Scaling Users, Products, and Growth Teams with Casey Winters | Greymatter

In this episode of Greymatter, Greylock growth advisor Casey Winters talks to general partner Josh Elman about best practices for acquiring and retaining engaged users, starting and scaling a growth team, and what he learned running growth at Grubhub and Pinterest. For more news and podcasts, please visit.

News.Greylock.Com hey, this is Josh Elman from Greylock Partners. I'm here with Casey Winters, who's a growth advisor with us. At Greylock, we talk about growth all the time in the context of startups, how they grow, how they attract users, how they engage users. And we're really excited to have Casey as a growth advisor working with a lot of our companies and able to help them figure out how to grow and how to work through all these opportunities around engaging users and retaining them. So Casey, your background really started in marketing. You have a marketing degree. You have an MBA with a concentration in marketing. You were doing marketing at grubhub and then you came over here to the Valley to Pinterest and you kind of transitioned from marketing to growth. Can you talk about this transition from marketing to growth? How did you get started and sort of, what do you really see as this marketing and growth and this kind of new way of engaging users?

Sure. Well, first off, thanks for having me, Josh. Excited to be here. Yeah, I'd say for me it started right after undergrad. So I started working@apments.com and started working in the online marketing department and everything that I needed to do as a marketing analyst and everything I needed to measure was totally different from what I learned about in school. It was things like paid search, affiliate marketing, SEO, email.

What'd they teach you in school?

They taught me about integrated brand campaigns and spending a lot of money on television ads and things like that. Right. So just none of that was really what was happening in my first job. So I basically took the time to learn how all of these new marketing tactics for tech companies worked because I had to measure them for effectiveness. And from there I just kind of started working on implementing them as I progressed in my career. At the time, we still called it all marketing, but it really was much different from what I learned about in school and what people think about marketing now. Much more analytical, much more experiment based and much more going inside the product. And I think as I evolved my career, moving on to grubhub and then Pinterest, I got closer and closer to making the product changes that would have the largest impact in tech companies growth instead of the more advertising based changes, which while they can be valuable. Generally aren't as impactful.

Awesome. It's a great way to think about it. This sort of transition of changes you can really make in the product versus just doing on the external or paid or campaign side. So you were the first marketing hire at grubhub. They were in Chicago, right?

Correct.

You helped the company grow from 30,000 people in three cities to more than 3 million users in or customers in 500 cities. So can you talk about the strategy of scaling locally? How did you think about this? Especially kind of at Internet scale, How did you target new markets?

Yeah, so figuring out which new markets to go into is basically saying, okay, we're online food delivery. Where are their delivery restaurants? Because restaurants aren't doing delivery, then we're not going to be very successful. So we would look at other sites to try to figure out which cities had a lot of delivery restaurants. And then we started to figure out other things that were really important for city expansion. Were people on Google searching for delivery. So there was their existing demand for consumers to have delivery. And then what was the population density of a city actually was a key factor because people tended to order delivery more if it was much more compact. Otherwise people would drive everywhere and just pick up food. So those became like the three main criteria. We looked at number of delivery restaurants, the population density, and then the number of people searching for food delivery online. And that basically created a list of the most dense cities in the U.S. san Francisco, great. Jacksonville, not so great. Then we just worked down the line. Originally we'd launched maybe one city a year. Then after a while we got so good at it that we could launch 15 cities in a few months.

Just help me understand Grubhub's model a little bit more. At that time. We're all probably familiar with things like DoorDash and Postmates that will just go to the restaurants and pick it up for. But grubhub really was bringing this sort of online ordering system to these restaurants that already offer delivery. Correct. So how did you think about making that sort of, what did you think the real advantage that grubhub brought at that time?

So you kind of need to think about it on the supply side and the demand side. So on the supply side for restaurants, the value that we were giving was we were bringing them new customers. So we didn't really talk about technology. We just said, Sign up with GrubHub and you will get more orders. And that was a very simple message for consumers that would actually place delivery orders. Most people thought that there were maybe five restaurants that delivered to them, a couple Chinese places, a couple of pizza places, and that was it. And the reality was that it was more like 150 different restaurants in a lot of these cities would deliver to you. So when people came to grubhub, they would see all these options they didn't know existed. Thai, Indian, sushi, and they would be able to order a lot more different types of food. One piece of the value was the variety. The other piece of the value is the convenience. So being able to see the entire menu, the price, being able to pick, you know, just press a button and have the options added in your cart and the order submitted, not having to like wait on hold and tell someone your credit card information, which is kind of sketchy. All of that made it a lot easier for people to place orders as well. So it was kind of this two pronged approach of variety and convenience.

And really what was the major tactic that you guys used to really grow? Was it a lot of paid marketing? Was it a lot of SEO? You know, was it, you know, social? Or was that even a big thing at the time? Like, how did you think about sort of like finding this new demand and connecting them with the restaurants through grubhub?

Yeah, so I'd say there were, there were three main things that were helpful. So one was definitely SEO. There were people searching for, you know, San Francisco food delivery. And if you could show up, number one, you would get a lot of orders because they were like, oh, I don't have to call. I can just press a few clicks and have food on the way. Advertising was definitely a big piece as well though. So we obviously did paid search on top of organic search. But we also started to do things like outdoor advertising.

Back to what you did learn in school?

Yeah, back to a little bit of what I did learn in school. So for Grubhub, 50% of its orders were placed between 6 to 8pm local time. And what's everyone doing before that? They're on their way home, so they're either in their car or they're on public transit. So we figured out pretty early on that if you bought public transit ads, you would get a pretty good response. But then it was basically like built in offline retargeting because every time they went home, they get reminded that they've used this thing. So we started. Every time someone placed their first online order, we would survey people asking how they heard about us, and we'd be able to build a cost per acquisition off of outdoor. And it turned out the cost per acquisition was incredibly low, and the lifetime value of those customers was higher than organic users.

And in terms of tracking with outdoor and things like that, you don't have any real tracking. But talk a little about these surveys and how much they actually showed to you, because that's, I think, something that a lot of companies figure, oh, we can track everything. And they sometimes forget the easiest thing is just to ask.

Yeah, exactly. So as soon as someone finished their first online order, before they could get the order receipt, we'd have a one question survey that asked, how did you hear about us? And then it would just have our marketing mix and it looked like you had to fill it out to get to the order receipt. So 60% of people did fill it out, and that is way enough data to be able to statistically measure the impact of a campaign. So later on, when we did tv, we did the same thing to measure the impact there. So basically we were able to run things that are traditionally considered brand marketing as if they were performance marketing. And if the CPA hit a certain goal, then we were fine with it and we'd continue to ramp it up.

Is that something you think a lot more companies should do, is just ask users how to do it? Do you actually work on that when you're advising companies?

Yeah, absolutely. So generally people think there's less tracking available than there is. Right. And if you spend some time thinking about how might you triangulate how this thing is working? There's usually a way. It may not be perfect survey data. 40% of the time people don't answer. But it turns out 60% is a really high sample. Right. So I always try to figure out with the companies we're working with, okay, what's the thing we want to do? How do we want to measure for success? What's going to make us feel comfortable if that is successful or not. I never want to go into a situation where we do a lot of work and then at the end we can't tell if it's a success or a failure. Then why would you start doing it?

That's funny. You and I were just having an email thread about app store marketing and trying to figure out how people actually found an app. And that would actually be an interesting way to answer that. And actually, I remember meeting you right as you were thinking about leaving Chicago and thinking about coming to the west coast. And we started talking about different growth jobs out there. And, you know, I'd been sort of in the, in the Growth world, working at Twitter and other places too. And we started talking about kind of this new world of growth and you got attracted to come out here and find a role for that. Talk about your process and how did you think about this transition from marketing to growth, from moving from Chicago to the Bay Area and landing on Pinterest?

Yeah. So GrubHub I think is a lot different from startups that are built out here. It was much more of a marketing driven company than a product driven company. In some ways, to its detriment, I probably sent 70% of my time successful multibillion dollar company. It worked out pretty well. I'm very happy with it. But you know, probably spent 70% of my time on more marketing based things or spending money and 30% of my time on organic things just because of limited engineering. We didn't even have a product team for the first four years and the first product manager actually came from the marketing team. So when you come out here you have kind of the opposite, like really strong product and engineering teams, very little investment in marketing. So a lot of the things that I was used to doing at the grubhub environment was were things that were sitting in product or these newfound growth teams instead of in marketing. So it just felt like more of a natural fit to be there. I didn't really think of the work I was doing that was changing a lot. Once I started doing it and figuring out that a lot of my job is like keeping engineers and designers happy, I was like, oh, this is kind of a bit different. So that was an adjustment for me, but it felt like more of a natural evolution. And I did look at running marketing teams and joining growth teams. And I talked to you and I talked to Dan about it and I think I got really great advice that you should go the place where you can have the most impact instead of saying that you're running a team or things that are based more around title or seniority.

Pinterest clearly at the time you joined was a really interesting network that was forming and still growing in shame ways. How were they thinking about growth when you joined and how did their thinking evolve and how did you help evolve that right at the very beginning?

Pinterest in the early days grew pretty dramatically through the Facebook open graph and through an invite only system. That is a great go to market strategy. It was great for growing the company for a while, but it didn't really scale. After a while, the invite only system and then Facebook open graph kind of went away as a strategy. It was a company kind of searching for its next act on growth. And it kind of backed into the fact that we have all this great content. We can start distributing it on other networks like Facebook and Google. And they needed someone that really understood how to do that. So I really started focusing on SEO there.

Okay, well, let's pause for a second. I'd love to dig in a little bit more that you say they had a lot of great content that was sort of there, talk a little bit more about what people were doing on Pinterest at the time. And the fact is, it was all staying very, very internal to Pinterest. Is that right?

Yeah, exactly. Pinterest started as this social catalog where you would sign up, you would install a Chrome extension, and then you would pin things into Pinterest to save to your own boards. But once thousands of people were doing that, it created this data set where you could start to recommend good content to people. And that became the much larger opportunity for Pinterest as this discovery tool for anything that you were interested in, really. So we had all of these people with great taste pin a bunch of content, and that created an opportunity to service content to the existing user base, and that became really the Pinterest product. But once that became the Pinterest product, we started asking, well, why can't we distribute this content outside of Pinterest to start to draw people into what Pinterest has? And originally, Facebook Open Graph was a good scenario for that. When that kind of went away, we started looking at Google as a potential way to distribute this content to get people introduced to Pinterest.

As you talked about Google and distributing on Google, I'm assuming what you really mean is SEO. And you've become. People have told me from Airbnb and other places that you advise. Your understanding of SEO is just you really get it, and you've been able to give a lot of people great advice about how to do it. Like, what did you start doing at Pinterest? And how does that translate more broadly for how other companies should think about it?

Yeah. So Pinterest had backed into a little bit of success with organic search, because people created these boards on specific topics like, you know, woodworking or Chukka boots or whatever. And it was really great content that they had pinned from other sources. So Google would start to show that to people that were searching for that on Google. The challenge was that a lot of people's boards were really random, and it was one person's opinion on what was interesting on a topic. But Pinterest had generated all this repin data on what globally were the most popular pins on pretty much every topic in the world. So one of the first things I worked on at Pinterest was creating these new topic pages that aggregated the best pins on each topic instead of just one person's opinion, and making those available to Google to index. And that worked really well. But the challenge that we were finding is that Pinterest is a collection of images. And when Google crawls a page with a bunch of images, they don't know what the page is about because they're looking for text to figure out what it's relevant for. So we started to think about how can we aggregate the textual descriptions that some people write and a lot of people don't, onto each one of these pins to give it more of a full fledged understanding to Google as to what each pin is about. So now if Josh and I pin the same pin, you'll see both Josh and I's description underneath that pin everywhere you see it on Pinterest. And that specifically helps Google understand like, oh, here are all the things this is relevant for. And then you start applying that in a lot of different ways and you start getting a page that has no text to a page that has a lot of text. That's obviously a very Pinterest specific strategy. If I tried to apply that to Quora, it would not help a whole lot because they have paragraphs of text on every page. So a lot of the future of SEO is figuring out how Google understands your specific site and what it needs to understand it better.

Cool. So how did those Pinterest topic pages, were those really the pages that started driving a lot of your SEO traffic? How did you get those ranked? How did you get those up to for certain queries, this is the question that every company asks. And Google feels like such a black box.

Yeah, there's really two main factors when you think about search engines. There's authority and there's relevance. Authority is driven by how many sites linked to you and how good are those sites. Pinterest is very fortunate in that regard, in that a ton of people link to their Pinterest content. It has a very high domain authority. So at least in the US that was not something we needed to worry about. When we started expanding internationally, we kind of had to rebuild that reputation from scratch. So we really needed to focus on the relevance problem, which is we have all these images collected. Google doesn't know what they are. How do we start to aggregate all of this Text. And then how do we distribute those pages so that Google finds them and starts to learn about them? So on every one of the kind of board pages that were already doing well with Google, we started adding links to these topics at the top, which were useful for users because if they wanted to really dive in on one of the topics the board had some content on, they could click on it and go to that page. But it also helped Google say, oh, this page is there. It actually has more relevant content to the page I'm currently indexing. I should go rank that instead. And it really took from. We started working on that at the beginning of the year and. And you know, so it's basically zero percent of traffic. And by the end of the year, we had gotten to 20% of our traffic from search engines was to those pages. So it was a really quick ramp up. And you could tell that Google knew Pinterest was powerful and important. It just really needed to understand a little bit what for. And then it had the excuse to rank the page as well.

And let's just be clear. I mean, it's an Incredible journey from 0 to 20% of your traffic in a year. But most startups try to operate on the orders of like, months. Like, how long ahead do they really have to start to plan to see SEO traffic? Because it isn't just something you can change and see results over.

And yeah, it depends. So Pinterest now can change and see results overnight because, you know, Google is crawling it all the time. We built an experiment framework just for SEO so that we can measure those traffic changes. But what I usually advise startups is that it can take three to six months to start seeing a return. But at grubhub, it actually didn't take that long. So grubhub, we had a similar problem in that we would launch a new city, let's say Los Angeles. And by definition, we'd have no restaurants there. We'd have no pages for it. So what we would do is we would send our best salespeople there. They'd go sign up 50 LA restaurants. I would create pages for Los Angeles, all the neighborhoods and all the cuisines. So if someone typed in Santa Monica Thai, if we had five Thai restaurants there, we would show them, and then we would go to press and we would say, hey, we want to give your readers $10 off their first order. Just send them to this link and write about Grubhub and you know, it'll automatically apply. And then after 30 days that that URL would Redirect to our landing pages for SEO and that would be all the links we'd need. And then we'd be ranking number one in like a month. So if you have a smart strategy to build authority like that, then you can, you can rank very quickly. If you don't, Google has to start to trust you over time and it can take that three to six month period. Usually, you know, press is a very big friend in trying to build authority. It may not be the only thing you focus on. But we really tried to figure out how we could speed that up at grubhub and we had some success there.

That's really interesting to find these authoritative sites to link to you. We could talk about SEO for ages. And this is why I'm super excited to have you as a growth advisor, because you'll be helping a lot of our companies on this. But my last question just on SEO before we move on to some other topics, is talk a little about mobile SEO now, because a lot of what you built, I think in the earlier days at grubhub and mobile and Pinterest was really still on the web and a lot of people were still on desktops. Obviously, mobile SEO has transformed things quite a bit. And how are you thinking about that these days? And what do companies need to do differently or just make sure they're doing? And how are users searching on mobile? A little bit differently, yeah.

So the strategy we talked about before with topics is kind of this aggregation of content that you can view on a desktop. Right. What Google is doing now with mobile specifically is they have this new initiative called called AMP Accelerated Mobile Pages. And basically what they're trying to do is create a browsable experience on mobile that loads really fast because mobile load times are usually pretty terrible and AMP pages do reload really fast. They load like sub 1 second. So the big opportunity now for startups and incumbents alike is to take the content that they have on specific topics, make it AMP enabled, and start to get exposure in these new experiences Google is building for mobile. And that's something that a lot of people are aggressively pursuing. The other thing that really mattered for Pinterest in mobile SEO was we did this early test around do we keep people on the mobile website or do we push them to the app? And it turns out people that get to the app activate at three times the rate as people who are on the mobile website. So how do you elegantly push people into the mobile app? Became a really big strategy for the Pinterest growth team this is something that Google has been struggling with. How much do they want to allow publishers to to do that versus keep them on the web? And you have to pay attention to the latest Google rules. Now you can't have a full page interstitial or things like that. But Pinterest, one particular engineer on our team did at least 15 different iterations on how to effectively get people to get the mobile app. And that had an incredible impact on our business. So that's something that I've really seen startups under invest in and most startups that I see, the activation rate on the app is better. So getting people to apply drives more attention. It's better experience for users and better for the business. And that's something you really need to think about with your mobile SEO strategy. So now it's like, how do I get an amp enabled page and then how do I take people from there to the app experience once they've had a little bit of a taste of the product that I provide is kind of the new strategy that people are trying to figure out.

That's actually a perfect transition. As we start talking about growth. We've talked a lot about acquisition, especially about SEO from your background, but talk about sort of the overall growth teams. At Pinterest, you were managing kind of the acquisition and retention teams. How did you think about kind of that whole funnel from this SEO and traffic into acquisition activation, into actually retaining people and getting them habituated?

Yeah. So, I mean, the first step is figuring out what's the goal of the growth team. And the goal is engaged users. Right. The goal is not downloads or signups, it's actually engaged users. And then you have to model out how each of these different parts of the funnel drive to that goal. And what we saw is, when I started is acquisition was a problem. Right. We no longer had this Facebook distribution channel that worked so well. So we were having trouble introducing new people to the product. But once they were introduced, the activation rate and retention was really good. And then once we figured out how to scale SEO, it kind of shifted back towards, okay, now we're getting a lot more people introduced. They're getting introduced with maybe a more marginal understanding of what Pinterest does. How do we expand their understanding so that they retain really well? So we kind of yo yoed back and forth between what the bigger problems were. So on the retention side, you know, you first start with emails and notifications, which I think is where most growth teams start. But you'll find over time that that is, you know, a 5 to 10% improvement on your retention. But if you're looking to really scale things further than that, you have to invest in making the product simpler to understand. So we had the activation team at Pinterest really invest in removing a lot of the complexity of the product over time. So we got people to the core value as soon as possible, which is, you know, cool content on topics that you're interested in. Right. And multiple topics, not just the one topic, you know, from Google. So a lot of investment went there over time, and I'd say that got the most investment on the growth team once we had figured out how to crack top of the funnel again.

And talk a little bit about sort of your thoughts on onboarding, like how much of it is trying to catch the person in one session, how much of it is trying to catch them kind of in some period of multiple sessions. You know, how do you kind of bring them back before they really get sticky and find those couple topics of interest and are coming back on their own a lot.

What we saw at Pinterest is that you really need to accomplish showing the main value in the first session, or else there's no guarantee there will be a next session. And that does not mean you need to explain how everything works. It means you need to show them the value and you need to ask them the questions before showing the value to make sure you can show the value. So in the case of Pinterest, that meant we had to figure out what topics you cared about before we could show you a quality home feed. So we did that, but everything else went away. We didn't try to explain where pins came from or anything like that. We just tried to do the minimal amount to get you to a quality feed of content you cared about. Then we basically decided, what are the key actions that's going to make someone understand the value of Pinterest over the long term? And then it's just all about reinforcing that and just telling them what to do next to reinforce that value. So when you land on the Pinterest home feed now on mobile, it'll say, scroll to view pins. And then when you stop, it'll say, click on a pin to get a closer look. And then when you click on a pin, it'll say that you can save this or that you can click to the source of the content. And that is literally the only contextual education you will receive. Everything about where the pins come from, how to add your own pins that you can share boards with other people, all of that is removed from the early product until we can make sure that you understand the really important parts of getting value and then we can slowly start to introduce those features later on. But we try to do that all in the first session.

I've often referred to that as a ladder of engagement of like, how do you actually get somebody up the ladder step by step as opposed to giving them everything at once. But I really like the idea that you talked about around asking them the right questions before they get into the product. I find a lot of people try to rush people into the product and actually if you just make it a little bit of a experience of kind of a back and forth learning questions, the product then offers a few more suggestions. Then all of a sudden somebody is like, oh, I kind of get this thing and it understands me because it's asked me a few things.

Yeah, you really need to back to that point, make sure the user understands why you're asking these questions. Right. That it's going to help them get value. And then you need to strike the balance of how long you're going to ask them questions before you try to deliver value. Because it is a funnel. Right. And the more questions you ask, the more there will be drop off. And that's something that the Pinterest activation team has played with a lot to try to strike the right balance between level, specificity and yet time to value.

I've often erred on the side of longer time to value, but more questions. And I've often seen that actually play out really positively with users that I've worked with. People are like, oh my God, don't add an extra step. And I've often seen that it actually works to actually add an extra step.

It's a bit of an irrational fear sometimes for sure.

We talked about this at the beginning and I think we rushed into so much of this great content about onboarding. But how do you actually think about measuring that kind of core retention, that core usage? How do you know somebody's really active?

Sure. So I think you have to map, you know, if you have a real world analog, you have to map to how often would people do this in the real world? And that's kind of the, the benchmark you start with. So at GrubHub, the benchmark was like people are already ordering delivery. How often do they do that? And let me map the retention of my product to that to see if I'm being successful. And that was people who are delivery once or twice a month was kind of the average. Our basic Metric of activation was 30 day repeat purchase rate. Did they place a second order within the first 30 days? If they did that, it seemed like they really understood the value and they appreciate it. If they didn't, they were probably at risk of churning with. Pinterest is a little bit different. You know, cataloging and pinboarding is a very niche activity in the real world. That's obviously more people are doing it on Pinterest than ever did in the real world. So we had to understand what is the unit of value that indicates that someone is having a positive experience on Pinterest and then what's the ideal frequency that we think someone should be doing it. And for a product like Pinterest, where it is, you know, finding things that you're interested in and topics you care about or doing projects on, specific additions to your home or to your wardrobe, we didn't think that was a daily use case necessarily, although there's a lot of daily usage on Pinterest, but we didn't think it was necessarily something you only did once every six months either. So we started with a monthly use case and it was basically, were you coming monthly? And then we figured out that's not actually a great metric because it's really easy once you start learning growth to get someone to come back monthly. It doesn't indicate that they really are sticky to the product. So then we switch to, okay, let's try to do weekly, which indicates that you're forming a habit and let's actually make it not that you're just coming back, but you're actually finding something to repin. So you found something that you really enjoyed. And while that is probably a smaller set of a larger amount of people that are probably getting value just browsing things, it was very clear that if someone's doing that, that they're finding value. And I think in growth teams and you probably have experience of this from your time, you set a metric for the company and. And if you get really good at that, you just make the metric harder. So if you're really good at making everyone monthly active users, then you switch to weekly active users and then you switch to daily active users. So I could see a time in the future where Pinterest does switch to a daily active metric, but that's basically that evolution of getting more and more sure that the key metric is a unit of value that every person is experiencing.

And you recently actually posted a great piece on Scaling a growth Team and you just kind of alluded to part of it too, where you Keep kind of moving out the mark of the metrics that you really want to optimize to keep increasing engagement and giving people value. Can you talk a little bit about the three stages of scale, as you wrote about in that post, and sort of how you think that applies to most startups?

Right. So there's basically three ways of thinking about growth. There is starting a growth team, there's evolving one once you've started one, and then there's like the mature ones that the very big companies like Facebook and now Pinterest is getting there as well. And when you think about starting a growth team, a lot of founders talk to us about when should they start a growth team? And my answer is, when you have product market fit. And your main question that keeps you up at night is not how do I get this product to work? But how do I introduce more people to it? That is when you should start thinking about building a growth team. And at that point, what happens a lot is people usually reach out to people like me and say, I want to start a growth team. Can you come in and be a VP of growth? Or something like that? And I say, don't bother hiring me. Just get like one engineer and designer to start working on a problem that's hampering your growth, and that's how you should really start a team. And what I found is that very few actually start with a designer, but it's really beneficial if that happens. So the core group that I would think about in starting a growth team is an engineer, designer, PM and an analyst. And they should sit together and they should pick one problem to focus on. I think at Twitter that was onboarding, if I recall, at Pinterest, when I started, we actually focused on conversion because we pretty quickly were getting more traffic from SEO. And then it was, how do we get these people to introduce to the real product, which is behind a signup wall? And if you can then show some success on moving that original metric, and I advise that companies that it should be a metric that can have a high iteration cycle and that you don't have a long lead time to learn if you're successful. So onboarding is a little bit harder in that regard. Things like conversion or emails are easier. Then you can go to the CEO and say, look at all the success we've had with this. And then if that CEO is good, he or she will say, how do I get more? And you say, well, let's expand into a few more problems with a bigger team. And that's usually what like the second evolution of a growth team looks like, which is you start working not just on conversion, but getting more traffic, activating traffic better, retaining that traffic better, or creating more virality. And that's what the growth team started to look like at Pinterest is we have these sub teams in these different areas focused on different parts of the funnel. And when you do that, the important thing is kind of to separate yourself from core product and say, here are the parts of the product that the growth team owns and here are the parts of the product that the core product team owns so that the growth team can move really fast without having to ask permission from some other product manager who has a totally different roadmap. And the other thing that happens when you start building a team like that is you have to invest in your own infrastructure as well. If you're going to try to maintain like your own emails, you need to own that email system and make sure the uptime is really good and all things like that. Whereas when you're, you know, just a couple engineers, you're just kind of hacking away on stuff and just trying to show improvement in a metric. So that's kind of the second phase of a growth team. And then the third phase is kind of moving away from this. There are specific things that are growth and there are specific things that are non growth to saying what is the biggest friction the business has and how can we go remove that friction and anywhere it is in the product. And that's kind of where the Pinterest growth team has moved to. And that's certainly how the Facebook growth team and other large team growth teams operate, where they just look at the holistic business, they look at anything that's preventing them from growing faster, making more revenue, having more user growth, et cetera. And then they dispatch a team to those areas and try to make an impact. So there's no kind of like separate real estate from core product and growth. It's whatever the biggest problem is, let's go attack that. And the mission is I need to get as many people as possible to the value this product provides. What are all the roadblocks to that? How do I remove them?

So on one hand, core product is let's move the product forward in a bunch of ways. And then the growth team is let's remove friction from people getting into what's already there.

Yeah, so I basically think there are three types of product development now. There is pure innovation, which is creating new value, there is improving the current value of the product and then there's connecting as many people to the current value. And I think that third phase is where the growth team really operates.

Well, I love the way you just laid that out. That was gold for kind of how people should think about growth teams. So let's go all the way back to your college, where you learned marketing, where you learned brand, you learned spending on television as marketing. What's it going to take for them to start teaching this growth, this new discipline in those classes? Is this going to become sort of the new way? Or is this something that's really unique to the type of companies that you and I have worked at or work on?

Well, you know, I think an investment thesis when you're in venture capital is that almost every sector is becoming the tech sector, right? So I don't think it's like, oh, this special niche that only a few companies care about. So I think, you know, Sean Ellis, who invented the term growth hacking, I saw Super Bowl Sunday, was actually teaching a class to l' Oreal on growth hacking. So I think across every industry, people are thinking about how does growth as a function make sense here? And I think it's broadly applicable to almost every company in terms of learning growth and learning traditional online marketing techniques versus brand marketing techniques. I've started to see it seep into universities, but the challenge in general with university education these days is that you don't have as many practitioners being the teachers, or if they are, it's been 10 years since they have practiced, and the rate of innovation is so quick that by the time they're teaching SEO, SEO might not be an important tool to grow anymore. So I think a lot of this is going to happen via on the job learning, just getting in and trying to solve problems. And there are specific institutions coming out that are focused on just these things. So Brian Balfour and Andrew Chen now have this program called Reforge that's teaching people a few years into their career how to grow companies. Then I think that's a very successful model, that it's very easy for a CEO to say, should I send one of my product managers to that course? And if they learn how to grow my business faster, then it's money well spent. And I think we're seeing a lot of online learning in general start to take over. So I'm thinking about, if I have kids by the time that they are in college, is it worth for them to go to a university, or should they take one of the myriads of online courses in the things that they actually care about and Is that a more efficient model? My guess is that's probably where things move, but you probably know a lot more about that thinking about those problems every day than I do.

It's a great set of thinking, though, about where this is all going to go. And really I'm interested in when this growth learning becomes more of just an art that's sort of taught in Silicon Valley to something that is sort of more broad and mainstream. So my last question for this that I'd love is, we talked about SEO, we also talked about social, and you mentioned Facebook in the open graph a few times. What are you looking at as the new ways of growth? What are some of the really interesting new platforms or new techniques that you're just starting to see right now?

Yeah, there's definitely a lot with messaging right now. And the challenge I'm seeing is back when we started working in tech, you had these platforms that were fairly open and they allowed, you know, companies to build really strong businesses off the top of them because they were fairly open. And now with these new platforms, they are becoming a lot more restricted. And it seems like the platforms are comfortable, you know, capturing more of the value instead of just creating a lot of value and capturing a very little amount. And you see that with Google as well. They're starting to serve a lot of their own results instead of, you know, 10 blue links. So trying to figure out how do you create a value exchange with these new platforms so that they're not just going to, you know, kind of shut you down very quickly. So the concept of, like building partnerships has come back into the fore. It's something I worked on very early in my career. And then it's like, I don't need to mess with talking to anyone else. I can do all of this stuff without talking to anyone. Now it seems like you need to build these relationships with Facebook and Google to really understand how you can create win win situations for them, because the default is maybe not that much of a win for startups anymore. So I'm thinking about that a lot with messaging. So with messenger and WhatsApp, what are the opportunities there? So that's been really interesting, I think. Is Snapchat going to become a viable platform? Is something that I'm thinking a lot about. I'm over 30, so it's still really hard for me to understand, but the kids love it. So how can you basically, in a mostly mobile first environment, create distribution channels where you can have your content or your message seen without having to spend money on advertising? Is a lot of what I'm thinking about, and it's become a very hard problem. So I think what we've seen is a lot of startups start with just doing a lot of the traditional app advertising techniques, like they're spending a lot of money on Facebook ads or whatnot. And that can be great if you happen to have a high lifetime value or something like that. But with a lot of companies that don't, it isn't super scalable or it's too expensive at scale. So thinking about, is there a way that Facebook would show you for free? So one of the things we're working on at Pinterest is can you create, like, topical Facebook pages that distribute content? And that's been a very interesting strategy that has some organic distribution. It's not 100% organic, but there is something. So those are kind of some two areas I'm thinking about. And I don't have a silver bullet, but it's an area where we need to learn more in order for the new breed of startups to be successful.

Well, I don't think growth ever has a silver bullet. If you've listened to this whole podcast, you've seen all the hard work that's gone into, all the work that Casey's done. So thank you so much for joining us. It's been really informative.

Yeah, thanks for having me.