Introduction

Patrick: [00:01:21] My guest this week is Sarah Tavel. Sarah is a general partner at Benchmark, working alongside past guests Bill Gurley and Chetan Puttagunta. Sarah has a long history as both an investor and as an operator. She was an early product leader at Pinterest before joining Benchmark. Sarah has become one of my go-to resources for topics like networks, consumer technology, and marketplaces among many other topics. I've used her framework for how to think about client engagement, company data, and marketplace liquidity and quality over and over again in my business life. I'm so excited to finally have her on the show. Please enjoy our conversation.

Insights From The Food Delivery Space

So Sarah, let's start with something that you've been thinking about recently that everyone will understand because they're users of the companies, which is the lessons you've learned from looking at the food delivery companies and just this space generally speaking. Why is this interesting? Why is it so incredibly competitive? Obviously it's a big market. What are the main takeaways that you've learned studying that?

Sarah: [00:02:17] Yeah no, it's such an interesting space. What a case study. It's been the confluence of so many different things that have just been happening in our market more broadly. The first thing that I just, you can't help but reflect on. And I think is something that I think about when founders come to me and they are looking to start a new marketplace, and all they can see is incumbents everywhere and not see where should I start a new marketplace, those types of questions. But you look at food delivery, and you had a dominant incumbent, which was Grubhub/SeamlessWeb at the time that had probably the way that they thought of the market, they had 60% or something of market share. They were the definition of a dominant incumbent.

So why now? I mean just recently, I think it was Vox that released some data that showed that DoorDash had actually eclipsed Grubhub in terms of market share. So it's a very unusual transition to happen in the market. So what was the vulnerability and why did this happen? I think that there was two things that happened in this market. One of them was actually a self-inflicted wound by Grubhub, and the other was a more macro thing happening in the broader market.

So what was a self-inflicted wound? What Grubhub did was that they conceptualized the supply. So restaurants were the supply side for those, I'm sure everybody has ordered on Grubhub. So how Grubhub limited their market was that they limited it to really independent restaurants that could do their own delivery. That was how they conceptualized the market. They didn't aggressively go after training restaurants to do their own delivery, or actually starting to think about how do we give restaurants that don't have their own delivery a means of doing delivery. They didn't really attack that aggressively. So what Postmates, and I think Postmates was the first to really realize was that this was actually an Achilles heel of Grubhub. Because if you could learn from what was happening in a ride sharing space and actually bring a third side of the marketplace into this marketplace, which is the delivery itself, you create this opportunity to onboard all these restaurants that weren't at the time doing delivery. So you could actually leapfrog the availability of the incumbent with a lot more restaurants, and therefore create more liquidity.

So that was a real insight that again, I think it was Postmates that had it. and they, and then Doordash and then Uber Eats went after this Achilles heel incredibly aggressively. One thing that's interesting is that this is something that's played out not just in food delivery, it happened in home sharing. If you think about it, Vrbo and HomeAway were started a decade. I mean, I guess Vrbo was a decade before Airbnb. But Airbnb did a similar move, which is that they figured out a different atomic unit of supply, which let them leapfrog the liquidity of these dominant incumbents. Booking.com did the same thing to Expedia, where they came in with a different cost structure, a fee structure, which let them dramatically increase the supply of hotels. Which again, let them leap frog.

Patrick: [00:05:25] What was that last one? I'm unfamiliar with that last one. So what specifically did Booking do?

Sarah: [00:05:29] So what Booking realized was that they came in and let's call it Expedia was charging 30% rake. Which meant that if you are a hotel, you're a small, independent hotel, you just couldn't afford to list your inventory on the dominant platforms at the time. And what Booking did is they came in and they had a different fee structure, which was a lot lower. Call it 10, 15%. And because it was so much lower, it actually made it economic for all these independent hotels to therefore list their supply on booking. And then again, it leapfrogged the incumbents with more availability, more supply, therefore more liquidity.

Patrick: [00:06:10] I was looking at the a16z 100 top marketplace things that came out a couple of days ago. And Airbnb I think is number one. I was blown away by the market share of the top marketplaces-

Sarah: [00:06:20] I saw that-

Patrick: [00:06:20] The concentration is crazy. I'm curious when you were looking at Hipcamp, and you can describe what Hipcamp is for people that are unfamiliar, how the thinking around Airbnb ... because it seems kind of similar to the story you're talking about, it's like a new unit of supply. So maybe describe your thinking through that investment.

Sarah: [00:06:37] So for those of people who don't know, Hipcamp is building a marketplace. And people will describe it as Airbnb for land, which is just the idea that landowners can post their land on Hipcamp and let people who are either hardcore campers. I don't know about you, not me. I'm a glamper. I need a shower and a roof over my head. So they also have people who will post their lodges, their tents, etc. One of the searches you can do on Hipcamp is you can find a place to stay near a hot spring. That type of specialized search. When you have a marketplace that's picking off a particular vertical, a particular niche, you can create liquidity in a way that the bigger marketplace can't. Because A, you stand for something different. And I can unpack that a little bit. But B, you can create these experiences that are discovery experiences that are more specialized, and therefore create more liquidity for this particular use case.

When you have a marketplace that's picking off a particular vertical, a particular niche, you can create liquidity in a way that the bigger marketplace can't. Because A, you stand for something different. And I can unpack that a little bit. But B, you can create these experiences that are discovery experiences that are more specialized, and therefore create more liquidity for this particular use case.

Another thing you can do on Hipcamp is find a place that has an RV hookup. The question that you always ask yourself when you're looking at these types of companies, and I asked myself this for hip camp was is this going to be a one marketplace kind of market? Is it really just Airbnb? Or can it look like something that happened with Etsy, which was that you had eBay and Etsy, and eBay try to compete with Etsy. But never was really able to own the niche that Etsy had in people's minds, and the engagement model and everything. So you ended up having these two marketplaces. And Etsy, as you know is a very successful public company. So we made the same bet for Hipcamp and their ability to really own this what feels like a niche and expand from there.

Building Marketplaces

Patrick: [00:08:19] We've got probably three or four verticals that we'll explore in this conversation. Since we're already there, I'd love to begin with marketplaces. And I know that this is something you've thought sort of endlessly about. You sort of have this hierarchy in your mind of marketplaces. I'd love for you to lay that out and describe why these make such interesting businesses, and therefore interesting investments.

Sarah: [00:08:37] Yeah. And if you don't mind, I'd love to go back actually to one thing, which is in the food delivery. Which is the macro thing that was happening. Because this is such a big thing, which is think about Grubhub and it's a public company. You know better than anybody, when you're a public company, how you're valued is so different than how private companies are valued. So you are in Grubhub's position, and all of a sudden you have these incumbents coming in your turf. And they're playing by completely different rules. They're playing by a set of rules that you would never even think to play by. They're going over the top for all these restaurants, they're spending money without any thought of contribution margin. It's all about this race to growth.

And so interesting to see how I can imagine being in Grubhub's position and thinking to myself, "All right, well how much money are these private companies actually going to be able to raise in order to kind of keep doing what they're doing?" It just seems so uneconomic. And then you have Doordash raising $2 billion, Postmates raising, I think it's $900 million. I mean, it's just such a different ball game.

And I think the outstanding question, and this I'll I guess segue to the marketplace about hierarchy that you just alluded to, is what equity value is actually going to be created from this incredible amount of spend at the end of the day. I mean, I think we'll see when the public markets vote if they get a chance to on Doordash.

So I raise that because it's interesting to think about one of the things that I see all the time when founders come in is that they're building these marketplaces is that they're running this race to growth. And I think about Postmates here for this race to growth, which is Postmates went headstrong into San Francisco. They went after big cities. They went after a very, very broad definition of what they were building. And they did build a lot of GMV. But the question is, did they build a lot of liquidity? And that ultimately is what you end up building, why a marketplace is so special is because you're building liquidity in a market such that the market can one day actually tip to you because you're building network effects. So I started to really think about this and just reflect on the boards of a few marketplaces now, and just getting to work with Bill and all those things. And start to really think about well given you are building a marketplace and you're a product leader at a marketplace company, what should you be focused on? And how do you maximize your chances of actually building real equity value in one of these companies?

So the way I ended up thinking about that, well the first step always for a marketplace, and this is what everybody, there's a lot of fantastic content out there about this, which is kickstarting the marketplace. How do you get something started? You solve that really difficult chicken or the egg problem. So people like Casey Winters, and Lenny Rachitsky, and Andrew Chen have all read and really great content out here. But to my earlier point, it's not about racing to grow. If your goal of a marketplace is to tip, then you have to start from there and work backwards. And how do you actually think about growing in a way that creates the conditions for tipping? Which is a nuance. And when you think about it that way, what you realize is that there's actually a few things. Just number one, you really have to start with a constrained market. You really, really need to focus. That's a part of the reason why at benchmark we talk about, we don't really care about how big the initial market is. Because there's actually something really, really right about focusing on something really small from the very beginning. Because that gives you the highest probability of getting to network effects really quickly.

At benchmark we talk about, we don't really care about how big the initial market is. Because there's actually something really, really right about focusing on something really small from the very beginning. Because that gives you the highest probability of getting to network effects really quickly.

The second thing is that you want to set yourself up for success. To the extent that you can find greenfield opportunity. So going back to food delivery, you've got Postmates going into San Francisco, Grubhub's already there. There's just a lot of competition in San Francisco. It's a really big city. Not as big as San Francisco would like to think, but it's a big city.

And then you have Doordash and what Doordash did is they went after the suburbs of San Francisco. There was no competition there. It was a pretty hard market for any restaurant to do delivery. So it actually made it a better place for DoorDash to start because they could prove out the core hypothesis and get to liquidity much faster because there was nobody else there that was competing with it. And the third is just driving. You can use product and policies to create that liquidity that you're really driving towards. So have you ever ordered from GOAT?

Source: Sarah Tavel's Medium

Patrick: [00:13:26] No. I'm familiar with it, yeah.

Sarah: [00:13:28] I mean, I love this founding story of GOAT.

Patrick: [00:13:30] What was it? I don't know the story.

Sarah: [00:13:31] The founders were working on another company, and it wasn't working. And I think at this point they were on fumes basically. And they were trying to figure out what the pivot was. And one of the founders had ordered a pair of sneakers off of eBay, opened the box up, they were counterfeit. And he's just like, "Oh, well this sucks." And that was the insight that drove them to start GOAT. And the thing that you can imagine being GOAT. Starting a company, you look at eBay, God knows how much supply they had. But people didn't trust the marketplace. So they used this assurance of authenticity to create trust, which ended up opening up this market in a really special way.

So it just shows you that it's not actually just about supply to create the conditions for liquidity. It's about trust, the product that you create that's delightful. So you're doing all these things. You're creating this playbook that is all the hard things to kickstart your transactions to get that first crank of your flywheel. And then if you figure out enough where basically people aren't leaving you, you're having your cohorts flatten out at some point so that things are starting to make sense. Then you go to my level two, which is about tipping your market. So I know, I mean you're a student of this too. How do you tip a market?

Patrick: [00:14:56] Can you define what that even means, what tipping means?

Sarah: [00:14:59] Sure. So the idea of tipping, and Bill talks, my partner Bill Gurley talks about this quite a bit in the podcast I believe he did with you. But the idea is that you want to get to a place where if you imagine a flywheel, that you're able to get this flywheel to start spinning, and building momentum, and have a network effect start to kick in. So that the market goes from you having to do all the work of pulling people to you to it actually starts to tip towards you. And all of a sudden, instead of you doing all the things that you figured out in your level one playbook that I talked about, they actually start coming to you, and it starts being-

Patrick: [00:15:39] It's like literally tip. Stuff rolling your direction versus you pulling it up the ramp, right?

Sarah: [00:15:43] Yeah. And your CAC starts going down because instead of having to pay to acquire all your users, they just hear about you. You become the game in town. So you want to-

Patrick: [00:15:54] You become a default effect.

Sarah: [00:15:55] You want to become the default. That's a really special place. So step one, at the core, it's a network effect that kicks in. And you really have to understand that. And a flywheel is in a way cliche, but it's such a great articulation of a theoretical question for most marketplaces when they're just starting out. Because they have no network effect. They can see how if this starts to spin, if I start to get more drivers, more consumers, that a network effect will start to drive the flywheel forward. So what this ends up looking like is that you maximize, first you have to identify and maximize these tipping loops that I think about.

So first type of tipping loop is growth. Everybody talks about this, and it's a really big part of building a marketplace. Because what you want is to be able to get to a place where you're a bigger and bigger percentage of whatever market that you're in. That's when the market starts to tip in your direction. It's really a function of how big you are relative to the market that you're in. And that's things like buyer to buyer referrals. It's people leveraging SEO. It's collaborative features. I don't know if you ever had someone share a caviar order with you so you could add, well caviar did this great job of I create an order. And then I have I can invite some friends to order with me. And then they collaborate on the cart. It's very smart feature. Doordash ended up copying it or doing a pretty good job of being inspired by it.

And then there's a second type of loop. I think about it's part of kind of the social product background is retention focused loops. And the nuance with retention focused loops that I think about for marketplaces is that they actually improve the liquidity quality of the marketplace.

Source: Sarah Tavel's Medium

So you think about reputation in a marketplace, which every marketplace out there has reputation some shape or form. As someone has an experience, and writes a review, or rates the supplier in some way, two things happen. Which is number one, good suppliers get more and more of the marketplace's buyers so that the good ones retain better and the bad ones turn out. Which then leads to this virtuous cycle of then more and more buyers end up having a good experience. So you're letting the natural transaction kind of improve the liquidity, improve the quality of the suppliers that you have. And at the same time, you get this win-win. Which is that the suppliers that you want to retain, retain better. And the suppliers that you don't want to because they're not leaning in or they're not doing a good job, churn. And you have buyers, more and more buyers having a better experience.

Patrick: [00:18:40] Have you had experience with different kinds of reputation scores? So there's the obvious ones, one through five rating or something like this. Anything really clever or interesting that you've seen for how to bake reputation into the system?

Sarah: [00:18:53] Yeah. I mean, I think that the most interesting thing is when reputation is in the beginning, it's the marketplace's best guess at what behavior they want to incentivize for their suppliers. But then what you start to see happen with the more sophisticated marketplaces and I'll take Airbnb as an example, is that they actually codify what the behavior is that they want with some kind of certification or badge. So a Superhost on Airbnb is an example. If you're a host and you're responsive, you have an average rating of whatever, 4.9 or I can't remember what exactly it is. There's some other criteria they give you. And these are all criteria that Airbnb probably analyzed the buyer experience, saw which buyers have good experience so they retain higher, and then use that to create the Superhost category. And then all of these hosts then have very clear feedback of this is what I need to achieve in order to be a Superhost. And if I achieve that, then I'm going to have a badge on my profile, which will increase my conversion rate. And I'll also appear higher in search, which will also increase my conversion rate. So you create a behavior that is of the behavior you want and aspirational for the hosts. And then also creates a better consumer experience.

Patrick: [00:20:17] It's almost like a power seller or something on eBay. You want to think about the perfect behavior of an ideal supplier and then work backward from what incentive structure would create that behavior.

Sarah: [00:20:26] That's exactly it.

Patrick: [00:20:27] Interesting.

Sarah: [00:20:28] Search ranking is another kind of tool that marketplaces can use. I touched on it there with the Superhost, but you think about Uber Eats for food delivery. They actually rank the restaurants. In the early days, one of the biggest things was how responsive that restaurant was to the Uber Eats order. And you create this behavior of fast responsiveness, fast delivery, which was an Uber Eats strength relative to the competitors. You make that the behavior, you help restaurants understand that that's the behavior Uber Eats takes into account in order to show up higher in search. And it creates that virtuous cycle of better retention for the restaurants that are leaning in because they get more orders. The buyers get a better experience because they're finding things that are faster to receive, which is better liquidity quality. So your job when you're running these businesses is to identify these loops and maximize them. And if you do it right, and you kind of get that flywheel momentum to spin on itself, that's when you start to best have a chance of having your network effect come alive.

Patrick: [00:21:36] Anything that we've missed in level two of-

Sarah: [00:21:38] Those are the, yeah.

Patrick: [00:21:40] So then what's beyond that, if anything?

Sarah: [00:21:41] So the big thing then is how do you win? And it's about, Schibsted did this analysis of how much bigger, what leads to profitability of a marketplace. And it's all about how much bigger you are than the number two. So the strategies that you use when it comes out running is very dependent on how much bigger you are than number two. Or if you're actually not bigger than number two, do you want to fight that fight? Or do you want to go somewhere else? An example of that is Postmates, where they invested I don't know how much money. Trying to win San Francisco, trying to win New York City. But the only place where they really emerged with some market leadership is Los Angeles. And there's a little bit of if you could go back in time ... and sorry, I shouldn't say that's the only place. But of the top 15 or so geographies that Vox has, yeah. So if you could go back in time and you're Postmates, do you wish you had taken those dollars, and not just dollars, your best people, your energies, and instead of putting it in San Francisco, double down on LA? Yeah. You wish you had done that.

Source: Sarah Tavel's Medium

Patrick: [00:22:47] To generalize all of this, even into a lesson that might be applicable beyond marketplaces. It's almost like in tech and digital businesses, there is no focus that's too narrow to begin.

Sarah: [00:22:56] So true. So interesting, even Chetan mentioned that in the podcast talking about software. Even within software and you think about open source projects, there are these little niches that they start with, and-

Patrick: [00:23:08] They grow so much bigger than you expect them to. Like for example, the shoe marketplaces. This is not something ahead of time I would have been like yeah, it's going to be an amazing business. But shocking how big little marketplaces can be.

Sarah: [00:23:19] And starting with Uber Black cars in San Francisco. You look back. And even Pinterest early days, I remember talking to a bunch of VCs who had confessed that they had passed on Pinterest. And a big part of it was that they thought it was going to be small. And it's always about the takeaway. Starting small, getting something really right. And then expanding from there, you just have a much better track record of success looking at other companies than trying to go big fast.

Patrick: [00:23:49] I love Bill's idea that when you go there early on, it has to feel alive. That if it doesn't feel alive, it's a big problem. And the only way to do that is incredible focus.

Sarah: [00:23:57] Yes. Yes.

Future of the Consumer Social Sector

Patrick: [00:23:59] You mentioned Pinterest, and I'd love to get your take on consumer and consumer socials specifically. So this is kind of something that maybe nobody's thought a lot about because there's some huge winners in this space that are now 10 plus years old. And then everyone spends a lot of time on Instagram, and Twitter, and Facebook, etc. What do you think the state of that part of the venture ecosystem is today? Are there opportunities? Is it again, hyper niche focus versus trying to be the couple of platforms to rule them all?

Sarah: [00:24:25] I'm an optimist on consumer. And I think that some of the lessons that we actually talked about for marketplaces, I can't help but project onto consumer. I mean, you've got these huge incumbents right now. You have obviously the YouTube, you have Twitter, Facebook, big blue. All of the properties that Facebook has. You could even say TikTok is an incumbent now. And of course, Google and Apple. I mean, it's a very tough ecosystem. And yet, in the same way that you wouldn't have necessarily seen that eBay was vulnerable in all these verticals right now, they're getting picked off in a lot of verticals. I think we're going to start to see a lot of the same with YouTube. We actually made an investment in a company that is trying to unbundle some of the beauty category from YouTube, with a different taxonomy and a different way of approaching the problem. We're looking at opportunities in unbundling LinkedIn, disrupting LinkedIn with a more specialized approach. Again, these are all essentially marketplaces.

And then of course in consumer social, I've written about how I think that Facebook is very much this, the Facebook experience first of all isn't going to translate as well to young people. That's why you've got them on Snapchat, and Discord, and playing a lot of games. But then I also think that there's something that is very isolating about most of the social experiences that we have today, the incumbent experiences where they're very much about kind of this lean back, looking at someone else's life instead of actively participating in something online. And I think that's what people hunger for. So looking at gaming and seeing is there the next Facebook actually going to start off as a game, what would that look like? What are the experiences? You see Minecraft, and Roblox, and Fortnite. There's something special happening there. So always, always looking.

Patrick: [00:26:20] Say more about that participation angle. So how did you come across that kind of thought and idea, and what have you learned about people's drive to participate versus observe?

Sarah: [00:26:29] I remember reading when I was at Pinterest, this analysis that Facebook did on their own platform. And they almost admitted to the fact that people didn't feel good about themselves after they used Facebook. And it was only when they actually participated in a conversation, in a chat when they would feel better. And you can't help but look at what is online gaming in many ways? It is not all online gaming, but the ones that are social are these active participating experiences where you're not leaning back and watching Netflix, which is actually kind of almost one of those ideas that people think is sad of someone. Home alone binging on a show. Whereas when you're playing Fortnite, you're leaning in, you're talking to your friends. There's something that's very connective about that. And all humans, what's the Maslow hierarchy? We just want a sense of belonging. And that kind of participation, that engagement creates it. So you see that with all the gaming stuff. And then you see it with Discord and even TikTok. I don't know if you've spent a lot of time in it. It's one of these apps that just-

Patrick: [00:27:45] I had to delete it. It was taking so much freaking time.

Sarah: [00:27:48] It is so amazingly done. And one of the things that you just can't help but see is that it is about, it's not YouTube where people are putting videos and then it's a post. It's actually very participatory, because you see what someone else does, and you meme it, you do it yourself. And it's about that engagement and about that modding of someone else's dance or whatever it may be. And then everybody has a chance of having their thing go viral. I mean, there's something really special to what they've done.

Patrick: [00:28:22] I'd love to hear more about this idea of the LinkedIn problem. I've heard a few people mention this where there's an opportunity to again, get more specific underneath a very general purpose professional identity layer that LinkedIn successfully captured. I know they had the same chicken and egg problems early in their day, and had interesting tactics and strategies to get that network effect going. How would you suggest people attack let's pick out random niche, anything that pops to mind. How would you attack a specific industry vertical if you wanted to build a LinkedIn-like network effect business that was social or identity based?

Sarah: [00:28:57] As a VC, I feel like the product market fit that LinkedIn has for me is perfect. It's everything I could ... it's not a perfect product. There's so many things that you see and you wish you could shake whoever is leading the charge there on the product side. But they've done an exceptional job and it's incredibly useful to me. But as you start to peel away and look at the penetration that LinkedIn has in other categories, it doesn't really make sense if you're a real estate broker to be on LinkedIn. It's just not useful to you. And it's not your network. You're not reaching out cold emailing other brokers. It doesn't make a lot of sense.

So let's take real estate brokers as an example. What do real estate brokers do? Well first of all, they're always sharing the latest listing that they're working on, new clients that they have. They're engaging with each other. They're getting a lot of leads from Zillow or wherever it is. Many of which they've paid for, but they're not actually going to use. And there are these things that happen amongst real estate brokers, where they actually do share leads with each other. And one of them converts. They get a share of the commission, but there's no network, right? There's no formal system. Yeah.

So wouldn't you rather if I had a great lead, but I can't take it because it's in Marin and I work in San Francisco, wouldn't I rather give it to a friend of mine who I know who gives me good leads and have a really frictionless experience to do that? Yes. Let's say there's opportunities in sales, there's opportunities in a lot of vertical industries. We believe in nursing is one of those. There's what you saw RigUp do. They created rich profiles actually for all of the workers. So it became their LinkedIn. We have a company Instawork that has a similar concept around how do you create profiles for these people that don't really have a place for their profile. So absolutely I think there's going to be something there, and we're always on the hunt.

Patrick: [00:30:56] It sounds like this idea, we were talking about Zach Kanter before. A great way to think about software is to just observe the world and find activities that are already happening, but are not yet smoothed over by efficient vertical software, and attack those spaces. Do you think that's a generally good idea? Yeah. What is the most interesting or exciting to you in the consumer space right now

Sarah: [00:31:18] A lot of what we discussed before, which is I do think YouTube is vulnerable, that LinkedIn is vulnerable, that Facebook is vulnerable. And I love the people who are attacking it with different vectors. And there's some things that are starting to happen in social and consumer that are interesting to see. I don't know whether they remove more friction for people or create new catalysts. But even the simplest, the new Apple feature where you can register for an app with Apple's sign-in credentials. I don't know if you've tried it. So much easier. Same with the Apple Pay. The mobile commerce experience has gone from being just impossible to all of a sudden actually working that makes it feel like okay, there's some technical catalyst that might start to make an opportunity work here when it was just too painful before. Even things like it's kind of wonky, but deep learning.There's this app Sway that came out recently where you could dance in front of the camera, and then it would form fit you to a dance. And it's just these kind of funny things where it's just interesting to me to see how founders and entrepreneurs are just experimenting on the margin of the new things that are possible. There's another deep learning example was that app that was the Russian startup.

Patrick: [00:32:39] FaceApp?

Sarah: [00:32:40] Yeah. The face app, right? I mean, but it's all using these same techniques. So does that end up creating a new experience that actually gets people to really want to download it?

Lessons on Scaling a Business

Patrick: [00:32:52] One of the things I was excited to talk about with you is the scaling experience at Pinterest. This is something that I've started to really think a lot about. Hopefully I'll have the problem of having to scale a sort of a software organization. Enterprise one, not a consumer one. But I'm sure a lot of the lessons are similar. You had this awesome quote, I saw your analogy somewhere about an army veteran smoking a cigarette. I love you to tell that little story and sort of describe the primary lessons you learned about scaling while operating within Pinterest.

Sarah: [00:33:19] There's so many. But I'll start with the analogy, which is when you've been lucky enough to be at a company that goes through hyper growth, there's so many lessons that you learn from that experience. And one of the biggest lessons I think is most directly applicable to being a board member for a company is that it's just not always up and to the right. And that every company, even ones that from the outside seem like everything's working, actually when you know how the sausage is being made, there are these moments of intense doubt. And I remember, and I can give countless examples. I remember when we were growing exponentially at Pinterest, and then Facebook shut us off in the newsfeed and acquired Instagram. And all of a sudden, our growth went from being exponential to being really linear. And it was this oh shoot, what do we do now? And it took us six, nine months to figure out a new way to grow. But there are these moments where we're just like, "Well, that could be it folks." I remember another time where we had no head of engineering. The morale for engineering was so low, and you just have these moments. And it feels like when you're inside that-

Patrick: [00:34:30] It's existential, right?

Sarah: [00:34:31] It's essential. You learn over time that if you just keep focused on your first principles, the vision, and you execute through it, you'll get to the other side. And my analogy to that was it almost felt like by the end of my time at Pinterest, I felt like one of those veterans that you see in World War II or Vietnam movies, where the veterans are in the trenches, and they're smoking a cigarette, and then a bomb goes off. All the new recruits that rush for cover, grab their guns, and the veterans are barely moving. They're just smoking their cigarette, because you realize that the small things aren't big things.

And when I think about what makes a great board member, it's being a stabilizing force to your founders during these moments where it may feel to them that this small thing, this bump in the road is existential. And sometimes, they are existential. I mean when we stopped growing at Pinterest, that was an existential threat to the company. But you can just execute. You have to stay focused and you have to execute through it. And I think that's one of the biggest benefits probably of having someone who was a former operator on your board is that they're less sensitive to those bumps in the road, and they know how to execute through them.

Patrick: [00:35:50] So when you think back on your time at Pinterest, what are some of your favorite interesting examples of maybe things that matter more at the hyper scaling part of a businesses era or life cycle, versus maybe the very early days when you're just trying to find something that people would care about in the first place? Is it very different?

Sarah: [00:36:07] Yeah. I mean the problems do change in so many different dimensions. I mean, even just thinking organizationally. When you read about a reorg, it sounds like, "This company had a reorg." It can sound like it's this something's not working. But actually I think about reorgs as kind of like sometimes you have to break a bone to reset it so it grows in the right way.

But actually I think about reorgs as kind of like sometimes you have to break a bone to reset it so it grows in the right way.

And there were so many times at Pinterest, I remember where we would reorg in some way. And we would actually execute better after it. Because as you go through hyper growth, you're growing teams, and it's a little bit like that. We have to defragment after a while and see okay, which teams are having to do a lot of meetings with other teams? Where are we having a breakdown in communication? Which teams aren't really aligned with the key OKRs that we're focused on and how do we realign them? And that's all about execution is either wrong person in the job or wrong org structure. Assuming you have the right vision. That's just thinking about organization.

On the product side, there's always this temptation to build for your users. I sometimes got into trouble when I talked about at Pinterest, when I would insist on ignoring our power users. Because if you want to build a company that is really big, if you build all the things that your power users are emailing you about. And your power users are just your most vocal users are the ones that may you be so lucky to have these people. Because they're your early adopters, they're the biggest believers in what you do. They're the ones who are the evangelists. But they also, because they love your product so much, they're always making requests to your team about new things to build, new features that only they would use themselves. It's a little bit why Excel has become the many, many sheet, many, many option product that it is, is because it's a power user feature. And the challenge is that if you build for those power users, the product gets more and more complex. And that makes it harder and harder to have that next new user come into your product and understand it. So there's that constant temptation of you want to build for the next 100 million users, not optimize for the users that you have right now.

Patrick: [00:38:32] Yeah. It's a fascinating conundrum. Because it feels like with power users, usually you can get faster wins. Because they're going to be happy. They're asking for something.

Sarah: [00:38:40] Just a classic example is one of the power user feature requests we always had was people wanted to rearrange pins on their board, every time this came up. It was the number one requested feature from our users, which is a silly thing to say, because it was probably 0.1% of our users were making this request. But it's still a lot, but it was a symptom of something else not working in the product, which was that we didn't have the ability to search your boards. And when they finally did release it, I think 0.1% of users used it. And it was a really hard feature to build.

Patrick: [00:39:14] What have you seen if anything, I'm always interested in pricing and digital businesses as well. Consumer social, surely Pinterest and companies like that have never typically charged their users. They've indirectly charged them with ads or whatever. Have you seen successful companies, especially in the more niche vertical, you mentioned RigUp as an example, that are more explicitly charging? Even if it's just basically collecting a network of people who may or may not transact in certain ways. What are your thoughts on that as a business model?

Sarah: [00:39:40] It's interesting to see things like the athletic, and ringer that are actually charging subscription models, having subscription models and charging people for content that you would otherwise think is free. I mean, there's so much sports content out there. I kind of can't believe how much sports content there is. And yet, here are two very successful businesses that have done so with subscription revenue models. I see more and more of that happening in consumer. There's another kind of social private network. And maybe I shouldn't mention the name because I don't know if they want to talk about this. But imagine you're a social company trying to charge a subscription too for that. Kind of recognizing hey, you don't want to do ads in this business, members only. And will that become something that's big enough? I think it's interesting because you do see more and more people are willing to pay for things, especially Apple again has made it so easy to do that. Removing friction, removing friction, removing friction. But it does feel like does that end up still capping the opportunity? It certainly hasn't stopped Netflix. So there may be some real opportunity there.

Patrick: [00:40:55] Yeah. It's interesting the payment friction concept with Stripe, and Apple Pay, and everything else opens up a lot of opportunities.

Sarah: [00:41:02] Yeah. This is kind of the history of consumer is that there's usually a technical catalyst that precedes the opportunity, platform change. A new camera on your phone, whatever it may be, it opens up a new vector for a company to create a new experience.

Patrick: [00:41:19] One of the things I've been thinking about, I'm curious if you've got an opinion on this is the role of search costs in all of this. So in some ways, these marketplaces and networks, all these things we've been talking about are nothing more than ways of just making it easier to find the stuff you want, and do so in a high fashion. Do you think that that's a good or useful mental model just to think about where search costs exists for someone finding someone else or something else?

Sarah: [00:41:40] Yeah. Well, what you're articulating is really about liquidity. So yeah, I absolutely think that whatever you can make easier, we're all lazy. Whatever you can make easier, whatever you can remove friction for, you're making a more efficient transaction. And that always opens up an opportunity.

Patrick: [00:41:56] Can you mention what you mean by the concept of a core action in a software or digital business? It ties back into a lot of what we've talked about for sure. But I think this is such a compelling, clarifying idea for business builders out there to identify and then optimize around a core action. Can you describe what that means?

Sarah: [00:42:13] Yeah. I mean so all the time when you have a product, whether it's a social product or a SaaS product, there's so many things that a user can do in your product. And the way I think about it is anytime a user clicks or taps on their phone or the computer, engaging with your product, they're using energy. And you want to direct that energy in a way that creates the most value for the system that you have. And usually, there's an action that's a special action because there's a single thing that you can do in your system that is most correlated with our user retaining. And also creates what I think of as the accruing benefits and mounting loss of your product. It creates a sticky product. It creates something that people want to engage with and stick with. It makes the experience better. And ultimately if you're in a really good position, actually creates your network effect.

And usually, there's an action that's a special action because there's a single thing that you can do in your system that is most correlated with our user retaining. And also creates what I think of as the accruing benefits and mounting loss of your product. It creates a sticky product.

So to give a couple examples of it in Pinterest, look you could do so many things on Pinterest. In the beginning, we were tracking them all. And we were tracking them equally. You could like a pin, you could re-pin, you could create a board, you could follow someone, you could comment. You could just spend a lot of time scrolling through pins. But what was the most important action? What we realized was that it was someone pinning something onto a board. Because if you focus on that action, if you complete that action, there's a 98% chance you're going to come back to Pinterest the next week. You're making our discovery graph even richer, because you're creating a new edge in our graph between the pin that you panned and all the pins on your board. So that helps the discovery experience get better on Pinterest.

Source: Sarah Tavel's Medium

And you're also leaving part of yourself on Pinterest. So you never want to leave it. A new recipe on your recipe board or place you want to go for vacation. So this single action gets all these flywheels, these virtuous loops in motion. So you as a product owner then or as the founder want to be hyper-focused on how you effectuate that metric, and make the trade-offs of okay, if you're a Pinterest and follows go down for an experiment, but re-pins go up, pins go up. Then that's the right trade-off to make. So again as you know, getting a product right is all about focus. And how do you reduce to the very core what's important? And this kind of focusing on the core action is one of those ways.

Patrick: [00:44:49] Can you say more about these beautiful ideas of accruing benefits and mounting losses as being baked into a product design, and how you think about that as an investor? So one that comes immediately to mind is Evernote. I was so reliant on Evernote for a long time. I now use something called Notion. And Notion made it remarkably easy, literally just pour it over the architecture.

Sarah: [00:45:10] I'm a huge Evernote power user.

Patrick: [00:45:12] This is a great example of I would have said they have enormous accruing benefit, mounting loss. I rely on it more and more. It's more and more valuable to me. But then this other company came along and made it basically trivial to port my structure over and then give me new benefits on top of that. So how do you think about that as a company moat, or what are the weaknesses to this accruing benefit idea?

Sarah: [00:45:32] It's super interesting to me that you said about Notion. I might have to try it. I'm also, I'm a huge Evernote user. And it's the type of thing where as you articulated, the more I put into Evernote, the better the experience gets for me. Because I do a search now on Evernote. And I'd be curious actually whether Notion's search is as good as Evernote's. Because that to me, that's where if I were an Evernote product manager, I would have just be making the search ... I mean, the search experience is already good, but it could have got even faster. It could get even better in the OCR. And I would just be spending all my energy getting that feature right. And I think when people complain about Evernote and think about moving onto something else, it's because Evernote has gotten really heavy. And they didn't just keep, it's all about getting just something, one thing really, really, right. And it was just too heavy. It kind of created friction in the experience. And then whenever you create friction, reduces essentially your liquidity with the experience and creates another opening. What Notion did so right that Evernote was never able to figure out was that Notion is more intrinsically collaborative. And Evernote try to shoehorn collaborative features into the experience with collaborative journals, but it never really worked. It felt too heavy. And because of that, it just stopped growing as a company.

Patrick: [00:46:58] I still think the notion of building into your product, something that makes it harder and harder to leave, because you're building up an asset. Basically in any consumer product, think about what asset the user is building or creating. That's the concept I get from your writing. How do you optimize so that that asset of theirs is growing, is compounding? Does that seem like a good rule of thumb?

Sarah: [00:47:19] The two tests that I always suggest is will your users say that the product gets better the more they use it? And will they say that the more they use the product, the more they have to lose by leaving the product? And if you pass those two tests, then you're going to create a really sticky product for those users.

I'll contrast that with an example of companies that haven't been able to do that is anonymous companies. Think about Secret, or Whisper, or Yik Yak. And there are a lot of reasons why those companies haven't succeeded, but part of it is that you had no accruing benefits or mounting loss. Because of anonymity, there's no identity, there's no followers that I accrue. There's nothing that's persistent. So you could delete, I remember having Secret when I was at Pinterest and I would delete it. And someone would say there's some gossip on it, you open it up, and it's exactly the same experience. So that makes it really hard to have sticky users.

Patrick: [00:48:21] We can release this such that it's no longer a stealth company, but you mentioned by email that there's a new company that you're an investor in. I'd love to hear the story there, what they're doing, and what got you interested.

Sarah: [00:48:31] I'm on the board of the company called REKKI, and it's a London-based company. It's actually a wonderful story where the founder, his name is Ronen Givon, was maybe masochistic enough to run a hummus chain in London with a couple of his siblings for eight or 10 years. And the guy was actually a computer science undergrad, and was always, these restaurants are really, really tough businesses and was always trying to figure out ways to improve the profitability of their restaurant. And because he was a computer engineer by trade, like in training, he built all these systems to automate as much of their variable cost as possible. The ordering, whatever they could automate, they built it themselves. They didn't use third parties, but it was still really, really hard to generate consistent profitability.

And finally, they realized that it was actually their supply chain that was the big reason why they weren't able to be profitable. And there were these little tricks that they figured out where if you order six lemons instead of five lemons, it actually ends up being cheaper because you get a crate of lemons instead of buying them individually. Working with different suppliers to figure out the best pricing. There were all these tricks. And once they started to work on that, they were able to build real profitable margins in their business and ended up deciding that that was a core insight. It's one of those earned secrets that people talk about, which is how do we help restaurants by recasting their cost structures in some way, and help them find suppliers for their goods that are better for them? And this is one of those incredibly opaque markets. Rich Barton from Zillow who was a venture partner with us at Benchmark talks about how there's a some generation of marketplaces that they're not necessarily transactional, though I believe REKKI will be. But actually what they do is they set information free that was once opaque.

Patrick: [00:50:35] Power to the people.

Sarah: [00:50:35] Power to the people. Amen. So that's what the opportunity that REKKI is going after, which is that they start, it's classic, solving the chicken or the egg. Started with a tool for the restaurant, the chef in the restaurant. I don't know if you're familiar with this, but really what happens right now when a restaurant owner needs to replenish their supplies is for the most part, they're calling up a supplier 11:30 PM after the kitchen's clean, and they've taken stock of all the things that they need. And they're leaving a voicemail on a supplier's voice message machine. So that in the morning, someone at the supplier will listen to all the voicemails, transcribe the order into their PO system and then their inventory system and do the order. And that's you might imagine not the best way to do things. There's a lot of friction with that process.

So by going after the buyers, which is the restaurants. And they start to use REKKI to place their orders for their products. And it lets REKKI then actually crowdsource the catalogs of all these suppliers. And they're now a double digit percentage market share in London and growing quickly in other cities and other countries. And it's just been super interesting to see because it's one of these places where there have been a lot of companies that have gone after this opportunity. But they've all started with the supplier side instead of the buyer side. So it just shows you that there's still such greenfield opportunity out there. Just because other companies have failed, doesn't mean that there isn't an opportunity. It's just about figuring out the right angle of attack.

Patrick: [00:52:19] I love this idea of legibility and setting information free as the method to create something valuable in an ecosystem. It seems like you could do that in any of the things-

Sarah: [00:52:28] Yeah. I mean, and look at Rich Barton's career. I mean Zillow, Glassdoor, there'll be others. I mean, OpenTable you'd even say was that too.

Patrick: [00:52:37] The information being what's available in terms of reservations?

Sarah: [00:52:39] Yeah. I mean, you would have to call a restaurant up to see if there was a reservation available. And OpenTable just set that information free and reduced all the friction to getting it.

Patrick: [00:52:50] One of the thing that I've struggled with a little bit personally is kind of knowing what data inside of a business to capture and act upon. And I'm curious how this relates in your view to this idea of the core action. How have you seen companies, so many tech companies gather a huge amount of proprietary information, beauty of software or digital as you can track stuff. You kind of know what people are thinking by their actions. How would you encourage people at different stages of a company to think about using their own data that they've generated?

Sarah: [00:53:19] Yeah, I think of it as these phases for data. So in the beginning, data is useful to track. But when you're just getting started and you're trying to find where that fire burns really bright, and finding the product market fit, data isn't going to show you the way really. It's about your first principles, your user research. And I think you want to have a dashboard, but you definitely don't want to be governed by it.

When you're just getting started and you're trying to find where that fire burns really bright, and finding the product market fit, data isn't going to show you the way really. It's about your first principles, your user research.

But then what happens once something starts working is that you start to have to build a data muscle. You have to go from being that founder or product leader that's gut driven, to being one that knows how to make decisions with data. And that actually starts with exactly what you articulated. Which is that usually, the problem that happens is that you've got too much of it. I mean these dashboards, I'm always amazed by how much data people look at right now. And I find it to be not useful. Because really, what you need to figure out is what is, sometimes it's the single thing that matters. What is that single thing, as you allude to. If it's your core action, then what's your weekly active users completing that core action on a cohort basis. And how is that trending over time? Everything else is secondary to that single thing to focus on.

Patrick: [00:54:43] I love that idea you have of everything else being empty calories. That's such a good idea.

Sarah: [00:54:47] And it's just the idea that again, there's so many things and people. I think the most dangerous thing for a founder, and it's contextualized by everything we talked about before, which is how important focus is. Is that if you end up connecting your ego in any way to a metric that isn't actually the most important metric for your business, it's so dangerous because you want that metric to go up. When really, you invest time and energy to make it go up, you report it to people because you think it's a metric that you want to go up. But you're missing attacks on your system, when really you should be focused on the thing that's most core to your business.

Patrick: [00:55:29] And so often, those numbers are bigger absolute numbers than the correction thing.

Sarah: [00:55:33] Exactly. And that's kind of, back to this liquidity thing, that's what I see so many marketplace founders do. They think GMV is a number. They want GMV to go up. But Postmates had really big GMV numbers for a while. But if you're not building leading liquidity, it doesn't matter. You have to get to that core of what is important. And then, you're building the muscle of learning how to make decisions with data and not just your gut. And then that kind of gets you to, so you're figuring out what the most important thing is. And then you start to really be able to make decisions. And almost you have to make decisions using data because your gut as a founder no longer actually scales to the user base that you have. Because you're not catering to the power user anymore. You're catering to this broader group. There's always nuance. That's a difficult thing to balance for a lot of companies, which is using data to make decisions, but not being blinded by that data either because you don't want to end up in a local maxima.

Patrick: [00:56:35] What, if anything, are lessons that you apply or ideas that you apply to SaaS specific investing that we haven't already talked about? Or, is it kind of a lot of the same concepts?

Sarah: [00:56:45] There are so many concepts that are definitely connected to SaaS. Because as we've talked about, it's about getting something really right. It's about making a great experience for your users and solving a real need. It ends up being a lot of what we talked already.

Patrick: [00:56:59] Yeah.

Sarah: [00:57:00] Yeah.

Patrick: [00:57:00] It makes sense. My closing question for everybody I think you know, which is to ask for the kindest thing that anyone's ever done for you.

Sarah: [00:57:06] I think anybody who's taken a bet on me, thank you. And it probably, I mean to the person, we actually were talking about this earlier, Jeremy Levine at Bessemer was the first person who really took a bet on me. And I was coming out of college. I was at a strategy consulting firm that was a startup. And then decided thanks to a friend of mine who mentioned venture capital to me. And I read a book about it. I was like, "That sounds really interesting." And yet I was this person. I was a philosophy major in college. I had always been interested in investing, but had had no experience in investing. And somehow for some reason, Jeremy took a bet on me and hired me at Bessemer Venture Partners. And not only did he take a bet on me and bring me in, but then he ended up taking a very active role in mentoring me through my six years at Bessemer. So I as an investor know now how precious all your time is. And you're always making these decisions on how do you allocate that marginal hour. And to invest that in someone as opposed to meeting another founder, it was a very generous thing to do. And I feel very grateful for that.

Patrick: [00:58:17] I cannot resist the philosophy bait since I studied philosophy in college. What looking back was the most interesting either philosopher that you studied, or philosophical idea that you explored?

Sarah: [00:58:28] That's so easy for me. I mean, I was a Kantian. And in particular, actually Christine Korsgaard was my thesis advisor. And she has this essay. This is the essay that got me to focus on philosophy, which was Creating the Kingdom of Ends.

Patrick: [00:58:43] I recognize the title, but can you describe it?

Sarah: [00:58:45] The idea of Kantian ethics, it's about respecting each person as an end in themselves and never as a means. And that if you choose an end for yourself, your own goal, that I respect that. And what Christine Korsgaard talks about is actually that what a friend is, is adopting your ends as my own. So that if you achieve the goals, the ends that you have for yourself, it actually makes me as happy as if it was my own goal that I achieved it. Reading this essay is what got me to focus on philosophy. I had been doing economics before, and then it just gave me a framework for understanding people and life, and so made in ways that were important to me. And I actually think about all the time now as a board member. Because really, what is an early stage venture investor doing? They're adopting the ends of the founders that they work with, and working to make those ends happen. Which is hopefully, and usually, the success of the couple.

Patrick: [00:59:48] What an awesome, interesting idea to end on. Love it. I can't wait to go read the essay.

Sarah: [00:59:52] I'll send it to you.

Patrick: [00:59:53] I really appreciate all your time. Every time we talk, I learn a ton from you. So thank you so much.

Sarah: [00:59:57] Thank you, Patrick.

Created by @mehakpbatra