Lenny's Podcast: Product | Career | Growth
00:39:09 · Lenny Rachitsky
[00:39:09] S1: Awesome. This episode is brought to you by Ahrefs. You probably know Ahrefs as one of the leading all in one SEO tools used by companies like Facebook, Uber, Shopify, LinkedIn, Pinterest, and thousands more. But Ahrefs is not just for big companies. With their new Ahrefs webmaster tools, you can optimize your personal website like a professional for free. Can scan your website for over 100 common SEO issues that might be hurting your performance in search engines, plus get advice on how to fix those errors. You can have it automatically browse your website's internal external links and get actionable insights from your backlink profiles. And you can learn what keywords your website ranks for and see how you stack up against your competitors. Visit ahrefs.com/awt and start improving your website's visibility. That's ahrefs.com/awt. Okay. So shifting to our last topic that I was excited to chat about is around network effects and marketplaces and SaaS and kind of how those things connect. I feel like you have the clearest way of thinking about and explaining network effects. So maybe just to start, can you just, like, simply explain what is a network effect and then the different types of network effects that exist?
00:40:20 · Casey Winters
[00:40:20] S0: Yeah. No pressure after you say that. Right? Yeah. So at its core, a network effect is when a product or a business gets better with more users or customers using it. And there's three types that I tend to focus on. So the first and probably the most well known is called direct network effects, and that's when every additional user makes the product better for all the existing users. So when someone joins WhatsApp, other people can talk to them that they... That wouldn't be able to talk to them before, and that makes WhatsApp valuable... More valuable for everyone. And then there are cross side network effects where there's two distinct types of users, and adding an additional user on one side of the network makes it more valuable for all the types of users on the other side of the network and vice versa. So when a restaurant joins Grubhub, it creates more selection for users to order food. And when more users start ordering more food, Grubhub becomes more attractive for restaurants to join so they can make more money from, you know, delivery orders. The the last network effect I focus on is data network effects. And that's when the quality or cost of something improves as more data is collected through the product. So when you save content to Pinterest, to a board, it gives Pinterest signal on the quality of that content, its relevance to other pieces of content since they're shared to the same board, and more signal on your preferences. And this allows Pinterest to better recommend more content to you as well as to other people who look like you and share similar interests. So, one thing that's tricky about these network effects is most people talk about network effects in the context of social network, not marketplaces or, you know, SaaS, like my background. And the thing that's really tricky in understanding social networks is they're traditionally described as these direct network effect businesses. But all of them have to become either cross side and or data network effect businesses over time. The only companies that stay direct network effect businesses are these pure communication tools like Messenger, iMessage, WhatsApp, and there isn't a clear path to make money with those. But you wanna sell ads? You're becoming a cross side network effect business. And, also, creators versus consumers creates a cross site network effect business.
00:42:36 · Casey Winters, Lenny Rachitsky
If you wanna get better at personalization, guess what? You're recommending content by leveraging, you know, data network effects. [00:42:44] S1: I love that. No. You you nailed it. I imagine Marketplace founders listening to this may be wondering, hey. How do I... Can I add on these additional network effects? Can I evolve towards one of these? And what you're saying is not only can you, you need to eventually. [00:42:58] S0: I agree. Right? I think, you know, they're a great form of defensibility, first off. Off. So do they make sense for every business? No. But in a lot of cases, in order for you to evolve properly and continue to grow, they can become [00:43:09] S1: a necessity. Are there any other examples of Marketplace companies that come to mind that did a great job evolving or adding one of these network effects or just, like, interesting? [00:43:18] S0: Let's take some inventory. Right? So I think Zillow is really interesting in that a real estate listing site is not particularly an interesting product, but the ability of saying, very rarely do, you know, both sides of this network, people who are buying homes and selling homes, need to connect. But I can find a way for the brand to connect with both of them by giving real time pricing information on the value of homes. And that creates a much stickier product for both sides. I think that was a a pretty genuine, innovation, you know, that other people have tried and and rarely been successful with. So that's that's one that definitely, sticks with me, you know, off the top of my head. Great.
00:43:18 · Casey Winters, Lenny Rachitsky
[00:43:18] S0: Let's take some inventory. Right? So I think Zillow is really interesting in that a real estate listing site is not particularly an interesting product, but the ability of saying, very rarely do, you know, both sides of this network, people who are buying homes and selling homes, need to connect. But I can find a way for the brand to connect with both of them by giving real time pricing information on the value of homes. And that creates a much stickier product for both sides. I think that was a a pretty genuine, innovation, you know, that other people have tried and and rarely been successful with. So that's that's one that definitely, sticks with me, you know, off the top of my head. Great. [00:44:05] S1: We've been talking about Grubhub a lot, and Grubhub had some pretty strong network effects and Yep. In large part thanks to the work you did. And one of the benefits of network effects is it creates a barrier to entry. It's hard to replicate and compete with a company where they already have all this network effect. But famously, Grubhub got disrupted by DoorDash and Uber Eats. And I need to talk about how that happened and why that happened. And it'd be cool to hear just, like, what do you think they could have done differently and what happened where their network effect got eaten eaten for lunch, pun intended?
00:44:34 · Casey Winters
[00:44:34] S0: Sure. It's a great question, and I think people do mistake any form of network effect as this perfect form of defensibility. They're the best form of defensibility, but that doesn't mean they're immune to disruption. And I think the main way this happens with cross side network effects that we typically talk about for marketplaces is when a disruptor dramatically expands selection. So I wanna be clear. I left Grubhub at the end of twenty thirteen before said lunch was evil. So what I'm communicating is more like public knowledge of watching these amazing companies compete rather than, like, being inside the fire at the time. Great. One important thing to remember is Grubhub was an asset light marketplace model. Restaurants did their own delivery through their own delivery drivers that they hired. And I wrote this post for, you know, Andreessen Horowitz a few years ago about supply strategies and marketplaces. So, in that, like, you really need to be comprehensive or have exclusive inventory. And, Grubhub strive to be comprehensive. We'd show every restaurant that delivered to you, even if we didn't work with them directly for online ordering. So what happened about a year before I left Grubhub was that these competitors started to raise their C rounds or series As to build delivery networks. Postmates and and DoorDash were the the first two that rose to prominence. Uber Eats would come later, but they actually started with a pretty different business model. And this is an extremely different business. You're hiring drivers. You're managing logistics. So we call this a heavily managed marketplace model as you are now facilitating the transaction versus just connecting buyers and sellers and taking payments like like Grubhub did. So on the one hand, you have Grubhub, this extremely high margin business, and these businesses coming up are actually negative margin businesses, at least they were for a long time. So in 2013, Grubhub acquired Seamless, its closest competitor at the time. And, you know, Seamless still had a lead on Grubhub in New York partially because it had this corporate program where law firms and banks and consultants got lunch stipends you had to redeem by ordering the food through Seamless.
00:46:47 · Casey Winters
So Grubhub operated in these dense cities and and suburbs like like like in New York because those were the only places that had restaurants that had delivery. DoorDash, when it started, operated in less dense suburbs and worked with restaurants that had never done delivery before, and they would provide, you know, the delivery drivers themselves. So this allowed DoorDash to grow without much competition in the early days. And these companies also took some real gambles. One of the things they did is they delivered from restaurants they didn't have an agreement with, and that caused controversy and lawsuits. But what it meant is when these companies did launch in the cities to go after Grubhub, it felt like DoorDash and Postmates and Uber had dramatically larger selection of restaurants to choose from, and Grubhub was definitely no longer near comprehensive. And, you know, back when we talked about cross site network effects, the more selection of restaurants, the more attractive it is to users, you know, and vice versa. So at that point, you know, Grubhub's put in a very peculiar position. It had gone public after raising $80,000,000 in venture capital and raising a $100,000,000 in an IPO, telling investors it's an extremely profitable, high growth, asset light marketplace. But now it's got disruption pressure from operationally intensive negative margin delivery services that were raising a $100,000,000 every six months from the private markets. So what Grubhub assumed is that these businesses were just structurally unprofitable and that VCs would stop subsidizing them eventually. Of course, that didn't happen. They kept raising more money. DoorDash eventually raised, you know, multiple billions of dollars in in this market. And part of what they used that money for was to lock up agreements with national chains, which Grubhub never worked with. It was always like the local mom and pops and promotions and discounts on the demand side. So all of this is happening, and then the pandemic hits. And all of those negative margins turn positive for DoorDash for the first time, if I understand things correctly. And on top of that, all this corporate ordering that we got from Seamless no longer mattered because no one goes to the office anymore. So this gambit really paid off, and DoorDash just took the market.