Getting Smart About Growth with Casey Winters & Uber's Andrew Chen | Greymatter

Hi everyone, my name is Casey Winters, growth advisor in residence at Greylock. At Greylock, we're always trying to learn from practitioners on how to solve real problems for startups and we like to bring these practitioners into grey matter to jam on how to solve these problems so others can learn as well. This week I'm excited to welcome Andrew Chen and we're going to talk more in depth about his blog posts about how growth is getting harder and some of the reasons why that is. Andrew, welcome. And why don't you introduce yourself?

Hey Casey, thanks for having me. So I am at Uber right now running the rider growth team and I've been at Uber for about two and change years. But before that I ended up working in venture capital at more Davidow Ventures. I've worked on my own startups, advised a bunch of companies, but one of the things that I've been doing is I've been writing a blog for the last 10 years which is about growth and metrics and marketing and all that good stuff.

Yeah, I think that's a little bit understating it, I'd say. It's the blog on growth and it's how a lot of startups or people getting into startups have learned about growth as a concept. You were one of the first bloggers to talk about growth in Silicon Valley and you've kind of watched growth come from this very niche thing only Facebook and a few early stage startups were doing. And now it's kind of gone mainstream. But with that, you've recently blogged that as it's gone mainstream, growth has gotten a lot harder for tech companies. And that's what we want to dig on with you today, is how growth has gotten harder. And what should entrepreneurs, product managers, engineers do in this new environment where everyone's getting smarter about growth, the incumbents are getting smarter about growth. It's not just, oh, I'm the only person thinking about this, therefore I win. Right? So one of the things that I learned a lot from your blog working at a non viral startup at grubhub was the concept of viral growth. So I'd actually love to start there. How have you seen viral growth change in the last few years? Has it become saturated? Is it still a viable way for startups to break out and become long term success stories?

Yeah, just to your point, I think 10 years ago when I arrived in the Bay Area, one of the first things that I noticed was this was sort of like golden age, right? 2007 you had the Facebook platform coming out, iPhone Platform was completely new. There was quite a lot of work still around contact importers for just like websites. That was still working.

Right.

And so you have multiple platforms that are kind of firing on all cylinders. You can get in with something that's pretty easy. You know, I think we all remember how the top apps in the App store were, you know, flashlight apps, because that was like, like so fun right at the time. And so what that means is that for startups that are getting created around that period, that these channels, these platforms are just completely white space and you can just go after them. What that means is initially there's higher response rates because the platforms themselves are growing really fast. That means that you can grow with the platforms. That's awesome as well. And if you fast forward 10 years, one of the things that I've written about is sort of the law of shitty clickthroughs.

Yep. Which.

Which is the idea that every channel, every platform, no matter what, has this decay in the responsiveness. So why does it happen? Well, it's because the platform owners often end up thinking that a lot of this stuff is spam, so they want to get rid of it. There's increased competition in early days of Facebook ads. Really great. If you were a mattress company, you could crush it just on Facebook ads.

Oh, there's so many of them.

And then all of a sudden there's so many of them, they can replicate your growth strategy. It becomes really competitive, you know, then those advantages go away. And so I think as a entrepreneur, what you're always looking for is you're looking for that opening where you can maybe find a channel that's a little smaller, maybe it's a little bit earlier in its cycle. And what that means is that when it's earlier, it's less competitive and you're sort of less weighed down by this law of shitty clickthroughs. And so I think what's happened over the last couple years has been that there's been so much consolidation, especially when it comes to viral growth, where all of a sudden mobile contacts just kind of like they're not working as well as email sort of contact inviting. There's a bunch of reasons for that. We can get into all of that stuff, but the idea is that it makes virality much harder and more competitive. And what that means is that now longer term, it's much harder to break through. So I think there are definitely new interesting channels that are available right now, but there's sort of been this change of guard, I think, especially for Consumer mobile products around this area.

Yeah, I mean, it's hard for me to think of a company that's broken out recently with like invite based growth. Right, right. Well, at Shady Clickthroughs, which is one of my favorite blog posts of yours, it kind of talks about not only that there's this kind of rise and decay over time, but certain channels have a quick rise and a quick death. And certain channels like SEO, like had like a 15 year run and now maybe it's finally on the decline for the emerging company. So it seems like you're saying viral's definitely hit like the decay side and maybe it's hitting the decay side pretty hard. I would love to hear from you, like, are there any pivots on that style of growth that you think are earlier in the life cycle or different types of companies that maybe are still able to leverage this if you know, you can't do it the way Facebook or Snapchat did it five years ago, 10 years ago?

I think one of the areas that I've become the most excited about as far as viral growth is viral growth in the workplace. Because I think what we're starting to see is that there are a lot of interesting trends. First of all, the continued adoption of the Google suite and all the Google technologies means that all of a sudden these things like Google Calendar, Gmail within the workplace, even documents, like if you think about what a document is a collaboration with, all the people who've ever edited and who care about that document are kind of attached to that. Right. And so I think there's a lot of really interesting social graph that are still yet to be sort of converted into the right products. So I think when you look at that, what that means is bottoms up, virality in the workplace using a lot of the techniques that the social apps used. But you know, in addition, sort of refining it for what you need to do in the workplace, which obviously can't be spammy, it needs to help people actually get their job done, you know, in a collaborative way. I think those are all like really interesting things. So instead of mobile contacts on your phone, you're looking at sort of workplace social graph instead of invites that are like, hey, I want to share photos with you. Instead it's like, hey, I just created a project and I want to collaborate with you. You have to tweak each of the value propositions. But I think there's a lot that's still left in making this work. Obviously Dropbox and Slack have done, you know, an awesome job But I would think a lot about, you know, for anyone that's working on products for work, how do you actually think about what you're doing intrinsically as a collaborative workflow? And how does one person who starts using the product end up bringing their entire team? And what are the advantages of, you know, having your whole team on? I think that all of that is super interesting.

Yeah. What's been interesting with those sorts of companies and Slack and Dropbox are certainly great examples of this is traditionally if you started a company like Slack or Dropbox, one of the things you did very early was you hire a huge enterprise sales team. And in both of these examples they were able to just create something that people found organically and then have a viral spread within an organization. With Dropbox it's through documents. And then of course on the consumer side they also had a good viral loop with get more space by sharing, which was fantastic. And in Slack it's just a product you have to use with other people, so the admin that creates it is incentivized to share it. And Slack and Dropbox have created sales teams, but they were able to grow 2 billion dollar businesses essentially without them. And those aren't just the only examples. I think Atlassian was a great example. Twilio grew entirely through developer evangelism and communication from developers at different companies, which I thought was really interesting. You can kind of see if you think about the law of shitty click throughs, enterprise sales has had a very long cycle and maybe is starting to go end up on the decline of that and now viral in consumers, maybe on the down cycle, but actually on B2B it's on the upswing, which is really interesting to see.

Yeah, I think that's right. And I think that's why within a lot of these workplace products, you know, if you just think about the raw number of growth levers that you can work with, it's actually an extremely rich environment because you can do virality, you can do referrals, you can do content marketing, you can do paid acquisition, and then by the way, you can tie all of those loops that you're creating also to sales. Right. And to all the traditional channels as well. Right. And so I think what that does is the more variables and the sort of more levers that you have, the more likely you are to be able to come up with a company and an offering that sort of uniquely taps into a certain set of those levers such that it becomes hard to copy. Right. Because I think a lot of this comes down to, you know, trying to create these proprietary channels that are like, durable and you can, like, scale over time with them.

Yeah. And what's interesting is now that some of those companies have gone into the workplace and have deep penetration, they're actually emerging as platforms for new startups to create value. The Slack platform, I think, is still very early in its life cycle, but I've seen some really interesting growth opportunities there where if the entire team at a company is already on Slack and your admin just onboards your app onto Slack, you can essentially email, like Slack, the entire group. I don't know if that will be the same in five years, but there's an immense amount of freedom to kind of introduce people to your product on top of Slack. And that was a platform that wasn't even available in the workplace four years ago. So there's some really exciting stuff there.

Another unique one that I see at work every day is at Uber. We use Zoom quite a lot.

Oh, right.

It's to the point where if your meeting request doesn't have like a zoom link, the first thing that happens is you get on chat or you hit reply all or whatever and you're like, zoom link, zoom link, zoom link. Right.

Yeah.

You know, what ends up happening is people expect that zoom link to live inside of the meeting request. And so I think that's a really interesting, like, I mean, I don't think they fully optimized it, but I think that's an example of something where they're actually kind of building on the workplace calendar as the channel and they could like double down on that. I mean, they have a Chrome extension. They could basically, you know, do a bunch of interesting, you know, or they could make it so you could plug into your calendar and it could basically add zoom links to like, basically every single event like that. I'm sure we would actually probably do that as an example. I think there's a lot of creative things that are yet to be done.

Yeah, absolutely. Calendar, I think, is a great example. I think that still hasn't been leveraged a ton. Shifting back to consumer. As we've seen viral loops go away, I think a lot of the emerging stories of consumer success over the last five to seven years have actually moved away from the viral loop, but they've still got a level of virality. It's just instead of around a person, inviting another person, it's a piece of content that goes and attracts other people. And I know that was certainly how we grew Pinterest. All of the Pinterest users, they either create content or they bring content in from the rest of the Internet and then they share it with the platform and then they frequently share it to Facebook. And in the case of Google, we would actually share it to Google and make sure Google indexed it.

And.

And that seems to be a viral esque channel that isn't as far in the life cycle. These content loops and Instagram grew entirely off of I create a better photo than my phone used to allow, and it gets shared to Facebook and Twitter. And then people ask why it looks so good and then they get Instagram. And of course, now Instagram's built its own network on top of that, which is its own platform that potentially is used for growth. I think some more recent examples in our portfolio have been Musical Ly and Medium, where Musical Ly was entirely built as a tool so that you could do cool stuff to share to Instagram and Snapchat. And then they just made sure that we got a ton of engagement on Musical Ly as well. So you're like, oh, I should be spending more time on Musical Ly. And they've had a very successful exit recently. And they started as an app where you could easily record yourself with live music to share these other platforms like Instagram and Snapchat. But then users started seeing more engagement on the Musical Ly than they started seeing on Instagram and Snapchat. So then they started spending more time focused on Musical Ly exclusively. And Medium is basically just a tool where you could post stuff and then the entire thing is that you're going to share it to Twitter and Facebook as soon as you're done. And that content has given them hundreds of millions of visitors per month, some of whom end up writing their own stuff. So what are your thoughts on kind of the content loop? Is that peak of the cycle? Is that also on the downswing? Is there still a lot of opportunity left there? How do you see those companies leveraging that working out long term?

Yeah, I think that's a super nuanced and interesting question because I think as we're talking about content loops, really you can kind of divide it into two separate things, Right? So one is there's the actual content creation tools, right? So in the case of Musical Ly or Instagram having some, you know, kind of really interesting new type of content that's being created, it feels like that can now sit fairly distinctly from the network where it lives. Right. And what that means is you can either decide to build your own network as. As a new startup, or you can basically try and leverage some of the networks that exist today with the risk that, of course, if you're too popular, maybe some of the networks decide that they're going to cut off your distribution a little bit. But the part that's so compelling to me from a trend standpoint is that people often talk about the idea that now we live in a world where we're carrying supercomputers in our pocket, which is awesome. But not only that, it's very clear that this is also the ultimate content creation device that's ever been made. Not only does it have camera and you can use your finger and you can type, it also just generates a bunch of passive information on your location and what you're doing, etc. So, as you know, Strava, which is a app that lets you track your running and cycling activity, they have a really interesting way of creating content, which is you're on your bike, you turn on Strava, and then you're taking a path while you're cycling. And what they do is they have a really awesome way that you can then take that map and you can share it to both your Strava feed and you can also share it to Facebook and other social platforms. And that is a really cool example of a really unique kind of content that only makes sense on Strava, that every time I see it in the Facebook feed, it's because somebody ran like a million miles or whatever and everyone gets on it and everyone likes it and comments how awesome it is. And I think that that is something that really can break through the noise because people just naturally respond to it. And so I think what's certainly true is that the originality of the kinds of content you're going to be able to make and the richness is just going to keep going up, up, up, up. I think that's awesome. Right? The part that I think is going to be tricky and super interesting is going to be that once you create the content, it ends up generating traffic on the next step of the content loop in two ways. One is it gets shared to these other platforms. And like we said, there's sort of risks with that. The other one is when you look at Medium, you look at Yelp and Zillow and so on, all of that is SEO. And you're far more qualified than myself to comment on this, but it's something where SEO, still completely important, viable strategy. But now that we are also now 10 plus years into optimizing SEO, it's not the Most straightforward thing, the number of tricks that work is far reduced. Do you have any thoughts on SEO in the content loop?

Yeah, absolutely. Well, just stepping back to the beginning of what you were saying is you have this originality of content that's created and shared to platforms and, you know, I think the SEO is certainly one component of that, but the Facebook platform used to be an amazing platform for that. And with Facebook, that is entirely how Pinterest grew. In the early days, we weren't sharing it to Google, it was just every repin was getting shared to Facebook and that was an amazing source of exponential growth. And then it just went away kind of overnight. And that was when Pinterest kind of figured out that this content was interesting for people on the Google search engine as well. But these platforms are maybe reacting a little bit quicker and deciding do they want to enable the growth of these other startups in a way that I think when we were coming up, we just didn't see them react that way or as aggressively as they did. I mean, it just felt like 1000 startups died overnight when Facebook changed the platform. Fortunately, Pinterest wasn't one of them, but there were a ton of others that were on the SEO side. I think you're starting to see Google act a little bit more like Facebook in the last couple of years. And when you think about Google strategy recently, I think it's pretty easy to understand. It's how do we on mobile where people have less time and they just want to get the thing done, replace 10 blue links that they have to do research on with just the answer or the solution to their problem. And in most cases they're saying we could build that solution ourselves. Why do we need to show some other site? And in some areas like local Yelp has actively said like, hey, this is anti competitive. We used to show up number one for these queries and now there's just a Google Local box in front of us every time. Has that done the work to make sure it's the most relevant result? Probably not. And I think Google's going to continue to do that category by category that people search the most on and eventually they're going to try to cover everything. Now in some case that creates an opportunity for startup because they will buy you to get better at that. But in a lot of times they'll just all of a sudden, oh, Google has a competitive product showing up in front of me on search results. I still think SEO is a channel that a lot of startups can grow on, but I think it used to be let me just create some great landing pages and have unique content out of Golding. Now there's ton of other people that are doing that and a lot of them have thousands more links than you. So you need to have some sort of competitive advantage on how you're going to build up the authority so that Google trusts you. Medium, I think has done a good job of that. They entice so much sharing of their content that even though they're a few year old startup, they have the domain authority that rivals people that have been at this for 15 years. But if you don't have that sort of way of like, how am I going to get a lot of of links to me quickly? It's just a years long strategy to build up that relevance and authority to actually get meaningful traffic. So for startups it's not usually the best strategy. If you get some distribution early on, it then can be a strategy to layer on that provides a lot of long term success. And Pinterest I think is a great example of that. Facebook growth early on built up a lot of attention, a lot of press, a lot of links through that. As that waned away, we kind of built SEO to come in as that second acquisition loop and, and that's still working pretty well for Pinterest today. When startups come to me and they're saying like, oh, I'm going to grow via SEO, I'm like, have you thought about this? Because it takes years and it's really competitive on almost every query. I think one example where it's not is if the queries you're trying to compete for are things that just started being searched. Like if you think about in the early days of cloud computing, it wasn't like Oracle had been ranking for that for 10 years because they didn't know it was a thing. Some of those query spaces kind of reset the playing field a bit. A startup can definitely attack that. But if you're saying like I'm going to go compete for restaurant queries, it's like, good luck, Yelp is there, grubhub is there and they're good at this and they have domain authorities that are insanely high.

So maybe to your point then would you say then that what that means is for a startup that is just getting going rather than thinking about SEO as kind of your primary thing because it's going to take a while slow, right? Instead you really have to have some other channel that you're going after. In the early years, try to use Those to kind of get you to scale, whether that's virality or paid marketing or whatever, and then use that to sort of like ladder into SEO as kind of like a future channel. Right. I think that that's usually what I've been telling people, but you know, tell me.

Yeah, I agree. What I tell a lot of entrepreneurs is there's a Kindle strategy and a fire strategy for distribution and a lot of times they're very different. I think you hope that they could be just the same thing. And the fire strategy is the expansion of the Kindle strategy. But with SEO, it's hard for SEO to be the Kindle that gets you your early users. But then it can be the thing that scales you to a billion dollar company once you have those early users, potentially, and there are some exceptions to that. Like Medium, I think has gotten a lot of traffic from Google pretty early on because they will just build a lot of excitement early even without thinking about like I'm going to completely optimize for SEO. It's not like they built their company around SEO, they just kind of got it for free because they had great content with a lot of links. And I think that would be an interesting thing for us to dig in on a little bit more. Is if you're an early stage startup versus a mid or late stage startup or an established company, what are the new kind of Kindle strategies? Right. If SEO is not going to be that enviro growth is, especially in consumer, not going to be that maybe a little bit more viable in B2B, how do you just get those early users? I mean, is paid acquisition the only option left?

Yeah, I think what you really have to do in the early years is to find places where your audience lives that really is not that competitive in terms of other players that are going after the same group. Right. If you're launching a new product in the cryptocurrency space, like people just announce that stuff on like Reddit, you know, or they just, you know, and end up going, you know, to these forums or mailing lists and, and just really going for these kind of niche audiences. Right. And I think you really have to win with initially these niches and to show that you're able to retain them and then sort of slowly ramp up from there, I think is the way to go. You can certainly go and like buy traffic. Right. In the early years that's also not a bad way to just spend 50 bucks or 100 bucks, start getting people on your landing page, start getting some usage So I think the danger there is to think about that as a somehow scalable thing that you're going to be able to do long term. Right. I think we all love to see the idea that spending money to get acquisition is something that happens as an accelerant to a business that's already working as opposed to something that you're going to make profitable right away. Because I think the reality is that especially in the early, early part of a product, whether you're working on a Marketplace or a SaaS product, you know, something that's able to, to monetize effectively, that you're just not going to be able to monetize that well in the early years compared to when you're able to fully optimize. So because of that, I think a lot of the traffic buying that you might do early on ends up being something that is more about testing your product as opposed to trying to scale it.

Yeah, I think one of the great things about paid acquisition is like if you have an idea, you want to see if you get product market fit for it, just throw up some AdWords or Facebook ads and you'll see if people click on it, you'll see if they use it. And I think it's been a really great tool for kind of validating products. I like the idea also of once you hit product market fit, but you feel like you need to get some sort of liquidity. If you're a marketplace, paid can be a just nice starter engine to get that. And maybe that unlocks some SEO distribution or unlocks some referral distribution. Yeah, I don't love it as kind of the core strategy of the business for distribution, even at scale. And look, certainly some companies have used paid acquisition to grow at scale very successfully, but I would like to see one of those organic loops driving the business and then you as the entrepreneur saying, oh, I'm going to throw a million dollars on top of that. So this loop operates 10% faster. It's going to accelerate this organic loop. But it's not just an arbitrage thing that I'm trying to make sure still is arbitrage 10 years from now. That's right. As Facebook and Google just ratchet up the rates.

That's right. Another way to say this is that in the early years of a startup, what you're really looking for proprietary kind of early high response channels that you can work with and it doesn't really matter whether or not they scale. And so what that means is if it's about content marketing, if it's about getting onto other people's blogs, mailing lists, Reddit forums, all these little things, they actually do add up and that's exactly what you want to see in the early years when you're just trying to find product, market fit and kind of getting your initial traffic. Now once that's working, you're retaining people, it's starting to really click. Then I think that's when you start exploring these effectively, like non proprietary channels that anyone can really plug in. But then you need to show that, okay, it's all working in a way that it's going to accelerate everything as opposed to trying to tap into those too early.

Going back to your thought about niche networks to find that initial audience, I completely agree with that. Do you have any tips and tricks on like, hey, I'm a new startup, I'm doing X. How do I find that niche audience to go after first and how do I find those networks where they're hanging out that I can actually reach them effectively? Do you have any kind of thoughts there?

Yeah, I think we've all gotten used to the idea that okay, you know, you start a new business and the first thing is you need some set of hypotheses on what your product is going to do, what kind of value it's going to create. And, and I love the idea that in addition to that, you need to sort of pretty fully bake your initial go to market strategy as well and make sure that, look, if you build something that's like some random photo sharing app where there aren't niche, passionate communities and you have to go right into the morass with like everybody else, it's like good luck, that's just going to be really hard. Obviously you can try and offer tactically some advice for someone who's in that position, but I think that's often why you want to go after these things that are segments, for example, hyper engaged, kind of like esports, like gamer type people. Right? Like that's a super interesting segment. Really passionate. If you're building tools for that group like Discord initially, how they initially broke into the market, that's super interesting, right? You look at something like Patreon, like The super passionate YouTube influencer kind of community, that's super interesting too and that you can imagine how that might extend beyond that initial foothold. But I think that's why the starting out with a niche is so powerful, is, you know, for that now how do you find that role? You know, I think first and foremost is, I think in the early years. You really do have to deeply understand your customers qualitatively. You know, I think you and I, Casey, we spent so much time like thinking about the metrics and the quantitative part of things. But like when you're just getting your whole thing off the ground, I think there's a really interesting question of like, okay, well if you're targeting, let's say this, the Overwatch community for games, where do they hang out, what tools they use and being able to really map that out in I think a really interesting way and then realizing it's like, oh yeah, okay, cool, I'm going to go after this Discord channel as my initial thing. I think those are the kinds of unique insights that makes the process of getting to market effective.

Yeah, it's funny working with a bunch of the companies in the Greylock portfolio how many times my answer to their question is just talk to users. And that just kind of tells you what you need to know. And I think Silicon Valley for a while just over rotated on. I look at the metrics and that tells me everything. And you know, in growth, it's a combination of qualitative research and the metrics. The metrics tells you what people are doing, they don't necessarily tell you why they're doing it and they don't necessarily tell you what the people who aren't on your product are doing. And that might be the most important thing, especially in the early days. One tool I've tried to use is once you have like, here's the type of person I'm going after, you can essentially build this map of their day and like, here's where they go get coffee, here's their commute, here's the Pandora app that they're listening to, here's where they go for lunch and then just think about like, okay, now that I know exactly what this person does on a daily basis, where can I insert myself that they might be receptive to my message? It's a bit more of a traditional marketing standpoint, but I think really helps some of these early stage startups being like, is there some sort of cheap way that I can get in there and just talk to this person about what I'm up to and see if that actually resonates. And I found myself using this traditional marketing tactic over and over again. I don't even know what it's called honestly, but I'm sure there is some fancy marketing name for it and it's really helped me build a map of the distribution opportunities for these niche networks at the start or even when you're trying to scale and trying to find some of those additional accelerants like we've talked about.

Yeah, I like that. I think often people end up creating these like, customer journeys that are very selfish to the company creating them. Right. It's like all about how people are going to use my product and how are they going to become more engaged over time. Right. You know, the reality, as you're saying, is that they're doing a lot of adjacent things, whether that's like getting a coffee. And then also very importantly when it comes to channels and platforms and that kind of thing is to really think about what are the adjacent products, you know, like if while they're using your product, you know, let's say again, let's go back to a workplace example. What other tabs do they have open?

Right.

Like, what are the other things that they're checking all the time? And you know, you quickly figure out it's like, oh yeah, that's why calendar is important. That's why email's important. That's why like Google Docs can be important. You know, embedding all these like, kinds of ideas inside the product leads to the kinds of breakthrough insights that happen in social. Ten years ago, when you look at like contact importers and you look at the way that like the initial Facebook apps were able to sort of like rapidly grow on the platform, they had these insights that not only did they go like, aha, like people are doing a bunch of this sharing and like, we can use this hook, but in addition to that, they really built a whole workflow around that to make it, you know, really obviously why you should, like, what the customer value proposition was and engaging in it. And so I think that's exactly the kind of thing you want to do as you're thinking about these other platforms.

Yeah, I totally agree. Moving on to another thing that I know you have a lot of experience with, that I'd love to get your perspective on is kind of in between virality and paid acquisition. There's this concept of like the paid referral. And you know, when I started working at grubhub, we tried this and it just felt a bit too early. Like it just, it never really took off as a channel, but it felt like five years later my Twitter feed was just a bunch of people sharing $5 off coupons for different startups. How are you seeing that strategy? How far is that on the law of shitty click through curve? Like, is that still a viable thing for startups? To be thinking about. Should you only think about it if you're a marketplace that makes transactions? Like, what are your overall thoughts there?

This is sort of a evolution, I think, of where mobile virality has kind of taken us over time. We started out with a bunch of stuff happening with email contact importers and like, just inviting people. A lot of the stuff that LinkedIn and Facebook and Twitter and all those guys pioneered. And then all of a sudden it's turned out that within mobile virality doesn't work as well.

Right.

And there's a bunch of interesting reasons for that. One reason is that consumers are just more mature. They're more sensitive with their contact information. They sort of know that phone numbers is this thing that, like, wow, if like a company has all my phone numbers, then they can do things with it that maybe I don't really want.

More so than email?

More so than email. Yeah, exactly. If a product delivers an SMS for you that you didn't intend, not only is that are there you know, legal issues, tcpa, you know, issues around that, but in addition to that, you know, that person's gonna write you back right away and is like, what is this? Right versus email is something. Because it's asynchronous, it's a little bit less sensitive, It's a little. It can be a little bit more FYI. And so what's happened, I think, is within, you know, mobile, all of a sudden, viral growth has gotten more difficult. There's legal issues, as I mentioned. There's, you know, when you use like, Twilio or Nexmo to try and deliver these invites, you have to be really careful about what, what you do, you know, that's really important. So then the question is like, okay, well, where does that demand go? And I think one of the things that people have been doing has been, as you mentioned, sort of this concept of like paid virality, which is incentivizing people using money to actually invite other folks. It's not just money. It can be things that are just, like, clearly valuable. That's an interesting distinction because, you know, famously, if you had to invite someone to Dropbox and you get some free, like, space as part of that, that's not money, but it is this incentivized virality.

It was very motivating for me.

Totally. And so, you know, these ideas have been around for a long time, but I think they've taken even more prominence in a world where we have both a lot of the other channels I think, being less effective, but then also on Top of that, this emergence of a bunch of on demand players, marketplaces, direct to consumer E commerce. I think all of those things have caused, you know, this to be a whole area and now consumers, I think are very used to it. I want to give a couple like nuances about, like folks that are thinking about incentivizing the reality. I think the first thing is that in a lot of cases you really have to think about the kind of customer it's going to give you. Right. Because the customer that you're going to get is going to be a customer that's incentivized by, you know, money. And so from an. If you have a premium brand and all of a sudden you are, you know, using incentivized referrals, then that also means that the audience that's going to come in from that, that channel is probably will perform worse for you as a result. Right. So I think that's like. So when you're thinking about your LTV calculations, you probably need to discount it a little bit.

Right. We certainly saw that at grubhub. With people that came in through any type of promotion, you're essentially selling them on getting a discount on food delivery. Instead of that ordering food online is better. That's right. So then the next time they want to use it, that value prop is actually gone unless you incentivize them again. And then you could get in this whole vicious cycle where you never actually make any money off. Exactly. The person. So I've certainly seen that borne out in the data. I don't think it's necessarily always true.

Yeah, exactly. Yeah, I think it's something to think about. I mean, interestingly enough, the complement to that, if you operate a marketplace and let's say you're doing something for that's a marketplace where you can buy and sell goods, then if you do referrals on the sell side because they're money motivated and that's the whole point that they're on the platform, that means your referral program will actually perform really well for that group. Right. So it's really interesting to think about like the psychology of using incentives and what kind of audience it brings together. So I think that's one part. The second thing I'll say is that you really have to think about cannibalization and fraud being an issue as well. Right. Because all of a sudden you have this money that's part of this thing. And so you want to make sure that should you really do it off of, for example, do you want to Give the inviter money when the invitee ends up signing up or do you want them to complete a series of real transactions? Right. So thinking about CPC versus CPA versus.

That whole thing, was that actually two different people or did someone just invite themselves with a new email address?

That's right. Or maybe you're getting people that you would have gotten anyway. Right. But what ends up happening is just as what E commerce is often told us, like when you show that little promo code and then an empty text field, people will abandon cart and go online and search for like coupon, like whatever. And then like go back and fill that out. That the same thing ends up happening when you make that referral code. A really, really prominent thing, folks that are thinking about referrals is that the structure of your referral ends up being really important. I think we've seen currently a lot of kind of like symmetrical referral programs. So give $10, get $10. And I think that that is a very, very early take on what the future of referrals are going to look like.

And Uber started that way.

That's right. Exactly. That's right. And what we're going to see instead is we're going to see first, you know, a rebalancing. Should it be 10? $10, $10, right. That's one aspect of it. Maybe they should be different numbers. Maybe these numbers should be actually hyper personalized. Maybe the structures themselves, you know, it's like, well, the invitee maybe gets $10, but they have to order a certain number of things in order to qualify. Right. Like there's all this kind of personalization that I think means that it's still early and driving that efficiency is still there. So when we talk about kind of where is paid virality in the law of shitty clickthroughs, I think it's actually pretty early. I think it's a channel where you can always decide to kind of like overpay in order to get it to work.

Right.

But I think in reality, if you're going to compete on making that more efficient, I actually think there's still a lot of legs to it.

Yeah, I think in grubhub times we were pretty anti promotion in general because of what we saw with the lower lifetime values. But we were really excited about was if we could show that a promotion got someone to try something new that increased their lifetime value. We were then very bullish on trying those things out. So we did. GrubHub was the first food ordering app on the App Store. So everyone was still ordering it via the web. But we saw that the people that did migrate to mobile doubled their lifetime value, essentially. So we just gave everyone $10 off their first mobile order to migrate them over. And that was actually very successful. Whereas the generic like get $5 to try Grubhub was a lot less successful because of a lot of those people just didn't come back and do it again.

Part of constructing these referral programs, if you understand that $5 on Grubhub, I think everyone would know exactly what that means, right? That means like cheaper food, right. So that's like a currency that everyone's willing to do. And you know, one more step away from that is something like free storage space. Because you may think, well, do I actually need free storage space? Like, you know, what is that really?

Yep.

And then all the way to, you know, people do referral programs within mobile apps like games, right? And let's say if you refer to somebody, then you'll get a thousand like gold coins for it. Like, what does that even mean? Right. And then the invitee is less excited about that. So I think that's something else to think about, is sort of like, how concrete can you make your actual referral rewards?

One thing I liked about what you said is that it's not that paid virality is really a separate thing. It's just really about understanding what is the motivation that you can create for someone to share. Sometimes that motivation is I'm going to make money. Sometimes that is the product's going to be better for me. Sometimes it's, I need to get this project done. And you are a critical path to doing that. And really stepping back and saying, like, what is the right motivator to get someone to talk about? This is the real high level, abstracted thing you need to learn. And sometimes that answer is going to be you incentivize it. But sometimes the answer is going to be you don't need to at all. You have the right incentive kind of already intrinsically in the user. And I think that's like a critical thing for people to understand.

That's right. And I think you go to that first. Like, I think if you, if you can do it, you want to have organic virality first and foremost.

Right?

So I think for the folks that are, you know, they have a new startup and they're building, you know, kind of these utility tools, right. I think one of the very first things is, you know, if I were to invite a friend, like, why, why does that make this a better product, or why do I do something and I share it with somebody else? Like, why is that better? I think you want to start with that as the fundamentals. In a lot of cases, you have something that's so much a utility that it doesn't make sense to bring other people on. A lot of fintech companies end up using paid acquisition because it's really the only thing that makes sense. It's a private piece of information that you're not going to broadcast to a ton of people. So, for example, Wealthfront, which is an automated investment service, they have a way that you can do referrals such that you can get more of your assets managed for free. But it's not really something that's like, oh, well, you know, let's compare Wealthfront statements, like, that's not going to happen. Right. And so for them, it makes a lot more sense to go with a ton of content marketing, a ton of paid acquisition, and these sort of like paid referrals as the acquisition drivers for their product.

Yeah, that makes sense. When you think about the virality or referral funnel, a lot of people that are trying this out, they just don't know if this can be a big thing for them or not. So as you think about that, there's actually a funnel there. There's like, can I get enough people to share? Will people actually open that email or text message or whatever? And then will people actually redeem it and become good users? What's the most important part of that funnel, I guess would be my question. Is there one that you tend to focus on? And if that's not working, don't bother. How have you thought about that in your career?

Yeah. So I think what ends up happening is that every time you want to optimize your virality, obviously the first thing you're going to do is you're going to break that down into steps.

Right.

And so I'm going to give you the naive version and then I'm going to give you how it actually works. So the naive version, what people used to talk about with viral loops is that you'd basically come in, a new user would come in, you'd get a certain percentage of them to invite their friends, they would send a certain number of invites, and then, as you were saying, there's sort of this redemption rate of how many people then accept the invites and sign up, and then they repeat that process again. And so if you think about it, it's basically a bunch of percentages. Like, if you start out with 1000 users that'll go down because only a certain percentage will invite, then only a certain percentage will give you their contact, address, book access. Right. But the multiplier effect is basically what happens is when like how many do they invite? That's the, the thing that kind of like makes the whole thing more a loop as opposed to a funnel that just goes down. So back in the day, what would happen is you could actually build a viral loop that was like a single shot viral loop, like just a new user comes in, does this thing, and then you can get a viral factor over one. Right, right. That was like an amazing time for extremely spammy email based viral loops. That's a lot of where kind of the viral coefficient, which is kind of the ratio that measures the whole thing, comes into effect. If you can get that viral coefficient to be over one, your viral factor to be over one, then all of a sudden you have this exponential growth. What has really happened though is it turns out that that is not a great strategy to grow your product. Because what ends up happening is you get a ton of acquisition and then people just fall out at the end and there's no engagement, there's no retention. So what ends up happening in real life, I think, is that you never get a viral factor of 1. In fact, what ends up happening is you come in and rather than doing a one shot loop, instead you come in, you use the product, and then a certain set of those people will then invite a couple people, and then you use the product some more, and then you invite another set of people and then you keep going a little bit more and then you invite more people. Right. And so what that means is that all of a sudden, I think in the real world, retention ends up actually being one of the most important factors. Because if you have really great retention, what that means is you have a lot of shots on goal as far as like getting people to complete the viral loop. Now, within the viral loop, what's the most important part? Because we're now in this more constrained environment where you're not inviting 200 people to your service. Instead, what ends up happening is that the acceptance rate of the invites is probably the most important part. So you want to see an acceptance rate that's like let's say over 10% for people who click through and sign up and then go from there. And then if you have great retention, you can eventually convert those people to more virality. But what ends up happening is that instead of viewing virality as this exponential growth curve thing, instead what you should view it as is. You have a really great product, really great retention, and then you have this reality that's kind of kicking off a bunch of frame free organic users and it amplifies the core of what you have as opposed to being the driver. I think 10 years ago when everyone was doing spammy Facebook apps and doing spammy email stuff, you could actually have it literally be the foundational thing that you do. But these days I think you have to be much more careful and sensitive about it.

Yeah, I think you had a really great way of describing it, which is retention creates the acquisition opportunities, whether it's through virality in this case, whether it's through the content that the retained users create, or whether it's through the money that those people actually make you that gives you this opportunity to go bring back more users. Whereas, you know, when I think when we originally were learning about it, it was just like, I actually don't have a product, I just have a viral loop. And that created all these companies that went like straight up and then straight down as they ran out of the base of users to invite. And that is not what you want to be building as an entrepreneur or product manager. You want to be building something that adds long term value. I think that's really smart.

And when you look at a product like LinkedIn versus Instagram, what ends up happening is the more frequency you have, the less prominent you can make your viral call to actions. In LinkedIn's case, because you basically use it for a couple specific use cases. That is exactly the reason why almost every action you take on LinkedIn, somewhere down the line they're like, by the way, do you want to import your address book? Do you want to add people? Right. There's a really good reason for that. And that's I think, perfectly designed for the frequency they have for something like Instagram. They can make it much more subtle. They can also tie it to obviously the Facebook platform. They make it really easy to invite your Facebook friends to Instagram. That's clearly one of the big things that they're working on these days. But because they have this really high frequency, they don't have to be as hardcore about every action leading to that because they know they'll have more chances down the line.

Yeah, that's really interesting. I've never really thought about that framework before. I think how I've thought about it is entrepreneurs, product owners come to me and they're like, I want to talk about referrals and virality. And I'm like, Well, a really good way that you can say if this is going to matter is you take the home screen and you just make it entirely focused on inviting someone and then you see how many people take that action. And if not a lot of people take that action, the rest of the stuff probably isn't going to matter. Right. Like you've made it as aggressive as possible and very few people are doing it. I think you have a much more sophisticated plan that we can now share with.

Thank you.

With everyone.

I mean I think virality is a catch all umbrella term for a lot of things that you can potentially do in the product. Right. Because if you think about the viral port, let's use Dropbox as a good example of this. Right Within Dropbox you could imagine they could have a straight up just like invite your friends to use Dropbox. They could also have a value exchange like get disk space. But also they can have something like share a folder with a colleague and then that person has to install Dropbox in order to open that folder or like an individual file. Or they could even get into specific things like, you know, use case things such as, oh, if you want to view this secure document, you have to sign up in order to interact with it or you want to sign this thing, you want to do an E signature thing like okay, great, you have to then sign up to do it. So Virality ends up being a catch all for basically anything that causes one user to generate another user as a result of using the product. And so I think a lot of the creativity around viral growth is trying to find the right angle such that people are actually excited to do this and see it as part of the value prop as instead of kind of a completely, you know, bland, generic like share this content with friends or like invite your friends. Like I think those especially because we're in a world where a lot of products have done that now. I think consumers are just immunized against it. Like they just, they're, they're just going to auto skip, you know, those kinds of flows.

Yeah, I completely agree. So I want to shift back to paid acquisition a bit. You made a really good point that like paid can be that starter engine and that's not just theoretical. I think more and more of the companies that we're working with, we're seeing that, right, that they're just starting with paid. Like some even without a business model or just paying to acquire users. I would love to hear about some of the trends in general that you're seeing with paid acquisition at Startups and growing companies. And what you think is most interesting.

What you're seeing with paid acquisition is that obviously there's the Google Facebook duopoly, but one of the things that they've done that's awesome is that the amount of kind of like targeting options and the amount of inventory they have makes it really easy. So it used to be with advertising that you actually had to, you know, deal with humans. You know, when you're buying ads it's very different. Now we're obviously all self serve and we use APIs now, which is great. Back in the day if you wanted to actually buy ads on any of these media properties, you had to actually call up a person and you would create an rfp, a request for proposal that would basically say hey, I'm looking to buy a million impressions or a million clicks or whatever and can you please write me back and propose how that's actually going to work? And so you would do that with a whole bunch of different media properties. Very time intensive. It looked like basically each thing was like a one off enterprise deal deal as opposed to, you know, these days it's amazing that you can just like pull out your credit card and in you know, 10 minutes get a campaign running on Facebook and Google that's super targeted and awesome. Basically every startup should allocate a little bit of money towards paid paid marketing just to basically like have that as a channel going. It could be a hundred bucks a day, it could be in a really small amount just to start iterating with it. I think that's really great. The self serve aspect of it means that you can actually learn important product things where you can, you know, try different kinds of ads and headlines and then expose people to different kinds of landing pages and really learn a ton about, you know, the viability of your product that way and the interest level. I think that's awesome. Within certain segments, like if you're going to talk about for example on demand, you know, sort of labor as part of it, right? Supply side like drivers and couriers and that kind of thing. Like there's pretty well established benchmarks at scale for how, how much that signup is going to cost you. And so you kind of like know ahead of time that like hey, if I keep optimizing things I'm going to be able to get to that. You know, same with like a B2B like a workplace like email, right?

Yeah, that's a really good point. I remember, you know, when I started at grubhub, they had you know, some paid acquisition running and it wasn't working particularly well. And I just told, you know, Matt, our CEO, it's like I'm going to try these five channels and I'm going to spend a month trying to optimize them. And I know what our target CPA is for our payback period and I'm not going to hit any of those payback periods in the first month. But there's going to be some where I'm like 100x away and there's going to be some where I may be like triple away. And what I'm going to do is I'm just going to turn off the ones where I'm like 100x away and I'm pretty sure I can optimize the ones where I'm like 3x where I need to be down over more period of time. And that ended up being true. But I think for people just starting out with paid acquisition, a lot of times they're just like, I'm not sure what I should be paying. Here's what I'm paying now. Can I get it down 10x? Can I get it down 100x? They just don't know. And I think certainly one of those steps is kind of like you have to really aggressively try all those targeting options and figure out which ones work for your product and if those are going to be scalable enough to actually go get real users. And there isn't just a really well defined playbook on like here's how that process goes, early stage startup, go do it. And maybe that could be a future blog post of yours.

Yeah, this area is so deep because there really is a learning curve. Right. And I think we're also at a point where it's not that hard to like put in your credit card and like have a couple different kinds of creative and have a couple kinds of campaigns going. But the next level of that, which I think is really interesting, is that if you know that you're going to be a product where paid marketing is going to be a major driver. Right. Like basically if you're a direct to consumer e commerce thing, like you're going to live or die based on your ad spend efficiency. If you're a marketplace, it's going to be the same fintech, these things that are kind of like low frequency, that picking up that intent is really important, that becomes super key. And so what that means is not only is there a learning curve and just as like getting your team going on the first little bit of ad experiments but then beyond that, there's all of a sudden all of those self serve platforms also have really deep APIs that then allow you to do this optimization one level deeper. Like, purely programmatically on, like, which you know exactly what you're talking about. Like which campaigns are going to be on, which ones are going to be off. Like, how long do you let them run before you turn them off? How do you generate, you know, kind of all the permutations? That's really interesting, really exciting, because it means that you can actually invest and use software and software engineers to create a meaningful advantage in your ad buying process.

Yeah, I'd say that's one of the most fascinating transitions I've seen in the last couple years, and you mentioned it previously, which is, okay, paid acquisition is not just a marketing channel to get more users. It actually helps you build a better product by what you learn on the targeting and the messaging and all of that. And then now there's these deep integrations with Facebook, AdWords, et cetera, where you actually not only get product insight that the product and engineering team can leverage, you actually get product and engineering horsepower that can make these things work better. So you're seeing these things that were usually just in the purview of a marketer, now being owned by product and engineering teams for the first time. We've seen this at Airbnb putting real engineering resources behind paid acquisition and wish, which is one of those E commerce examples you mentioned where you can buy very cheap goods from China and just have them shipped to you for low cost if you're willing to wait a couple weeks. They don't have a marketing team. They just have an engineering team that spends a lot of time on their Facebook advertising because it's so critical for them. And I think you'll continue to see that happen. Right. As paid becomes one of the primary channels that people need to scale to win, it becomes more and more core to the DNA of the company. And therefore product engineering design are all thinking about it instead of just saying, oh, that marketer off in the corner is going to take care of it for us.

That's right. When you look at kind of one of the foundational ideas behind this whole new growth function that's emerged over the last couple years, it is this thought that it's not like marketing and product are going to be able to be separate things. And it's like, okay, product, you guys build what you need to build and then marketing will promote it. But instead it becomes highly integrated and maybe it's a little bit like kind of software eats the world. Right. But a lot of the functions that maybe you would have in previous years had dozens of kind of media buyers kind of sending emails, like getting on phones, trying to do all these partnerships can be things that are completely just purely executed by software. And I think that's really exciting future because I think what it means is that you can really invest in it, as you were saying. But the other part is I think it allows startups to have real differentiation as well. Because if then you're adopting some of these APIs or you find ways to buy these niches of segments on Facebook and Google ahead of everyone else, then that can be a real advantage in your go to market.

Yeah, I mean I've been doing paid acquisition for a very long time. I wouldn't want to go compete with wish right now they have a big competitive advantage in terms of what they've learned, how much they've done, how much they've tested that even me with a lot of expertise, there's no hope that I could go and catch up with them very quickly. That's a real sustainable advantage in a way that paid acquisition a lot of times historically is not. It's been like, oh, you're spending a dollar with Google, so am I. Big whoop right now. It's not the same. They've actually made that core DNA at their company and they've built a huge head start on anyone that wants to compete with them at that level of scale.

That's right.

I want to go back to the Google Facebook duopoly and there's a couple pivots on this topic that I'd love your perspective on. One is that that's just where all the advertising money is spent, I'm sure at Uber, at Pinterest, at all these companies that are doing paid acquisition is going through Google and Facebook. But that also means that companies are desperate to find anywhere else to spend money. So there have been some emerging platforms in the past few years. Snapchat, Twitter, Pinterest, Reddit, Quora and LinkedIn for B2B. How do you think about evaluating those new channels as a startup? Is that a thing to kind of rush into and try to make work? Do you want to sit back and make sure other people kind of test it out for you to learn? How do you approach that?

I'll make two observations on this. First is when you're a startup, it doesn't really matter where the majority of ad dollars are spent because the reality is that you just should care about where your ad dollars are going to be spent. And when you're small, you can spend them with anyone and get, you know, as long as there's results, that's exciting. In fact, if anything, what you would argue is that an early startup you would almost rather have a few kind of proprietary partnerships that are getting you traffic, maybe from like other media properties that are in your space that aren't that competitive before you like go onto Facebook and Google. Those are the things that scale. But that's going to be the place where you're going to deal with obviously the most massive competition.

Right.

So I think that's one framing. The second thing I would say about that is when you are a startup that kind of like doesn't know if it needs to use paid acquisition to compete, then I think going with kind of the tried and true is perfectly fine. Right. But when you can see a really clear advantage, I mean, I think one example would be in E commerce and you're like, oh, there's going to be this new Pinterest ad system that's going to come out or you know, some new feature of this ad system, you should jump on that right away because that's like a really clear kind of product channel fit. Like, you know it's going to work. That's definitely worth exploring versus like if it's a completely random new thing, it's sort of like, well maybe like let's let all the other folks who are professionally testing this stuff see if it works first and then go from there. I think that's why a lot of times you're watching what the affiliate people do and what the lead gen people do because they're in the business of trying out all these different things. In the early days of even the Facebook ad system, it was really clear that it was gonna work because you could run even really small campaigns and you could watch other people run campaigns and it was like the traffic was good and so you kind of knew that like, okay, this is going to be a thing.

Yeah, it's interesting. There's this risk versus reward for these new channels, right? In that there aren't best practices that are available yet. Yet because of that there isn't that much competition and you can maybe get a lot cheaper cost per click or cost per acquisition. And you have to kind of do the risk versus reward math for the different channels and for the different stage of company you're in. If you're an early stage startup, there's so few things that you could Spend time on, like, just the risk of distraction is really high. If it doesn't pay out yet, you're sure if you're Uber, you're just willing to try any new thing at Susan that comes out because you have the resources, you have the money. That's right. And you're always looking for that competitive advantage.

You know, one thing to acknowledge is that for a lot of these platforms, when they first build their ad systems, like, they don't know whether or not it's gonna work.

That was really true for Pinterest. Yeah, exactly. Right?

Yeah. And so I remember this was like a long time ago, but I remember there was an interesting blog post about, I think it was Joe Krause, who had worked on Excite, which is like a pre Google search engine, talked about how they had tons of search traffic, they just didn't know how to monetize it.

Right.

So back in the day, they would just throw up a bunch of banner ads and they would be like, well, how do we actually monetize this traffic? Well, and it turned out that banner ads was not it. And it took obviously several iterations with Overture and then ultimately adwords before they could make that happen. And I think, you know, for the Pinterest, when, you know, they started things out or when Snap is, you know, they're continuing to iterate on things, I think it's going to take a while in some of these cases to actually figure out what's the right segment of advertisers and what's the right, like, format of advertising.

Totally. I mean, I think Pinterest was a bit naive when it launched the ad platform. At first it's like, oh, people are here telling you what they're interested in. It's just going to be perfect for advertisers. Then we launch it. And the first few campaigns didn't work super well. A lot of people at the company were just crushed. It's like, guys, this is an iteration process. We're going to learn, we're going to build something better. And then we started finding, oh, some advertisers are having success. Why is that? Oh, what are they doing? What's the creative thing that they're trying? What is the targeting they're doing? And then eventually have built something that's working a lot better. Now, four years into it, I think what ad supported platforms tend to underestimate is how long it took the Google AdWords model to actually work. I remember trying Facebook ads and being pitched for them in 2007, and I was like this is a joke because I'm used to AdWords with all the features and the AdWords editor and all that stuff. And they just didn't have any of that. And then they just Facebook kept iterating. And then of course when Facebook was going public it's like, oh, you don't have a mobile ad strategy. Turns out they figured that out just fine. It just took a little bit of time. As an advertiser trying to leverage those platforms, you're figuring it out with them. You might figure it out earlier than the platform does. And in that case you can create a tremendous amount of value for yourself because you have very great cost per acquisition, very high lifetime value, all of those things. But the reverse can also be true in that you just get no value and then you feel like you've wasted a bunch of time and money.

So maybe one way to think about making this actionable would be that if there's a new platform that's coming out and it is really easy and self serve to just put in to plug in like 100 bucks a day, that information is probably worth it because it might work just right off the bat. Once you go beyond that then there's a question of like, okay, how much do you really enjoy invest? And I think that there are these ad platforms, like I think, you know, snaps video ads is one, one example of this where it's not quite clear. And so I think because of that, you know, those things are much less interesting to a new startup. I think as a startup you really have to do kind of self serve, small budget, very performance marketing oriented and like go after those as, as new things that you try.

Yeah, I think a very dangerous word in startups is the word awareness because you can justify doing a lot of things that you won't see in the data help you, that in a lot of cases aren't helping you. What both you and I try to echo in almost everything we do is that the most important thing to drive is product driven growth. And that's the thing you should be focusing most of your time on. And then performance marketing can be this accelerant on top of that that functions very similar to product driven growth. You can measure it, you could scale it. Brand is a super important thing, but it's typically not a thing you're investing in until you feel like you're really, really on product driven growth and performance marketing. And then it's an additional accelerant that maybe even creates more of a moat long term. And in consumer you really shouldn't be investing in brand justified spend until you're a very large company and you're very secure in B2B. You might do it a little bit earlier because the lifetime values are so much higher. But it's very dangerous to be like, oh, I'm just going to try Snapchat brand ads and in two years it's going to pay off. I'm not going to be able to measure it. That rarely works for startups.

That's right. Yeah. I think there's the Y Combinator thing where they talk about how startups are default.

Dead right.

And so once you get to a point where you're kind of like, okay, maybe this is actually going to work, this is awesome. Then very quickly after that you can justify things that look more like brand spend because then you can sort of like amortize them. The cost of that over a long period because you're like, okay, well you know, my company's probably going to be around for like the next 10 years. So now I can do things that have a longer ROI window than like literally getting people to buy clicks. But I think there's a special time in your company when you're able to start thinking about stuff like that. I think that's much more like a Series B, Series C. Like, you know, if you're going to think about that at all.

Yeah, makes sense. One thing about paid acquisition I want to dig in on with you because I know you've seen it, is kind of the irrationality. You could call it irrationality, you could call it strategic advantage. Maybe smarter people than me will call it game theory about paid acquisition and competition in that you have people seem to be playing a different game. If you're a startup and you're trying to make your money back in four months and then Uber comes in to compete with you with their war chest, they don't necessarily have to have their payback period in four months. How are you seeing the competitive dynamics play out in paid? Are you able to avoid the kind of direct one to one or just it's going to get there eventually. And trying to figure out the advantage earlier is kind of who wins. How do you think about that?

Especially in the world of direct to consumer like E commerce. I think we've seen that in like a really big way kind of play out, you know, where the really naive way to just describe it at like a high level, what's really happening is you have a startup, they think that they have a certain ltv, right. And so they can Justify that. Because either they're doing a longer payback period, as you're saying, they might think, like, oh, no, no, we're going to get people onto this thing and then we're going to sell them other stuff. You know, that's like another point of view. Or maybe they have a really rosy point of view on what their churn looks like, because the early cohorts were really good. But then they don't necessarily know that, by the way, like, by the time that you wait out 12 months or 24 months on a paid channel, that's going to be very different than the folks that you got organically in the first place. Right. So you have all these kind of, I think, asymmetries. And then I think the harmful asymmetry is the one that basically is, I know that if I can show traction, that I can go raise more money. Right. And it's just like land grab mode. And like, sometimes you have an investment community that's willing to go for that. Right. Like, I think we had this whole phase of all the monthly subscription boxes, like, you had one work. So everyone was like, let's try to build the, you know, Birchbox for X. Right. Obviously, once Uber got started, everyone was like, all right, let's do Uber for X. And we just sort of know that the unit economics aren't going to quite work out at the beginning, but we're going to, like, invest into it.

Yep.

But I think as a more mature unit economic positive player in the market, that, like, looks really crazy to you. Right. You're basically like, oh, my God, like, all these people, like, they're pouring all this money in and you know, it can. And the dollar amounts that they're spending don't work for us in our model. So it's hard to match.

Right.

But also just that they have maybe just a different business than yours or they have a different strategy than what you're really going for.

Yeah. I think this was really fascinating to watch in the food delivery space in the last five years. So, you know, when I joined grubhub, there were about four competitors around the same size, all with essentially the same business model. Right. Which is like restaurants that already deliver food. We're going to match them to people ordering food. In grubhub, we were able to have some strategic advantages with our supply side, with our brand. Our CPA was a lot lower than our competitors on things like AdWords, and we were really proud of that. Nowadays, if I saw a CPA a lot lower than my competitors it might actually make me scared. And I think that kind of happened once I left grubhub and kept in touch with the team is after grubhub went public, all these delivery startups came out the doordashes, the postmates, the caviars of the world. And the person I was talking to that took over acquisition after I left, he was like, they're just spending so much money, so much more than could possibly have a positive outcome for them. And I think he saw it as dumb money and that it'll wash out of the system eventually and grubhub just has to have a little bit longer payback period. I realize that we should probably define that payback period for those of you who don't do a lot of paid acquisition is if I spend 20 bucks to acquire a user, how long does it take me to make that 20 bucks back off of them in revenue? So Grubhub would target a six month payback period, which I think is pretty common for mid stage startup. Early stage startups that maybe don't have LTV data are going to try to make their money back on first transaction. And then now Grubhub with 10 years of LTV data can probably stretch that out to a year or potentially even longer just because it's more predictable. Grubhub was like, oh, is this dumb money? And I think if you're doordash or postmates during that time, you're like, I just need to provide liquidity and I'm willing to pay any amount of money to make sure that my supply gets demand and they stay on the platform because like I'm paying for these drivers. I paid to acquire restaurants. So the fact that I'm spending 5x what Grubhub is spending on a consumer, like I don't care, like the business doesn't work if I don't get that spend. And that was initial change during the business model. And then Uber Eats came in and that's even scarier because, you know, Uber has these competitive advantages. It's not that they're being irrational. So they have, they actually have cheaper cost of capital. Right. And it's weird to think that a private company would have cheaper cost of capital than a public company. But if grubhub basically doesn't show profit growth, their valuation drops. Right. And then Uber amplifies their core business. Right. So that's just a huge advantage that grubhub doesn't have. And I think it's important that if you're getting into this game that you understand the games that all the competitors are playing. And in the case of Grubhub, like the early game when I worked at, it was easy to understand everyone's playing the same game and all of a sudden Postmates and DoorDash are playing a different game and Uber Eats is playing a different game. And it may change the ability for you to compete on some of these channels.

Yeah, that's right. The on demand model is with a two sided marketplace and with multiple product lines is really different than like a standalone like food business. Right. And so if you're Uber, you have a pretty different set of variables that you're playing with because as you mentioned, you know, after somebody comes in through Eats, that means that we potentially can upsell them to, you know, more trips.

Right.

Or vice versa.

Right.

Or you know, the other point of view is like, well, maybe we get a ton of like organic Eats users off of people who are riding with us. And so that what that means is that, you know, that's a really low cost of customer acquisition. And then the ad spend that we get is like a different set and then the blended amount is still actually looks really healthy.

Right.

That might be another way to view it. Right. And so I think there's, there's a lot of different points of view that you can take on this stuff that makes it an asymmetrical comparison and it makes it much harder to just say, oh yeah, these dollars that are being spent are smart or dumb. Like they're just different in the marketplace.

Yeah. Which I think raises a point that you wrote in your blog post about growth getting harder. Is that a real strategic advantage is investing in your business model earlier to figure out your LTV earlier. And that can allow you to justify a lot of things that, you know, companies that don't really know how much they're making, they just can't do. So I'd love to hear kind of if you have any thoughts on if you're a startup and you want to do that, what are the best practices? Is it just you have a totally different type of business, you start, you start B2B. You don't worry about consumer, like how do you think about that?

So if you look at this problem from the kind of growth first kind of lens, right. What you're going to quickly realize is that in order to compete with other people for the same audience, you're going to need some kind of structural advantage in the amount that you're able to pay. That's one version.

Right.

So you're able to pay more than others and we'll talk about what things could lead to that. The other version as well is if you're able to utilize certain APIs and channels within these big platforms in a way that maybe others aren't. Right. So in the first version, you know, the structural thing is that, like, I think that is why for a long time people were super excited about the subscription boxes. Because what ends up happening is you have this longer LTV period that you're able to deal with. And so then you could initially, in a world where most people weren't doing subscription, if you came in and said, oh, you're doing X dollars selling razors on the Internet, well, if I have a subscription version, I'm able to then pull forward potentially a year or two years of LTV and completely just crush you in terms of bidding and like really do that. So I think that's like one version and I think we can, you know, subscriptions, clearly one thing. The other version would be you are able to cross sell more effectively. You look at all these, like travel companies. A lot of what the folks within travel do is they're often cross selling all of the adjacent services. They're making their money in all the other services.

Rental cars.

Rental cars, like all that stuff. Right, right. So if you can come up with kind of a unique insight like that, you can go now. The problem is a lot of times they're less proprietary than you think. And so then over time it sort of washes out. Everybody can build a subscription version of what you're doing and it gets difficult, but hopefully by then other effects like brand, which maybe gives you higher conversion rates, maybe you figure out your referral program, maybe you figure out some of these other channels, they can help create your version of your growth moat.

Right.

Based on that, then I think the other piece is if you can go after Traffic or use APIs in a way that maybe others aren't. And I'll use a cautionary tale for this, the story of Teespring, which obviously early on they were basically creating a lot of custom T shirts and they did a lot of really clever things on Facebook in particular around taking users like topics that they had liked, pages that they had liked, and where they went to school and all that stuff and advertising, it would be like this hyper personalized thing, really, really cool. Right. And so that kind of thing can give you a huge advantage. I think the cautionary tale in their case was that this was basically only offered on the Facebook platform. And once that stopped working, then there weren't a ton of other channels to go to after that.

Yeah, it's fun to hear you talk about structural advantage. And then in travel, like I think we've all heard about the booking.com story, which is they just got way better at conversion rate optimization than all of their competitors and that allowed them to bid More aggressively on AdWords and other channels because they could actually close the deal, others couldn't. And that has been a key driver in Pricelines stock.

If you were one of these online travel agencies, OTAs, you would actually do this arbitrage. You would like go to the hotels, these mega hotels and you would say, hey, I will pre buy a bunch of your inventory and then I will sell that to my audience and you know, and then, and then take money on the market. And the hotels love this of course, because they basically get this kind of like pre booking, you know, inventory that's already locked down, it helps them create some stability in their revenue, etc. Etc.

Right.

And that was great. When booking came along, they basically said, okay, well we're not going to do that. We're just going to be a lead business. The margins are going to be worse and we're just going to send leads to these hotels and they're not going to do the arbitrage game. Right. And that did a couple things like that were pretty interesting. One was that it meant that you all of a sudden when you're coming into booking, like they would collect your credit card, but it was all actually kind of like optional because they didn't really need you to buy the thing, it was just a lead. They obviously pioneered a bunch of those fun techniques where it's like five other people are looking at your listing and you're like, oh no, I better like transact. That's a great example of a structural kind of asymmetric advantage because if you're Expedia in that case, you actually don't want to convert your entire business to a low end lead gen thing. But then booking is able to not only have lower friction in the funnel, but they also can offer a wider variety of hotels because they don't need to just do these deals where they like buy inventory off of Hilton. They can also do lead gen for really unique, quirky bed and breakfast kind of things cause those guys will buy leads as well. So it became a larger volume, lower margin business. And I think that's the true version of a structural advantage. Because it's so structural that your competitors don't actually want to copy you until it really works, but it gives you a advantage in how much you can spend and what kinds of terms that you can go for, because all of a sudden, you can go for, like, tail terms. You can advertise a different kind of hotel inventory. So I think that's a really interesting use case, and I think that's a really interesting thing. If you're a startup and you're thinking about your bigger competitors and where they're spending their money, you obviously want to try the same channels and the same things as them just to get the benchmark. But I think you also really want to think about, like, well, what are the things that they don't want to do do, or they're not willing to do because it doesn't really match their business model and really articulate the growth strategy around that.

Yeah, I mean, there's so many good things about what they did that are still working for them, like, 15 years later. It's a really great story. I want to shift from a bunch of things people should do to some things people should avoid with paid acquisition. You talked about Teespring maybe a little bit over reliant on one tactic that was something that Facebook wasn't going to necessarily continue to offer. I think there are lots of other examples of paid acquisition kind of gone wrong. Are there any that kind of come to mind for you that you'd want to caution the audience, don't make this mistake? And it's really clear how not to.

I think the big thing is paid acquisition is something where it's easy to get addicted. Right. You buy a little bit, you're kind of like, okay, cool, this is working. My revenue numbers are going. It's easy to get addicted and to suddenly start saying, well, I'm only getting to this scale, so I don't want to have a bad quarter, so I'm going to, like, push the LTV window. You know, as you were saying, your payback period, you know, instead of being six months, you're going to say, well, what if we made it like nine months or 12 months? And then you're at 12 months and you're like, well, maybe like a little bit more. And, like, it's very tempting to do that. It's easier to make that kind of a decision than it is to make the decision to put a team of people to, like, optimize your whole product. What ends up happening is because it's the easier decision. It's very tempting to, like, go for that. So I Think that's one one aspect. I think the second part is it's easy to also become short term about it where there are just so many examples of direct to consumer e commerce companies even actually going back to, you know, the living social and groupon days.

Right.

Like these are all stories of basically paid marketing arbitrages that exist for a year and then go away and then companies like collapse, like hundreds of people lose their jobs in many cases because of these kinds of decisions. Right. And so one of the points is that really understand where paid marketing is in your set of levers, right. Is it really a durable proprietary thing that you're going to be able to use over a long period of time, or is it just like this wedge into the ecosystem either early on in your startup in order to get the initial traction, or maybe as a complement to your other stuff, but really understand where you're going to be able to compete on acquisition and really be intellectually honest about where that is. I think that's really hard.

Yeah, I think there are a few good case studies on where that's kind of gone wrong.

That's right.

One that I'd like to bring up is kind of what we're seeing played out with Blue Apron right now. I think a difference that I don't think is well understood with companies that use paid acquisition is whether your core business model has a network effect to it or not and how that impacts paid acquisition. Grubhub, Uber Eats, these are examples of paid acquisition. But there's a core cross site network effect where you have an early user that orders a lot of delivery, you go acquire them, they order from restaurants, you're able to use people like that to go acquire more restaurants, which then makes the product interesting on the demand side for people who aren't as hardcore delivery users that end up ordering just as much as that initial person because of the addition variety. So the cross side network effects allows both the supply side and the demand side to expand over time and maintain really healthy lifetime values. And then compare that with something that on the surface seems like a core substitute, like a Blue Apron plated. These food delivery subscription businesses. The difference between those two products is that Blue Apron doesn't get better as more people use it. So Blue Apron is a classic case of they found these core customers that worked really well, they retained really well, therefore they had a high lifetime value and then they could acquire them even at cost that maybe grubhub or Uber Eats couldn't afford. Right. But then they just kind of ran out of those users, and then they have to expand to people that aren't as qualified for the service, which means they're not going to have as good a retention, they're going to have lower lifetime value, they're going to churn out at a higher rate, which then means they need to go replace those people with new people who are even less qualified over time. So you have this trend of every user you acquire is worse than the last person you acquired, and then it kind of doubles in that they get more expensive to acquire, even though they're worse, which means you need to go to more and more fringe channels. So it just adds complexity, it adds cost, and it increases churn, and you run into this situation where you run out of the market eventually. So if you look at the, you know, Blue Apron subscription subscribers curve, it's a pretty exponential curve, and now it's actually gone negative. And so if you don't have network effects tied to your paid acquisition, it's not the end of your business. But if you're modeling it as if you do, you will run into some trouble. Right? So Stitch Fix, I would say it's very unclear if they really have a network effect. I'd say they probably are claiming they have a data network effects. As they get more users, they get better about picking items. But what did they do? They grew very organically. They didn't spend a lot of money on paid acquisition. They used Pinterest as an organic channel. They used referrals as an organic channel. So they were able to really maintain a low cost and they were able to have an innovation in terms of how much people returned because you got a cheaper rate if you retained all the clothes that they shipped to you. So unlike a trunk club where returns were like 50% of a box, they had very low returns at Stitch Fix. So they innovated on the model and kept these more organic growth levers instead of just relying on paid as a channel the way that Blue Apron did. And so far, their IPO appears to be a lot more successful than Blue Apron. So thinking about all those structural issues, as well as how the network effects come into play and are you really creating a sustainable loop or not is super important. And I imagine it's very painful working at Blue Apron right now and just kind of learning that live while the market learns it with you.

Look, if you don't have a structural advantage on this stuff, it's very hard to add it in later.

One last thing I want to talk about with paid is we talked about a lot of examples where there's transactions involved. Certainly our experiences are in marketplaces. What about when it's not a transactional business? Do you feel like paid acquisition can be a real lever for those styles of businesses or do you need to think about something else? How do you think about it?

Yeah, so I think when it comes to different kinds of acquisition channels is which things are loops that sort of can scale and sort of like self perpetuate and which things are kind of these linear channels that are like, they're nice to have but you can't really scale them. So the example of the former would be like content can be both a loop and a channel, but like if you wanted to create one as a loop, you would use probably user generated content, or you would need something like what Zillow and Redfin does, where they create just a massive number of landing pages. But the whole idea is how does each user who comes into Yelp, for instance, there's a certain percentage chance that they're going to sign up as a Yelp user and ultimately create more content, which Google will then index and the whole flywheel will turn. And so that's an example of something as a loop. So I think the point of this is that if you have a main thesis and a main loop around acquisition and how your product works, and hopefully over time you can add multiple, like LinkedIn started with invite virality and they've added more SEO and now they've added even more stuff. What happens is you can use paid as an accelerant for that, right? So, you know, maybe early on you don't have a lot of content, or maybe early on you don't have enough users to make a referral program work. But what you can do is you can like, you know, sort of juice some of that stuff and like get that content created. In the first place that's one great place to use it. Then the second place, as we've been talking about, is that a lot of these products end up being networks of some kind. Right? Whether that's something like a GitHub where people come together to share projects and files, or it's a completely private thing. Dropbox is a completely private network and it's in the workplace, et cetera. What ends up happening is that a lot of times you need some kind of critical mass in order to get the whole thing going. And so you can imagine using paid marketing. I've heard that there's many of these growth teams, for example, at Facebook, where they typically don't use paid marketing. They may use it in specific cases where they're trying to juice growth within a particular geography. As a for instance.

Yeah, I think they're actively using it in India, Indonesia, Nigeria, stuff like that to make sure they get the base.

That's right, exactly. And if your thing has network effects and there's a bootstrapping period, then that's awesome for that. And you know that applies to marketplaces. Of course, that applies to user generated content things. It can apply to a lot of different aspects. So I think you want to make sure that in those cases though, that where you're spending the money and how you're spending it is very careful.

Right.

That works well in especially. We were just talking about a couple of these Southeast Asian countries. A medium sized budget can go a long way right within that. And that's awesome. Right, because then you can really like juice the thing. Now if you're trying to build something that is for the developer community or something that's for the design community, it's going to be really expensive and really hard to bootstrap that using paid acquisition. So for something that's more narrow like that, especially if you're targeting like the US GEO plus something that's like much more expensive, all of a sudden your acquisition is going to look more like a couple hundred dollars per activated signed up US professional worker. And so then all of a sudden you're like, okay, this is actually going to be kind of hard to do. And that then becomes not a great kind of bootstrapping solution.

Yeah, I agree. So I think there's kind of bootstrapping as well as kind of the acceleration. And I think at Pinterest we traditionally have grown organically. Right. But it was totally worth it to kind of pay to get the first users in Japan or something like that to get people who would contribute content. And then also what we found is once that content loop is working, that the content gets distributed to Google or Facebook. You could actually pay some money to boost that content even more on Facebook and especially in international countries. That return was absolutely amazing and well worth it. You would make your money back very quickly. So I think BuzzFeed kind of pioneered this. We're going to create awesome content and then we're going to actually spend money to push it out to get the initial viral loops going on that content and then kind of let it play out from there. And the level that they've scaled that over time is pretty immense and kind of upworthy was kind of another one that was really well known for that. And I think we're totally seeing that be a viable channel. I mean, the jury's out on how amazing those content businesses are going to be. But they've been able to scale and grow and I think do some of that pretty profitably. You can't just think about, oh, I'm a content business, therefore I can't spend any money. You might need to think a little bit smarter about it. But using to bootstrap viral growth of a particular article or to build the initial network I think makes a ton of sense.

Yep.

Cool. One scary element that I want to talk about is there's kind of this other element of the Facebook Google duopoly is kind of them as the boogeyman that is going to eventually just compete with you if you're successful, especially if they're seeing you continue to spend more money on their platforms and grow really fast if you're in a different company. And I'm sure Uber has had the fear of Google competing with them and whatnot. How do you think about that? Does that worry you? Are there things you should do to avoid that? How do you think about that as a startup?

There's been absolutely that narrative that we're in a space now where even if you do well on consumer, these companies will kind of chase you down and they'll clone you and they'll copy you. And the reference point I think that everybody talks about is Snap, right? Like Facebook and Snap and how aggressively now Snap stories are now integrated inside of Instagram and Facebook itself and WhatsApp. And now recently I just saw in some news that they're going to allow Instagram stories to be cross posted to WhatsApp. So that's another leverage point within the ecosystem. It's hard for me to look at that and think it's a wildly compelling case because I think it's vastly oversimplified and disregards the fact that, yes, that is happening and Instagram looks like is making a dent in Snap. But we also have to remember that before that Facebook had poke, they had slingshot. Slingshot, yeah. They had paper, they had all these labs things and none of them worked. And the only thing that ended up working in a big way ended up being taking one of their primary properties and integrating Snap's feature set deeply enough that, you know, it was so obvious, like. And so that becomes a huge product bet. I think that's totally fine. But that is a big gun with very few bullets. You know, it has like two bullets in It, Right. And they just shot one. The way I look at that is it's actually an awesome time for innovation. If you have a lot of like consumer startups that are getting scared off from it, then the folks that are actually in the business with a really unique investment insight and unique value prop, actually it's going to be less competitive. It'll probably be a good time to start a company, right?

Yeah. I think when you look at all the companies Facebook and Google have tried to compete with over the last five years and what their successes look like, I think it boils down a lot to has the company they've tried to compete with really understood their own loops and had that strategic advantage or not. And when we think of Snap, we don't think of it as a company that really understood and tuned its viral loops and built moats around its users. It's more of a thing that just blew up. Whereas you look at other things that Facebook and Google have tried to compete with that have really had more built out an understanding of what makes their ecosystem unique and what their loops are. I think Facebook and Google just haven't been as successful. Facebook recently competed with GoFundMe, which is one of our portfolio companies. And I think GoFundMe is doing just fine. It's across 5 billion GMV because GoFundMe has a core loop that continues to work and they've built that mindshare that this is the way that you raise money for causes. At Pinterest, we kind of faced competitive pressure from both Google and Facebook over the years. Google famously copied a lot of guided search functionality into image search. And Facebook and Instagram have built save buttons, they've built collection features, they copied a lot of the Pinterest core product. And Pinterest continues to grow fairly aggressively. It just crossed 200 million monthly active users. But that's because Pinterest understands its acquisition loops, it understands its core engagement loops and these salvos from these other companies, it's hard for them to break these existing loops. But if the loops don't exist or you don't have ways for them to drive additional users to get introduced to the loop, I think you're much more risky proposition of if Google or Facebook start to care, they can start to take that stuff away. I really love your assessment of the gun with a couple bullets. When you think about how much work it took just to take one company down as the Facebook behemoth, how many times are they going to put that level of effort into killing your little startup? It's unlikely.

The other way you could talk about Snap is you could say, hey, these companies will probably let you grow really fast, become a public company, make yourself very wealthy, and then they'll chase you down. For most entrepreneurs, getting to that stage is already hard enough, much less worrying about the long term ramifications.

I think if you're an entrepreneur and you got to go public as a $15 billion company, you're like, yeah, I'm okay.

Yeah, exactly.

That's fine. If now it's going to be hard.

It's a public company and I think this story gets told quite a lot. Like even when you're talking about like Google and I mean, how many years has it been told that like Google Drive is going to kill Dropbox and yet that's, you know, not happened. Right. It's like I think one of the things that people undervalue is that, you know, network effects are just, I think, you know, categorically undervalued. It is just really hard. Once you've built a network that is like thriving and you have decent retention rates and all that other stuff, it'll just keep going and going and going. And maybe what may happen is that one of the bigger companies, if they execute well enough, they may be able to cap some of the upside. I think that's probably more the danger. But it's really hard to get entire networks of people to switch from one platform to the other.

Yeah, it's not as if Snap has lost all their users. It's really more that they've lost their ability to acquire new users at an aggressive level. Clip. So I think it's a really important point and travelpox is a really great example.

I think Slack will be another interesting test of all this as everybody sort of views enterprise chat as a feature of all the things that suites need to provide. So Google Suite and obviously Microsoft and all their product offerings. But I think based on this theory, what you would say is that actually probably Slack will continue to really continue to crush it on their existing user base. And then the question is maybe the attrition from some of these other products into Slack will slow over time. And I think that's the real risk. I think one of the most interesting moves I think that Facebook's done is they have with regard to the whole Snap thing is that they are more aggressively expanding the whole Stories concept and all those things into the developing market because that's where you where a lot of where they have expertise. Like the fact that some of these features are now part of WhatsApp will make it hard for Snap to, once they get their act together around low end Android and they're into those markets, I think it'll still be really hard because then all of a sudden it's like, oh wow, there's actually an incumbent and they're the challenger brand as opposed to the other way around.

Yep, absolutely. And yeah, if you want to learn more about Snap and their user growth challenges, we had a great podcast with Julie Zo a while back talking about that in more depth. Well, thanks again, Andrew for taking the time. I think we learned a lot today. We talked a lot about the law of shitty clickthroughs and how each of these channels has a lifetime in terms of how they're going to work and times when they're going to be emerging and really easy and then they're going to have decay periods where they get harder. That I think really drives a lot of the decisions that you need to be thinking about in terms of how you're going to grow your company. I think we talked a lot about where virality is on that curve in different capacities in that it's not necessarily about am I trying to invite people or am I trying to give people money, it's about really understanding the framework of what's going to motivate people to talk to others about this product and then tuning a program around that. How the bar for that has gotten really high for consumer apps because there's just been so many programs that have tuned this that people are starting to tune it out on the sharing side and on the redemption side. But in B2B, that's actually alive and well and becoming a more emerging channel to replace some of the early enterprise sales that those startups used to do, which is very exciting. Another thing that I think we learned a lot about was how to think about paid acquisition. Should it be the core of how you grow your business, or should it really just be an accelerant that's built on top of a more core loop? And also, if you're really going to go after paid acquisition, what's your structural advantage versus how you're going to beat the competition? What are the things you're going to be able to do that separate yourself for the short term and the long term? We talked about in food delivery, the advantages that Uber has versus some of the incumbents, and in travel, how booking.com really was able to separate itself on multiple advantages. You really need to think about these things to be successful in the long term for paid acquisition. So I hope this was really helpful to you as you're thinking about growing your business. Don't dismay growth is getting harder, but it's not that growth is impossible. You just need to think a little bit more strategically about how to make it work for you.