Off-Platform: Building Tools to Discover the World

Phillip: [00:00:00] Hello and welcome to Future Commerce, the podcast about cutting edge and next generation commerce. I'm Phillip.‍

Brian: [00:00:04] And I'm Brian. And today we have a very special guest with us, Jeremy King, Senior Vice President of Engineering at Pinterest. Welcome, Jeremy.‍

Jeremy: [00:00:13] Thank you very much. Great to be here.‍

Brian: [00:00:15] We're excited to have you.‍

Phillip: [00:00:17] Glad to have you. Yeah. Pinterest. Pinterest. Congrats on the move. ‍

Jeremy: [00:00:23] Yeah. Just four months now. So brand new.‍

Phillip: [00:00:26] That's great. now. Good for you. You've come up on the show before, but for people who may not be familiar, could you introduce yourself? Who is Jeremy? What's your story?‍

Jeremy: [00:00:38] Thanks. Yeah. My name's Jeremy King. I'm the head of engineering at Pinterest. I've just been here a few months. And I but I spent, previous to this, I spent eight years running technology at Walmart, ran a team called Walmart Labs and helped Walmart go through their digital transformation. Previous to that, I spent some time at a couple of startups, and then I was at eBay for about seven years and went through sort of a re architecture and going through their massive scale back in 2000 through 2008. So I've been around the valley for a long time.‍

Brian: [00:01:19] I love eBay. I love the classic eBay story. It's awesome. When I first heard your talk at NRF, I was like, oh, this guy's got something interesting to say. Some of the takeaways from what happened at Walmart, what you were able to accomplish there. How did you how did you end up at Walmart? What led to that?‍

Jeremy: [00:01:46] You know, it's a great story. I was at a small startup called Liveops, and it was sort of cratering because we had to split the company into two pieces because of a number of challenges with investors and where the market was going. So I decided to move on to somewhere else. And frankly, I got this this random call. And, you know, most of my jobs around the industry have been through referrals from other friends, and this one truly was a recruiter that I happened to pick up the phone. You know, in my business, you usually don't pick up the phone because it's usually a recruiter or a vendor trying to sell something. And this woman asked me, "How would you like to work for a Fortune Five retailer and work on e-commerce business? I literally had to look up who is in the Fortune Five. You can't underestimate... Walmart's been in technology news for the last few years now. But you think about eight years ago, literally, there were crickets. There isn't literally nothing going on. They didn't talk about it and they were really a vendor specific. So with incredible support, I love... And frankly, it was one of those things that I didn't know whether it was gonna be a six month job and I was going to run away in frustration or I'd be there for years. And with incredible support from the top, Mike Duke as the CEO of the time, and then Doug McMillan, had just incredible support for technology and transformation. And as a result, we built Wal-Mart Labs with some incredibly talented people around Silicon Valley, and then incorporated the Arkansas teams and the India teams together into a single organization that is now, I would say a world class retail technology company. They've always been really good at technology, but often were really good at integrating vendor solutions and scaling them. But now they're really a builder and can not only use open source technology, but build their own on technology with the best of them. So really proud of the team and you know, I've made up a ton of friends. And now there's a big ex-eBay Mafia team there and all that, as well. So it's a great talented group around the world now.‍

Phillip: [00:04:01] A few months ago, we had Wayfair on the show.‍

Jeremy: [00:04:06] Yeah.‍

Phillip: [00:04:06] And so we were talking about what open source meant to Wayfair. In the midst of that conversation, my little segue was it's amazing how adoption of open source and internal engineering culture can change the entire perception of a brand. And when I think of Walmart and what Walmart Labs was able to accomplish and its broad contribution back to engineering and open source, that's what reoriented my understanding of what Walmart does. And I might... I'm in the minority from an engineering background. So those things matter to me. But when I see, you know, I think most people equate the broader consumer reorientation around something like jet.com and all the other DTC acquisitions. I think that leadership from a technology perspective is the other part of the equation. And so I think that it's a monumental task to take on. I'm not sure what you if you have anything to say about that. Is there a philosophy for you around leading engineering with an open culture?‍

Jeremy: [00:05:18] Well, you could argue that it's all a talent play. But I mean, where do you go find the best engineers? I mean, Wal-Mart sort of got to the point where, from a scaling standpoint alone, you couldn't... Most companies don't build products for Walmart scale. And so every product we build, we'd have to customize or rip the guts out of it or, you know, figure out what the you know, how to scale the underlying technology plays. And so once you start doing that, then you're sort of bound by, you know, another company in that sort of thing. But and so you sort of make the leap. "OK, I'm going to have to get a whole bunch of engineers anyway to rip these things apart. Why don't we use a technology that's relatively open and not only figure out how to scale it, but make it secret sauce...only our capabilities there." And so I think the engineers that are strong, and this is a generalization, of course, but want to not only use the latest, greatest technology, but also contribute back. And one of the biggest advises I give to folks who say, "Oh, we want to use open source technology," I ask them, "So how many of your engineers have contributed back to the open source tech stack?" And often I hear crickets. "We don't have an open source contribution policy." Like come on, guys. This is the kind of thing that generates great engineering talent and good references to the world. The best engineers know who are contributing to these platforms and that sort of thing. So that's that's one piece. The second piece is making sure that you use technology, you know, just going open source for every single product across the technology stack isn't a great idea. There's lots of great vendors out there that have great solutions for you. That isn't your secret sauce. So, you know, don't try and build everything, but pick those things that are secret sauce for you and build it. And you need to have a great team for that. And oftentimes that leads to open technologies across the board. Right? So that's the advice I give them. You know, oftentimes people are sort of going one way or the other. And it's really a hybrid approach anyway. You know, even integration between vendor products can be connected together with open source tech. So there's a lot to think about there. And most companies, you know, these days have to have a new respect for how much technology plays in there and their success or failure.‍

Phillip: [00:07:49] Especially those that are digitally native. Right? ‍

Jeremy: [00:07:52] Yeah. I'm more talking about companies that are going through transformations, right? Absolutely. Most digitally native ones have already respected that. And I'm sure there's a ton that have grown up on sort of the digital commerce platforms and then go, "Oh my gosh, we're getting really big. What are we going to do?" You know, "These bills from these vendors are getting too big," or, "We're starting to out scale it or we want new features that the vendor doesn't supply. Now, what are we going to do?" So I advise a lot of companies on how to take that next step, as well, once they get big and they need something more specific.‍

Brian: [00:08:27] Yeah, I think you hit on something really important there, which is if you are using technologies that are not open source, you're typically beholden to the direction and the decisions of a single organization. Whereas if you go with an open source technology, you have a community of people around it that can help steer it towards the end goal. And it's ultimately sort of a stronger play and a more long lasting play and ultimately a more enterprise play than a closed source technology. I think, with that, like you said, pulling in tech that's SaaS or closed or whatever, and finding the right spots for it, like that's huge. I think that every retailer should be thinking about how this.‍

Phillip: [00:09:22] Yeah, this hybrid approach.‍

Jeremy: [00:09:24] Yeah. And I like to say, it's hard to outsource innovation. Right? So pick the places where you want to be innovative and insource those things, and then outsource all the other things that are not your secret sauce or not going to be your innovation platforms.‍

Brian: [00:09:38] That reminds me of what Chris Homer, the CTO over at that ThredUp. He said the same thing when we had him on the show, that you need to try things internally first and then you can always pull in a vendor and pull in the right technology or whatever afterwards. But run smart, thoughtful experiments, internal first.‍

Phillip: [00:10:06] There's an interesting... So you take all of that background and sort of the scale of Walmart and you bring that to your new role over at Pinterest. I'm curious about how you make the move to a company that's not retail centric, or at least isn't thought of as retail centric, and does that give you any a unique perspective to launch a new channel for what is effectively, I think, considered to be just a social media platform? I wouldn't say "just." It's one of the largest social media platforms.‍

Jeremy: [00:10:38] Well, we'll talk about social media versus discovery, in a sec. But I agree with your point. And frankly, when I decided to leave Walmart, I was there for a long time. And I had a great team. But based on my conversation earlier about, you know, life changes and things I wanted to do, moving to San Francisco and that sort of thing. I started looking at boards and started talking to VCs about other companies that I could help advise. And I got introduced to Ben and the Pinterest team. And what I really saw at Pinterest was not only just, you know, a great company, but a unique opportunity to scale the team, build technology that drives value for both the retailers and shoppers on a product that people already love. The brand is really synonymous with inspiration and positivity. And there's a lot of pride that comes with that in the company and externally. And we just announced that 300 million people come to Pinterest every month. And so I knew that we were going to have scaling issues. And how to how do we grow this grow this thing? And so you can imagine our goals or not only to continue to grow and scale us, but also to make it more shopable. And I knew, given my background, long history and shopping, you know, I could make Pinterest more a shopable. And that's the number one thing pinners say about the platform is, "I found this product that I truly love. Where do I get it?" Right? And that's the challenge that I'm here to try and help.‍

Phillip: [00:12:19] That sounds like it's the kind of challenge that someone like you is uniquely poised to be able to rise up to that challenge. There are very few people, I think, that could rise up to lead that. We've talked a lot about "audience first." Which I think we would... And maybe this is a bit naive of us or ignorant. I'd love to hear the distinction between a discovery platform and social media. What do you think you're doing at Pinterest that you have a natural tie in to the decision around acquiring a product versus just discovering it? Like, that shop versus buy mentality?‍

Jeremy: [00:12:59] Yes, sure. I think it's important to think of Pinterest as more of a discovery engine than social media. You know, people come to Pinterest to discover ideas for themselves and not really to share with friends. There are some places where people will share travel ideas or recipes or, you know, I want some fashion advice for one or two people. But in general, we're not a social media platform. They come to the platform with a mindset saying positive, future oriented focus is on their self. It's really a different commercial opportunity for retailers. And people have always come to Pinterest for shopping inspiration. And as a result, we don't really need to create new behaviors. When you're on one of the social media platforms, you're sort of interrupting their social with a commerce injection, and they spend tons of time on how to make it non disruptive. Pinterest is a discovery engine. And oftentimes, as I mentioned, people find something they love and like, "OK, I want to move to either build or buy mode." And we really are trying to use computer vision and search to focus on taste based personalization. So you'll see great ideas based on similar results. But styles and we're constantly making the shopping technology smarter using this computer vision technology. And it's, if you've been using Pinterest lately, there's lots of sort of pinch and zoom capabilities. So you can take a picture of a whole room and zoom in on a particular chair and it'll pull out all the chairs that look like it. It's a wonderful way to sort of deepen your searching and shopping experience.‍

Phillip: [00:14:50] I personally have used Pinterest to communicate with my stylist at Stitch Fix, which completes like some sort of a bingo card somewhere about this sort of like the modern fashion and brand discovery process.

‍Jeremy: [00:15:10] It's amazing. We heard of a whole movie set, a TV set, where the producer was using Pinterest to effectively build every set. And these kind of things are amazing. And they're collaborating with, you know, small groups of people. It's not sharing this to the world. It's a collaboration. I hear about people sharing it with their architect or their designer for that. People with fashion advice, people talking about travel tips... Hey, I took a European trip with some friends this summer, And we had boards for every city we were going to. Here are the restaurants, here are the places we want to visit... And it's a great platform for that type of small team collaboration.‍

Brian: [00:15:51] So it's interesting. So clearly, you know, Pinterest has been a discovery platform  since its inception, and you've been sort of building out the commerce capabilities as you go. And so where would you say that the evolution of Pinterest's view of its role in actual purchasing has been? When you started, it feels like it was a little bit less focused on commerce, a lot more on discovery, and now we're definitely in a place where you're very commerce focused. And then where would you say you're heading?‍

Jeremy: [00:16:31] Yeah, it's a great question. And back to my point about the number one request from pinners. So we talk often talked about "pinner first" mentality inside the company. Pinners want to know, when they're done with the discovery part of their of their journey, and they want to now either build or buy, we need to help them out. The good news is we spent years working with retailers and shoppers to understand how they work. And as matter of fact, as you can imagine, I've been talking to the Walmarts and the Home Depots and the Lowes and all the Targets of the world as I've come into Pinterest, and oftentimes retailers are surprised to find that they already have somewhere between 10 and 50% of their products already in Pinterest because people have pinned the things that they like, and their catalogs are already largely in Pinterest. So we really think that mobile shopping is really still in its infancy. And retailers are looking for ways, you know, as I was at Walmart, we were looking for ways to reach new customers. And there're tons of places to shop when you know specifically what you want. I know I exactly want this brand and this target. You can go and find it and everyone knows where to go. But if you don't know and you're looking for inspiration, where do you go? Personalized recommendations. Their options are pretty limited.‍

Phillip: [00:17:54] The department store, Jeremy. {laughter}‍

Brian: [00:17:56] I was thinking the exact same thing. We just talked about how the department store is sort of been replaced by discovery platforms like Pinterest. It's amazing.‍

Jeremy: [00:18:10] Well you even think like you're getting catalogs in the mail still... Like why are people getting catalogs in the mail still? Because they're still trying to build inspiration. And that's what people come to Pinterest for. They're planning a home remodel, or a new look, or outfit for a wedding they're coming up to, or are they're gonna buy a new car, or things like that. You know, content from brands in this space and ads are additive at this point and even welcomed, you know, a not intrusive like they are on other platforms. And we've made the discovery and the shopping experience seamless so that you're not being interrupted very much when you're not in this sort of shopping. So we have things like catalogs and product pins and things like Shop the Look, so you can shop for pins. When you when you save a pin that you like, you can zoom in on a blouse or a skirt or a tire on your car. You like these rims? Where do I get these rims? You know, and you can buy from that retailer and just a few clicks. So that's sort of the the way we think about it. It's how to discover first and then once you found what you love, go get it right. Or go build it, too, which is great.‍

Phillip: [00:19:20] And I think you have an interesting challenge there in that, unless you can correct me if I'm wrong, Pinterest is not the merchant of record in a lot of those transactions. It's a marketplace of various companies, brands and manufacturers that can fulfill those sorts of orders. And so that marketplaces is a very non-trivial challenge to have to solve.‍

Jeremy: [00:19:47] Yeah, luckily I built a couple of those, so...‍

Phillip: [00:19:51] Yeah, one or two. ‍

Jeremy: [00:19:52] You're exactly right. We're talking about hundreds of millions of pins that are catalog items, actually, and finding things like dupes and things like that. Like do I want to show a hundred of the same items versus showing slightly different versions of them? And that's where visual search in our tech comes into play. But it's also helping small companies get discovered. 97% of our top 1000 searches on the platform are unbranded. So think about... "I want a new coffee table." "I want a new convertible or a sports car," or "I want a white blouse," or "I want a groomsmen tie," or, "tuxedo" and things like that... This is a place where those companies can now introduce themselves to consumers for the first time often. Right? And in a platform like other retailers, you're limited to the brands, of course, that they're showing. And so this is a place for where retailers can really get discovered by very specific use cases across the board. And we have all kinds of hobbyists, like woodworkers, and artists, and boy, there's dozens, if not hundreds, of places where people can get discovered by new brands.‍

Brian: [00:21:09] Phillip. Phillip, I feel like this is the ultimate passive shopping.‍

Phillip: [00:21:14] Yeah.‍

Brian: [00:21:15] People are... So there's this idea on the show, Jeremy, that we've talked about quite a bit, called passive shopping, where people are like sort of browsing passively and then something will sort of kick them into gear to make a purchase, whether it be the right offer displayed at the right time, or something happens in their household where they have to make the purchase, or whatever it is. And I'm just thinking about how you've described shopping on Pinterest. And it seems like there may be occasions where someone will go very actively to Pinterest to make a purchase. But oftentimes it sounds like they want to do a home remodel and they're looking for the perfect thing. And then when they see that thing finally pop up, they find it on Pinterest, they're like, "Oh, my gosh, that's it. I'm gonna make a purchase now."‍

Phillip: [00:22:02] But I sense that there is something different that happens here. So in the passive commerce model, which shout out to Jason L. Baptiste for coining the term, and I think we had Richard Kestenbaum talk about the Viant study at some point here that talked about that. A lot of that is in two ends of the spectrum, it's like highly considered purchases where, "I want to buy a bike, but I don't know what bike to buy and I'm going to think about it for a very long time." What I think Jeremy's also talking about is that there is a unique opportunity of "I'm always looking for the next sort of shirt dress." I'm not looking for a shirt dress. {laughter} But like, "I'm a big fan of shirt dresses, and I'm always looking for short dresses." And that purchase mentality might happen many times in my engagement. It's not just one considered purchase, which I think is a unique experience because any other shopping experience would rely on the conversion at the moment of the product consideration. Pinterest happens to be a place where people go to be inspired all the time.‍

Brian: [00:23:10] Yeah. Good point.‍

Phillip: [00:23:11] You know, it's super interesting.

‍Brian: [00:23:13] Inspiration shopping.‍

Jeremy: [00:23:13] Yeah. And this is what's so great about Pinterest because it's so customizable. And you go look at people's boards... And it's so funny when I tell people I work at Pinterest, there's so many times, I say eight out of ten times, I get like, "Oh, my God, I love Pinterest. Show me your boards." Right? "If you don't mind, you know, show me your boards." And it's all about... One woman I just met recently, had a board she called "Sparkly Things." One of our boards that she used all the time. And it's funny. And and it turned out it was full of a bunch of things that were like from the ocean, that sort of thing. And it also had a bunch of sort of heavy metal rocker guys in it. And I was like, "What?" And she's like, "These are things that make me sparkle." {laughter} And I was amazed at how she was using this every day, and she had maybe 10 boards that had been pinned in the last 24 hours. So she's an active user on this thing. I can't tell you how many people I run into that show me this kind of thing. It's amazing.‍

Phillip: [00:24:11] So we have a very diverse audience that listens to Future Commerce. And, you know, we have a very global audience. We have people at many ends of the spectrum who are in Fortune 50 retailers who are also digitally native brands. I'm curious what kind of advice you can give from your vantage point about criteria and making decisions around technology and maybe even around the team outfit and culture and your perspective on having built that a few times. What kind of advice could you give to them?‍

Jeremy: [00:24:46] Oh, boy. You know, it's funny. I'm going to write a blog post on this recently. And the premise of the blog post is that, you know, I worked for a 300 person startup before I joined the world's largest company. And I tried to bring the startup mentality into Walmart. As a matter of fact, I used to call it the world's largest startup, Walmart Labs was the world's largest startup. And I was successfully able to do that inside of Walmart and in various areas, including, you know, just the environment, culture and the open source tech as we talked about. But even things like product mindset,  customer first mentality, and these sorts of things. And then the hitch on the blog post is that now that I've come to Pinterest, what are all the things that I forgot that I should have done? And so now that I've been at Pinterest, there's a number of things. But I can't tell you the foundation of things that make a company really technical in mindset are things like access to data, for example. 61% of Pinterest employees have done a query in our big data system in the last 60 days. That means everyone in the company has access to our data systems and questions are answered with data not by, "Oh, I think we should do this. I think we should do that." It's like, "Oh, no, the data shows that this is working, and this is not working." Right? And that's HR, that's legal, that's the product teams, that's the executive teams, you know, everybody has access to the data and use it on a regular basis. And this is one of the tenants that we were trying to put in and did in a lot of ways successfully put in at Walmart. So that's one thing. The other thing I talk about is just basic culture things. Do you have access to... Are your engineers and your business people sitting together, for example? Or you have some engineering building that's miles away or maybe in a different building and nobody ever talks to each other? You know, the ability for you to iterate quickly on ideas that the business have and to react quickly, it helps a lot when you're sitting next to each other, or your product people are sitting next to your business people. If you have your product and engineering people in one city and your business people in another, you better make some real hard choices about how you do collaboration. Via weekly video meetings and things like that. So I talk a lot about those thing. So there's about 100 others, and I'll put a few and in a blog post I'm going to publish shortly, so...‍

Brian: [00:27:25] Well, we'll definitely have to repost that. ‍

Jeremy: [00:27:26] Yeah.‍

Phillip: [00:27:28] I'm curious too... You're around that and team building. You touched on... Or we touched on open source before. Are there any sort of new initiatives or new trends in team building, especially in the engineering side that you're looking to implement? I know there's a big remote work culture that has developed in the past few years like.‍

Jeremy: [00:27:54] Yeah.‍

Phillip: [00:27:54] What is your take on that? And I didn't prepare you for this question in any way whatsoever, so...‍

Jeremy: [00:28:02] Oh, it's fine. If you look at my background, Liveops was all about building a culture around remote work. The company had 20,000 people that basically worked from home for, you know, hundreds of corporations. And so the whole idea was to effectively build this remote work place. And back to my point, and one of the things I am going to write in the blog post, is one of the things I not only forgot about how great collaboration tools are, but they're only great if everybody in the company is in it. At Walmart, we implemented Slack in the engineering group, and is great. But at Pinterest, not only are using Google Docs, which is phenomenonal, and I'd forgotten how great it is, but Slack is used by everyone in the company. So I can get to my product leader, I can get to the CEO, I can get the legal, I can get that everybody in one second. And in corporations, oftentimes, you have one team in one tool and one team in another tool... And I forgot how fast, you know, decisions can be made if you can just get connected to everyone. You don't have to wait for the phone call or the email back or version 5 of the PowerPoint you attached your email, and things like that. And it's a relief that I'm in a company that works this way, and I'll pressure companies to make that hard choice, because it is hard to make those kind of changes in a corporation.‍

Brian: [00:29:28] It's interesting. You're using technology to bring together your company. I love that. And I feel like Pinterest does the same thing for pinners, and what you're really doing is you're building communities on Pinterest, and I think it's amazing. And so I guess, sometimes we've seen technology that can sort of maybe push people apart. You know, how do you see technology enhancing communities or does technology really kind of create an illusion of community? How do you see technology playing into community building? And then relate that back to commerce. How does commerce fit into that?‍

Jeremy: [00:30:07] Yeah. Great. And this is one of the reasons I was inspired to join Pinterest. Our goal is really for us to find inspiration, so you can bring an idea to life in the real world. Makers, right? We don't want to keep you online all day. We want you to find great idea to do work, and oftentimes it is with a community, and make it happen. They'll go to a maker space, or they're trying to learn how to reupholster something.... And that sort of thing. You get out... We were talking about people who've rebuilt motorcycles and they're now into motorcycle clubs, and these sorts of things. Community on Pinterest happens when people get inspiration from those overlapping interests, not necessarily in a social way. These ideas that people find can strengthen their real world communities. And we've heard just awesome stories about teachers creating lesson plans for their classrooms and new hobbies, new careers. It's a great place to build community and again, off platform. Leave and go discover the world.‍

Brian: [00:31:13] Off platform. That's really interesting. So what you're saying is Pinterest is inspiring community, not just on Pinterest, but actually in real life. That's so cool.‍

Jeremy: [00:31:24] Yeah. We want you to go do stuff. Right? You know, and that's the... Ben... I remember my first dinner with Ben, and that was one of the hook points for me. It's like, get out there and go do things. Like use your hands. I'm a big sort of home improvement kind of person. I did a lot of work on cars when I was a kid. And that sort of thing. I want to go out and do things with my hands. So it's great.‍

Brian: [00:31:48] Totally.‍

Phillip: [00:31:49] You mean that it's Ben Silbermann, right? The CEO.‍

Jeremy: [00:31:51] Yeah. Ben Silbermann. CEO.‍

Brian: [00:31:53] I actually... I've remodeled my shed into my studio office, so I'm totally on board. I love it. I love it. I think you're right. I think Pinterest really does a lot of inspiration for actual projects, which is so cool. It would be so cool... Like now I'm getting visions of Pinterest shared workshops down the road someday, where people can go into their community and build stuff. That would be really cool.‍

Jeremy: [00:32:24] Yeah, absolutely. ‍

Phillip: [00:32:25] It's sort of like the antidote to what I think a lot of these sort of digital wellbeing mindsets are right now, which there's like an open question of the constant plugged in nature, especially for younger kids. I know it's something I track in my household, and I'm very aware of it. This summer, my youngest daughter, she 7, discovered that we had plastic instruments from a bygone era stuck under the bed. She found my old Guitar Hero and Rock Band plastic instruments, and they've been wanting to wake up at 5:30/6:00 in the morning every day this summer to play fake plastic instruments. I cannot wait for school to start because as fun as that is,  and I'm a musician, I want my kids to be musicians, but I also am very aware of the fact that they're engaging more with technology. Their schoolwork happens on screens. Everything happens on screens. And I'm very concerned about it. It's interesting to hear that other companies, especially those that are commerce oriented, are focused on trying to help people to be assistive in their life, to have real world experiences, and not just to be glued or stickied to the platform for dopamine hits.‍

Jeremy: [00:33:47] Yeah.

‍Phillip: [00:33:47] I think a lot of people are very concerned about that right now, and I commend you guys for championing that.‍

Jeremy: [00:33:54] Yeah, it's great. It's, like I said, one of the main reasons I chose to come here. So it's a great part of the platform and people just love this platform. So you love working in a place that people love your product.‍

Phillip: [00:34:05] Yeah, for sure.‍

Brian: [00:34:06] So true.‍

Phillip: [00:34:07] And speaking of love, you're a big public company now. Right? That's a big part of the story. You have to be careful what you say nowadays.‍

Jeremy: [00:34:19] No kidding. ‍

Phillip: [00:34:19] But I think that story is very, very timely. And we really appreciate all your time today. We always ask our guests, and I thought I'd... And we are called Future Commerce. We would love to hear your prediction, maybe not just at Pinterest, but what do you think the next five years looks like from a technology and commerce perspective? I'll seed it a little bit to maybe help you orient your brain. You know, some people are very concerned about, what deep fakes might mean. Some people are very excited about voice. Some people are really always looking at the way that technology can be used to enhance the customer experience. Everybody wants an experience nowadays. I'm curious what you think. What does Jeremy think?‍

Jeremy: [00:35:10] Yeah, I spent a lot of time thinking about this. These disruptions that are going to help. In our case, you know, I used to spend a lot of time thinking about more physical problems that would help dramatically lower technology costs in the retail stores and things like wireless power and the ability to integrate robotics more. But I'm now back to into more of a digital view of the world. And frankly, when I look at what makes Pinterest great, it's really all the machine learning and tech that we use. 80% of the expense of Pinterest is run on CPUs, which is unusual in most retail environments where you have a lot of storage of data and that sort of thing. When you come to the home feed on on Pinterest, we're jamming hundreds of calculations a second to to make sure you're getting the right experience that is personalized for you in real time. And so I'm very interested in what's happening in the TPU or the the tensor processing unit, the GPU space that's allowing us to use compute even faster and more real time. So spending time with the technology players in this space to make that experience. And that's really, as you can imagine, pushing us harder on the machine learning platform KDD, which is the sort of mecca of machine learning, is happening right now. And one of our chief scientists got an award, along with a few of his co-writers, on a long running paper that they used to talk about machine learning. And so companies are investing heavily in this space, and it's getting easier in some respects to use machine learning for your properties. But at this kind of scale, it's getting even harder, especially when you're looking at the breadth that we have. One person's pin might be because they like a very specific part of a pin and another person might be completely different. Right? So how do you... So you have to make it very specific.‍

Phillip: [00:37:22] Context, right?‍

Jeremy: [00:37:24] Yeah. And to your point about customer experience, it's all about making a beautiful pinner experience. And you know, Ben and Evan, who are the co-founders of the company, are obsessed with the beauty of the platform. This is why people like this platform, it's a beautiful set of UIs sitting on this massive machine learning engine. So I think about that a lot. Machine learning and consumer experience and personalization more than ever.‍

Phillip: [00:37:52] I'm curious what you think about... You talked about the decision at Wal-Mart Labs, the build versus buy sort of conundrum in that there was a decision that you have to build. There's no one who builds what you do at that scale to suit. So some of that...it's not an out of the box thing. Do you have that similar conversation around machine learning/machine vision investments at Pinterest and platform partnerships?‍

Jeremy: [00:38:21] Yes. Yeah. Right now, the cloud providers we're using... We use a very generic sort of compute platform. We're not using their their ML engines or their algos. All that's developed internally. And I think, again, we've chosen...that's our secret sauce. You know, we need that in order to provide that experience that's unique to our consumers. I think it will get easier and easier. I've talked to the other... The cloud providers are trying to build generic machine learning models that will help people optimize things like supply chain, and merchandising, and pricing, and that sort of thing.‍

Phillip: [00:38:58] Right.‍

Jeremy: [00:38:58] People have been trying to do that for years. But I think if you can do that... But those kind of things might not be your secret sauce, you know. So great, outsource those things to third parties. Awesome. And then figure out what the things that you're that you're doing are unique to you.

‍Brian: [00:39:16] One criticism we've heard of personalization based on AI, and we talked about a little bit on the show, but not a lot actually, is the concern that you get sucked in to a set of recommendations and almost like a filter for any given platform that's so based upon your previous experiences that it doesn't enable you to get out there and try something new. Have you thought about this and thought about how to factor in sort of getting sucked into personalization holes, if you will?‍

Jeremy: [00:39:54] Absolutely. This is a huge, huge, important factor for Pinterest, as you can imagine. To some extent, at eBay used to call this serendipity, which is how do you sort of figure out when to plop something in that is not relevant? You know what I mean. It's a relevant score. The relevance teams go crazy because that is not relevant. But it turns out it is relevant. And this is what's great about Pinterest. Pinterest is made of a bunch of boards. So when you pin something, you pin it to a board. So I might pin my bike shot to my mountain biking board. Somebody else might take my exact same pin and talk about the trees that are in the background, or the sky that they like, or maybe they're looking for a very specific...they like my bike seater, the color combination of something like that. And they'll put landscape photos, or they'll put mountain bike, awesome mountain bike, pictures or something like that. And we use that, essentially, to train the machine learning model. And so as a result, we can see what people are inspired by in a totally different way than if we were doing it on our own. So in a lot of ways, the more pinners we get, the more inspiration we get, and the more angles we get from a variance from a single pin. It can be looked at in 20 different ways. And that way we can add some serendipity with the machine sort of controlling it, if you know what I mean.‍

Brian: [00:41:32] So yeah that's interesting. There is a social component here. It obviously it is a discovery platform, but the social component really, I think, allows for that variety. And as people contribute more and more to the platform, it will enhance everyone else's experience, as well. That's interesting.‍

Jeremy: [00:41:51] That's absolutely right. Yeah, it's like a big tagging engine. We have a giant machine, a human led tagging engine that helps our algorithms get better and better.‍

Phillip: [00:42:05] Yeah, not to depersonalize the human experience, but you've got a really effective Turk platform there.‍

Jeremy: [00:42:12] Yeah. ‍

Brian: [00:42:12] {laughter}‍

Phillip: [00:42:12] That's really what you have, and they're highly engaged.‍

Jeremy: [00:42:16] Yeah. Absolutely.‍

Phillip: [00:42:16] And in many cases you have experts of individual industries and categories which you probably don't have in any other shopping platform. They're contributing in some way. I think it's super interesting. I'm gonna have to think about...‍

Jeremy: [00:42:33] Absolutely. I was talking to a pinner that builds mobile home designs. And she was actively talking about how our ranking works and how she gets raed. And I've just been picking up the phone and calling pinners just because I want to learn more about it. She's out in North Carolina. She lives in a mobile home park. She's helps people design their mobile homes when they're moving in, and it's just amazing. You know, all the different things you do for small spaces and what do you do with curtains, and all this kind of stuff. It's incredible. So there's an angle on every single thing you can you can think of on Pinterest.‍

Phillip: [00:43:06] Let me pick at that for a second. So, I had a conversation not two days ago... This is very timely for me. I had a conversation with someone that I do some consulting for and they said, "We fulfill two to three thousand orders an hour. And people keep telling us we need to talk to our customers more. So we implemented an on site chat." And I said, "But that's a robot. Everybody knows it's a robot. You need to talk to your customers and you need to pick up the phone and talk to them." And they were like, "How do we possibly do that?" I said, "You get a person to pick up the phone and just call them. That's what you do." And they said, "But we can't ever do that at scale." I'm like, "Well, you're never going to do it, period, with that kind of an attitude." You need to know what your customers need and what they want. And I'm curious... When you hear someone, from your perspective, doing product or product business discovery, or some sort of like a use case and research, really, internal research for yourself... How do you operationalize that? Do you operationalize that? Or is that something that Jeremy King does?‍

Jeremy: [00:44:20] This is funny. I learned this from Meg Whitman back in eBay days, when we got to... I forget what milestone we hit. It was some big milestone. And she had a great idea. She gave every employee in the company five customers. I think it was a 10 person list, but you had to call five of them. And it was really just to thank them for using eBay. And the stories we got back from people were just incredible. Number one, they all couldn't believe that somebody from eBay was calling. And that's what I find, when I call from when I was at Walmart or Pinterest or wherever. When I pick up a phone, I have to kind of convince that I actually am from Pinterest, but they will unload on you in good and bad things. And I happen to know CEOs all around the world now that I've worked both from eBay and Walmart days. And I told them just give me the scoop. How is Pinterest working for you as an advertiser and that sort of thing? You got to... And your point is right, it's hard to scale. But when you find something, just pick up the phone and call them. It's not that big of a deal. And once you get over that, you'll be shocked at how much great data you can get. The customer perspective is super important. Ben and Evan have been going on what they call a "pinner visits" and they'll go visit pretty heavily user pinners in their homes all around the U.S. and around the world. I forget where they're going next. I think it's France. And they'll go visit folks in their homes and sit down with them and talk to them about what they use it for, and that sort of thing. It's a fantastic idea for any company to get in touch with their customers and talk to live. So I'd highly recommend it. Yeah.‍

Phillip: [00:45:57] And that... When you're thinking about actionable sort of advice, I think anybody in any size organization could apply their personal touch and make it personal with the customer regardless of scale.‍

Brian: [00:46:12] So I'm thinking back to something you said about, you know, machine learning and AI sort of being the future for Pinterest. And I'm curious. So I've noticed you guys have been putting out quite a few... Even within the four months that you've been there, you've already started to really kick out some really cool shopping features. And looking ahead, you know, to three to five years from now... What kind of cadence are you looking at releasing features related to today? I mean, it's kind of a broad question. I'm curious. I think you're in a really good release rhythm right now. And so I'm curious how you're going to maintain that. How are you going to apply that to this new technology?‍

Jeremy: [00:46:54] Yeah. Our vision really for shopping is to be able to buy anything you see on Pinterest or get a personalized recommendation for something just like it. So we're investing heavily in making Pinterest more shoppable, again in the shoppable areas, not the discovery areas, as well as innovating in visual search and surfacing the right content to the right people at the right time. Just last week we introduced some new features to help pinners scroll through in stock products alongside of pins, and we've launched catalogs and shopping ads for retailers a few months ago. And in the next few months, you'll see some more. And we're talking about, you know, adding hundreds of millions of items that are from catalogs all over the world. And again, not always just to show buy, but to be organic pens as well. And oftentimes people... I'll tell a company, "We want your entire catalog." And it's not the same as the social media platform, where they're just sort of "Here's our spring styles" or "Here's the latest, greatest stuff." No, we want your entire catalog, but you probably already have 20% of your items already being pinned by somebody else. So we want to pull your entire catalog in, so they could become pins that can be inspired organically before we even get him to discussing ads. So as we grow and more and more people come to Pinterest for inspiration, we take this very seriously. And like I said, our number one request is, "Once I find something that I love, where where can I get it?" So that's the goal for Pinterest over the next couple of years.‍

Phillip: [00:48:32] I think that's a much better answer than my cynical response to Brian, which would have been, "Well, they're a public company now, so every three months you will hear something of some value from Pinterest." {laughter} This has been really great. I'd love to have follow up in another year or so and hear more about the story.‍

Jeremy: [00:48:59] Yeah. Great. ‍

Phillip: [00:49:00] I'm just fascinated. Love this journey. And I always love hearing you share. So thank you so much for coming on the show.‍

Jeremy: [00:49:07] Yeah. Thank you, guys. It's great to meet you guys.‍

Phillip: [00:49:10] Thank you for listening to Future Commerce. And we want you to lend your voice to this conversation. You can do that at FutureCommerce.fm or anywhere podcasts are found. Apple podcasts, Google podcast, Stitcher Premium and Spotify or on any smart speaker device, because voice is not dead, yet. You could do that by saying the phrase "Play Future Commerce." All right. Well, thanks, Jeremy. Thanks for coming on.‍

Brian: [00:49:33] Retail tech moves fast...‍

Phillip: [00:49:34] Future Commerce is moving faster.

‍

‍