've seen people implement our conversion APIs, and these are large, sophisticated advertisers, as they implement our conversion APIs, you know, they're seeing 28% lift on attributed conversions. If you're an advertiser, that's huge.
Mm-hmm.
That's 20% lift in attributed conversions, you know, as we're moving to privacy safe tools. That's one example of how much opportunity there is, and we've gone from, you know, hundreds of advertisers on that to thousands of advertisers getting to that, and we still have a lot more to go.
Mm-hmm.
You know, we're making the core user experience better, making the monetization, you know, the ad formats, the intent to action better, and then making the measurability better for the advertiser, and there's sort of a compounding effect of those things all happening together. Last thing I'd say on this point on ad tech and ad stack, I think that we're, you know, as an industry, we are still relatively early in the journey to privacy safe measurement and privacy safe advertising. you know, Apple has made their changes, and that was a big shoe to fall when that happened. you know, Google has spoken publicly about the changes that will come for Chrome and for Android.
Yep.
They haven't happened yet. When they do, that will be another really big shoe to fall. When that happens, I think you'll see a much more clear dichotomy between platforms that have intent expressed directly on the platform versus what were the platforms that didn't have intent expressed on the platform and were really relying on sort of tracking people around the web.
Mm-hmm
to know what they're interested in. For us, we're on the high intent side of that platform. To state it very clearly, Pinterest doesn't need to track you around the web to know what you're interested in. People come to Pinterest and tell us what they're interested in, you know, I think that advantage is shining through more now that we're really investing in the ad platform, really investing in the intent to action. I think as you have, you know, Chrome and Android start to implement their changes with cookies, you know, going away, I think that is going to be not just from an ad measurement perspective, everybody's in the same boat with the industry rewiring to privacy-safe ad measurement tools. Everybody's in the same boat on that.
The industry is gonna go through a rewiring to conversion APIs and clean rooms and things like that. Where everybody's not in the same boat is on the ad relevancy, where it's really gonna matter, do you know the user directly through your own first-party data? Do you know the users? Do you have intent expressed directly on your platform? Platforms that are relying on tracking users around the web to know what they're interested in, I think there's another really big shoe to fall. , you know, part of why I'm excited, you know, about what we can do and how we can continue to differentiate as an ad platform is we don't need to follow the users around the web to know what they're interested in. They express their intent directly on our platform.
Okay. Another one on adding advertisers to the platform. I know you've spoken in the past about third-party networks and advertising networks that other platforms use to introduce more advertisers into the auction market. Just sort of a two-parter. One, sort of remind us philosophically about how you think about partnering with some of the networks to add more bid liquidity. How should we think about the impact to your bid liquidity if you were to add some of these partners to the overall auction market?
Yeah. Yeah. Well, first thing I'd say, pretty fundamental shift in our business From where we are now versus where we were two quarters ago. Two quarters ago, we were supply constrained. We grew supply by 15%+ in Q4, and we did that while increasing ad relevancy and increasing user engagement. We've laid a lot of foundation with things like whole page optimization that let us go think of ads as relevant content, particularly in commercial context. You know, getting engagement overall to double-digit plus, you know, big supply unlock to say then supply can grow faster than user engagement because we can think of ad as relevant content in a commercial context, lets the supply grow even faster than the overall user engagement.
That puts us in a place now where we're, you know, you know, we're no longer supply constrained. In fact, we have more supply than we have demand, that opens up the possibility to say, Oh, how can we ingest more demand because we have excess supply? Fundamentally different than where we were two quarters ago. When we think about how we bring in more demand, of course, first-party sales is always our preference, and we have a great sales team. We will continue to invest in selling directly. Even the largest, densest auctions in the world augment their auction with third-party demand. There are always gaps in the auction to fill. We're pretty much the only meaningfully sized ad platform out there that I know of that isn't using third-party demand.
Right.
As we bring that in, I think, you know, and we've started some experiments with this. We've talked about retail media networks.
Mm-hmm.
We've already been doing work with retail media networks. We've seen really good results from that. We've not, you know, announced any other specific partners or any other specific deals. What I have said is that I think about third-party demand as a meaningful opportunity for us and one that we can take action on in the near- term.
I think in our last earnings call, I was asked like, Well, hey, Bill, does near term mean like it could be 2024 or could be this year? I think of it within this year in terms of something that, you know, we can start to ingest more of, especially since we already have the proof point with retail media networks, and we are no longer supply constrained, and we've done the work on whole page optimization so that we can think about dynamically flexing up ad load when we have relevant ads to display to users in ways that are gonna be engagement positive. Those were sort of the prerequisites needed to go do those kinds of deals. Again, no specific partners that I'm gonna talk about.
Oh, yeah.
You know, in the long- term, like other platforms, I suspect that we will work with many third-party demand sources.
Mm-hmm.
We've now created a lot of the foundation required to do that. I think of that as a meaningful opportunity and one that again, you know, I think of it as a meaningful opportunity for this year, not something that, you know, we have to wait till next year to see.
Got it. Okay. The opportunity, if we're just sort of thinking about the impact on the ad business, the primary benefit would be it helps bid density, and then theoretically, wouldn't it also help you upsell some of those advertisers to your first-party sales over time, and there are certain verticals of advertising you under-index toward now?
I think, you know, one, it helps with relevancy, to your point. By the way, you know, that's been a focus for us.
Mm-hmm
is to increase the ad relevancy. Part of the. We talked about this with whole page optimization. You know, supply increased 15% plus faster than engagement overall, but we did so while increasing the relevancy of the ads. The more we increase the relevancy of the ads, the more, particularly in commercial context, you can drive, you know, ad load far beyond where we are. I think from the analogs that I would have for the potential ad load for commercial context is a multiple of the ad load that we have today in those commercial contexts. I think, you know, that's sort of how I think about, you know, the medium to long-term side, that opportunity is quite meaningful.
A lot of that we will solve through our own sales, but bringing in third parties I think can help accelerate some of that. To your point of like, well, how do you think about that as a, an on-ramp? You know, that can potentially be an on-ramp, you know, for first-party sales. I think the bigger on-ramp for us on first-party sales is actually gonna be what we're doing around the organic side.
Mm-hmm
Where I spoke about this in past lives of using, you know, the free and organic shopping listings as a way for, you know, retailers to see, oh, my products are actually getting traction, and if I, you know, promoted them, I could get even more.
Mm-hmm.
I think that is the best glide path into more of those direct sales, more of those direct relationships, and so we're leaning heavily into that as well.
Your, your background when you were previously at Google, sort of puts you in a really unique position around AI and sort of, you know, the current zeitgeist that is AI. We were talking about your old presentations you've given.
Yeah. Yeah.
I remember some of them. Maybe just talk to us just to sort of level set, what are some examples at Pinterest where you're using AI, where you think the market may be underappreciating it, and then as you sort of look across the platform, what are some opportunities where you can see continued improvement in either user experience, advertiser experience, or engagement from using some of the AI tools?
Yeah. you know, we are huge not only consumers of, but creators of, AI capabilities, particularly around computer vision. you know, just to give a sense of, like, where we are in terms of the quality of the outputs of that, if you look at things like related items on our platform where you would, you know, look at a product and say, you know, say you're looking at a pair of shoes, and you're like, Well, that's sort of in the direction of the kind of shoes I want, but not exactly the ones. Like, show me ones like it, you know, Pinterest is able to give you really good lateral exploration on that, and the user feedback on that is consistently at 95%+ relevance.
To be able to take just from the image alone and say, Oh, here's a bunch of other items that are comparable to that, but that help you explore laterally, not just exact matches, but ones that are related to it, you know, 95%+ relevance for that is, I think, really fantastic. I was looking at this even before coming to Pinterest and, you know, saying, Okay, that's really, really, you know, leading edge for computer vision. To sort of peel that back a little bit and say, well, what's underneath that, what a lot of people don't realize is, like, these next generation AI capabilities have been advancing sort of, you know, quietly behind the scenes for some time. You know, Google's talked about large language models, you know, powering search for quite some time, and these things.
At Pinterest, you know, similarly, behind the scenes, a lot of our improvements in relevancy and engagement have been through improvements in our AI. Some tangible examples of this, we moved from CPUs to GPUs. For anybody that's sort of, like, deep into this, like, will appreciate the importance of that. Like, what has that unlocked for us going from CPUs to GPUs? Our models have grown by 100x. After switching from CPU to GPU, models have grown 100x. You know, we're processing billions of interactions from users, so, you know, the AI is only as good as the, you know, the inputs that it gets, the user signal that it gets. We get billions of interactions from our users, maybe 50 million+ user interactions per second that we're cataloging now.
When you look at, like, okay, we've got a model that's 100 times larger than what we could've done on CPUs now that we're on GPUs. We get 50 million+ user interactions every second that we can use to train those models. That's what's getting us to these outcomes like, you know, 95%+ relevancy on related products, and that's where, I think Pinterest is quite unique in the ability to solve the sort of walking the bazaar aspect of shopping, which, you know, is the biggest part of shopping in the physical world, but as yet unsolved in the digital world. I think that's, you know, one of our biggest differentiators as we now come in behind that and make it so that not only can you walk the bazaar and discover, but you can take action on it as well.
That's where I see huge go-forward opportunity in the AI behind it. You know, we feel really good about the progress there. I think there are other more nascent things, like Shuffles.
Mm-hmm
Where, you know, we've seen really, really great engagement. You know, Gen Z has been our fastest growing demographic across the platform overall. Shuffles is definitely driving a lot of engagement with Gen Z, but when you look at that experience of being able to say, oh, I can take an image, pull just the one product from that image, and then recompile that with another set of images, that I've pulled, you know, specific products from other images, not only are we seeing that people really enjoy that mode of creation, we've recently started to bring that format from the Shuffles app into the main Pinterest app.
Mm-hmm.
You know, some of the most compelling user engagement is. Some of the most compelling examples of user engagement on that are on commercial use cases where people are putting together outfits or putting together rooms, and each individual object in this collage that's been created is actionable.
Right.
That's behind the scenes. There's a lot of AI and ML that's powering that. Again, I think, you know, the rate of advancement in AI is, you know, quite compelling. You know, there may be some cases where it's a wholly new experience. I think there's a lot more cases where you're getting really great improvement in existing experiences, and particularly on platforms that, you know, one, have good AI and ML capabilities, which Pinterest does, and platforms that have, you know, a proprietary data set, and the user interactions to train those models. I talked about the 50 million user interactions every second. It's also, like, what are the users doing on our platform?
Yeah.
They're curating. You know, you know, the AI is only as good as the inputs that it gets. Like, well, what are users doing on our platform? They're making product associations. You know, which drapes go with this sofa, or what handbag goes with what shoes. The AI can't figure that out by itself.
Right.
You know, we have users curating those things at scale that really feeds our models and why we think we can be quite differentiated here.
All right. Well, it'll be fascinating to continue to watch it evolve.
Mm-hmm.
Bill, thank you very much.
Yeah. Thanks, Brian.
already. Thank you.
Thank you.
Thanks. That was great.