Siqi was the founder of Serious Business, which was acquired by Zynga. At Zynga he served as the product manager, and he is now working on his new startup, Hey, Inc. He is an expert in virality, successfully growing products to millions of users.
→ Mafia was created as a game on Facebook almost by accident
→ His expert in virality and successful growing products
→ He learns about growth from being at Zynga
→ The game grew quickly to 100,000 daily active users
→ The success of Mafia led to her co-founding another game called Friends or Sale
→ Working full-time for a search engine startup
→ Mafia generated significant ad revenue, with his earning $1,000 a day from it
→ The experience with Mafia didn’t give her a lot of knowledge about viral growth or growth hacking
→ The game grew so fast that the metrics platform was never used and instead the focus was on scaling the game
→ They made the viral loops a core part of the game mechanics
→ And a whole lot more
→ Hey Inc
Bronson: Welcome to another episode of Growth Hacker TV. I brought in Taylor and today I have Siki Chin with us. Siki, thank you so much for coming on the program.
Siqi: Thanks so much for having me. I’m really excited and.
Bronson: Absolutely I think it’s going to be a great interview. You’re an entrepreneur. Start your own company. Sold that company, starting a new company. You kind of been there, done that, and you’ve grown a lot of things. So I think there’s gonna be a lot of value in this interview for people. But let’s start at the beginning because it seems like you kind of got into the growth stage almost on accident, and maybe I read that wrong, but it seems that you kind of fell into this growth stuff. Talk to us about Mafia. How does that. It seems like the first thing that really grew that you were a part of.
Siqi: Yeah, it definitely was an accident. So I started the Mafia about a month after the Facebook platform came out. And actually a week before it came out, I was playing a game of diplomacy with some white commentator friends I had when I first moved to the city, and we spent a whole day and all night planned. It was like a 12 hour game. I think it was like a 2 hours return. And right about then was was when they announced the platform. And I thought games was going to be a pretty big category there. And I thought a game like diplomacy would have been a really good fit for the platform, given that is social and it’s asynchronous. So I wanted to build diplomacy for the future platform, but that ended up being too difficult. So I said I built Mafia because I had to do his vote, but I had to think sort of like game mechanics in terms of, oh, it’s like it’s social involves, like people interacting with people. Yeah. So that took about two weeks to set up and I launched it and invited some of my friends to play and it just pure organically from there. Those were the good old days of the Facebook platform where pretty much anything you could grow if you had this little bit of our opportunity, particularly that was when Facebook invites and requests for still an API call so you could just do it on a back end. So that grew to about 100,000 daily active users pretty quickly, and I was at the time working full time for a company called Parasite, which was a search engine start up. And on the side I was making like $1,000 a day or more in like ad revenue. And so word got around that eventually my co-founding said, Hey, let’s build something together. We built the next game, Friends or Sale.
Bronson: What did you learn about viral growth during this time? I mean, is the lesson just be early on new platforms or is there anything more fundamental that you walked away from learning from Mafia?
Siqi: So at this stage, there really wasn’t a whole lot of growth hacking earnings or growth learnings because the platform itself was so easy at a time when you really didn’t have to know a lot or spend a lot of time thinking about it. At the same time, I was only one person and I just built this game as a hobby, so that wasn’t quite yeah, I didn’t quite get a lot out of it, but I was very surprised that we were able to get, you know, 100,000 daily active users just from invites. But for our next game, you know, the process the built in lobby has got did get me thinking about what are what is the infrastructure required for like success on a social platform. So for the next game, the first thing that Alex and I did is we built a metrics platform before we built a game. So we built this dashboard that plugs right into sequel with our standard tables, and we’re able to track growth, the referrals, retention. And from there, we built friends for sale. But the funny thing is we built this. We spent a bunch of time building the metrics platform and we ended up never using it because, you know, when you tune it, right, and the APIs are fine and then you end up growing faster than you can scale it. So we’re like, okay, well we’re not going to scale this metrics platform and we’re going to scale friends or scale because now we’re getting 100,000 users every day and or having trouble even staying up.
Bronson: Yeah. Now that’s so interesting. I just had the CEO of Bump on the program and he said the same thing. He said he doesn’t actually have good data about his company because they’re growing so fast. They don’t they don’t really care about it. So it’s like it’s so interesting that when things really, really work, metrics don’t matter as much in some kind of counterintuitive way. So it’s interesting to hear you say the same exact thing. How many people eventually did you acquire with friends for sale?
Siqi: So at our peak we had it’s close to 5 million daily users. We’ve reached just about everyone on the Facebook platform at the time. This is late 2007 reached over 100 million unique users. Wow. And monthly active users. I think we’re at at our peak 15 to 20.
Siqi: Monthly active. And this is the only advertising or growth hacking we’ve ever done on really just advertising, was we? Alex and I each invited like 30 of our friends to the app. That’s all we’ve ever done. And it’s just from there.
Bronson: Wow. And so, you know, with the first with Morphe, you said there wasn’t a lot of takeaways because it was so easy. Is it the same thing with Friends for sale? There’s just not a lot of takeaways because it really was just invite 30 people and watch the fire grow.
Siqi: So I think we were inspired a little bit around game mechanics for distribution and also because Friendster Cell Test stuck around for a long time, eventually was financing our entire company. We ended up working on it for a lot longer. So. So there are a lot more learnings from the sale. And we actually built out this dashboard we call Serious Infection that I’ve shown around in presentations before that we scale to supported that kind of infrastructure. But the first thing that we learned and friends sell is like, well, this is a hypothesis we had before we built Friendster cell, which was that if we can fully make use of the face of social graph and make it feel like your friends are already users, which we felt was a core strength of like the platform itself. Right? Like it solves sort of the cold start problem of when you build a new social product, how do you get anyone on it and how do you not have an empty room? And I think this actually I talked to a number of social workers and they call this technique this Francisco hack, because we were actually the first people to do this, which was that make all of your friends look like you’re already played in a game and you can just go and interact with them. And every interaction ends up being an implicit social communication, right? So the really, really simple mutation that we had was okay in before to face a platform if you were to build something like friends or cell. And for people who don’t know friends or cell relatives, it’s basically a game where you buy and sell people in this market economy and their prices go up and down and you you make them their pets. And before two phases, like you try to make a game like that, then you probably make people register. And then when you go in, you can see like all of the million users that you have who you can buy and then maybe you can see like two of your friends if you connect with Facebook, right? But with the Facebook platform, we just made all your entire social graph purchasable, right? Like it looks like every one of your 500 friends are playing already. They all have a price and you can click a button and by clicking in money you actually just bite him directly. I don’t know, back in his environ notification and he notified with this really mysterious notification, Hey, your friend Ryan just bought you for $500. You wanted to buy him. So that works out really, really well. And people came in and just through the process of playing the game, we were able to grow. So the two learnings are one is, hey, it’s really, really effective to make it seem like all of your friends are already playing and be able to directly interact. And two, if you can make your viral loops core to the game mechanic self. In our case, our game was all about, you know, buying your friends. So if through that action you get a foreign application, then that’s like a really strong organic growth channel.
Bronson: Yeah, that’s great. And you mentioned something there. Also, you mentioned infectious, that dashboard that you that right.
Siqi: A serious infection because our company was called serious business.
Bronson: Yeah. So tell me about that.
Siqi: So serious infection is a metrics platform that we built about a year into serious business. And what it did is so the most useful thing was that it did a B testing for viral copy automatically. So that was like really interesting. So what we did is we wrote our own tackling language and we use a simple language for request, copy, notification, copy, all our viral copies. And so in our code, all we had to do was have an if statement and have one version of copy in, in one version, a copy, another version copy. And then it would go through our server and our server would or able was able to detect that these two copies are different and automatically create an experiment for it automatically generate results. So you know, it just know like a B testing set up to just you change the text and you automatically create a B test, which was really, really convenient, really, really powerful. So we were just able to like quickly create a statement with like ten different copy and it would automatically have the experiment completely set up. And eventually we wanted to get to a point where it automatically determined the results too, so we’d be auto optimizing. So that was a main feature. But it supports it was effective. Contagious. And the Canadian guys actually told me that they base contagion off of what we shared about our own metric system.
Bronson: Yeah. Is that a platform still available for use or is it just something you used internally back then?
Siqi: That’s not the IP that Zynga acquired after we were acquired.
Bronson: There you go. So let’s talk about that. So your company series business was eventually acquired by Zynga. Now, you were at Zynga during the height of their growth when they were just, you know, taking market share like crazy, growing like crazy. What did you learn about growth from being at Zynga? I mean, you already had growth kind of in your veins when you got there. But what did you learn from maybe their culture or their philosophies or their unique way of viewing the world that really kind of added to your repertoire?
Siqi: Wow. So much. I mean, Zynga, they were obviously the best in the industry, so they knew just new things people didn’t know. But in particular, I think there’s sort of this a cultural difference between Zynga and I think most smaller startups in the space. If you look at the independent developers, we sort of grew up probably bootstraps and we were able to get this small viral app to a million users just organically. And, you know, for a small team with not a lot of capital to get that kind of distribution, you it’s hard to not be influenced and get religious about it and I think slight in particular and but quite exemplified this. But you know this is endemic of all the independent other developers, which was that, oh, because we are good at growth hacking, we are religiously against paid user acquisition. Mm hmm. And I think the main thing about Zynga is that they are not in any way religious about anything. So when they launch something is paid, user acquisition is growth hacking. It’s game designed to make sure users stick. It’s everything. They throw the book at this product and they make it work. As a small startup, sometimes you can’t afford to do that depending on the business model. So the other thing I’ve learned is that you need to match your user acquisition strategy to the market that you’re in. So for a purely social product that relies on massive, massive network effects and you have to reach scale before you can see meaningful revenue like, say, a Twitter or Facebook, then you probably want to dial down on the page user acquisition and tweak the growth hacking lever way up. But for something like a social gain or content business in general, where you have a limited window to monetize your users because you know, it’s Farmville or CityVille or any game is not going to be around forever. The way you can expect, probably something like Twitter or Facebook is then you need to be able to monetize it quickly. And because you can can immediately monetize these users, then it’s worth it to get paid user acquisition early. So we didn’t understand the market. We were right. Like if we were to do it again, we would double down, pay you for this, your acquisition and juice up our lifetime value and try to grow that way.
Bronson: Yeah, no, that’s great insight. Is there anything that you kind of brought to the culture at Zynga? I mean, I know you said, you know, that you learn so much from being around them. What did you bring to the table? You were like, hey, guys, consider this. This might help.
Siqi: So two things. Well, I think the most important thing but this is something I don’t I didn’t bring fresh out from the outside. Right. Pressure in outside. I think the independent developers were probably better on average and Zynga as pure virality and what I mean by that is just new user acquisition installs. The thing about Zynga that most people didn’t know at the time was that Zynga was a player using these social channels for retention and engagement. And like that simple fact is really the cause of Zynga success above and beyond everyone else in the industry. Like I, you know, I’m pretty plugged in with most of the independent social gaming developers at a time. And that point was just something no one was aware of, really. And in fact, if it be if any of those people were aware of it, like we, a lot of those companies would have been competitive. Yeah, it’s so sometimes it’s the smallest things and sometimes, like, it’s obvious now about it. It’s I no one was thinking that way. I think the bigger yeah. But so that’s one thing we brought from the outside. But after being Zynga for about a year, the one thing I didn’t try to influence in terms of product strategy was the importance of retention. Yeah. And also you mentioned I had some counter-cultural views around product design, so some of those opinions started. It bubbling up too. So the limitations of metrics. How? I have this favorite thing, which is retention is actually the best characteristic because it’s sustainable. So that’s something that’s always able to influence a little bit in microlensing.
Bronson: Yeah, we’re going to dove into that a lot more in a little bit because I’m really interested in your views on those things. One more question about Zynga. You know, at one point in Zynga, you had a few different roles there, but at one point you served as the director of product. Do you think that product decisions affect the growth potential of a product or are there always more factors that matter more? How core is product to growth, I guess, is the ultimate question.
Siqi: I think it’s incredibly important to grow. So it depends on. So growth is sort of this time dependent variable, right? Like growth over what time period and growth how quickly and how sustainable overall quality variables for one of the variables for that term. So if you want sustainable long term growth, that will get you to Farmville numbers in a platform environment that isn’t Farmville like doesn’t have all the low friction like conversion or channels and product is pretty much everything.
Bronson: Mm hmm.
Siqi: And you can see the the you can see the contrast directly. Right? So if you look at mobile product services and products, right. Mobile products tend to be highly polished. The success of what seem to be highly polished, highly engaging. But the same time a mobile, you don’t have these like low friction distribution channels, but B But as a result, you have to rely on the inherent quality, your products to retain and get distribution through word of mouth, which is its own kind of our reality that people don’t like to talk about. But with when you have these highly ubiquitous social channels and incentives become, oh, let’s try to reduce the friction and why these growth plans, which is possible and often at the expense of product. Yeah.
Bronson: Now that’s great. So now let’s talk about some of your counter-cultural views on design. I’ve heard you say before that you think metrics can bias our design decisions. What do you mean by that? How do metrics bias our design decisions?
Siqi: So what I’ve seen in practice is that so some contacts are is a serious business was an intensely intensely data driven company and in fact some I was asked about at a conference about how can you measure everything online? What if you can measure everything? And my answer to that was there’s two kinds of companies, right? One is, you know, you look at what you can measure and because you can measure everything with just sort of important data and go with our gut. And the other type of company is, well, you can measure everything so and so let’s try to measure as much as we can and more and more over time, actually, eventually we’re going to get closer to measuring everything. And I said, serious business. Definitely the latter kind of company. I definitely wouldn’t build the company that way. So the reason why I say that is because once once you are intensely data driven and I know like you can say that, oh, you can try to avoid these problems or like saying these are sort of natural symptoms. Are you going to have to actively, actively work against a lot of it? And it’s hard to even be conscious of it when you sit in the middle of it. But the side effects are one is it sort of replaces thinking and communication and debate in a team, right? So if you’re talking about a design decision and instead of like talking through the pros and cons of it, how it affects the overall product, because I don’t think product decisions can be made just in isolation. They affect sort of this whole you just know, oh, let’s just test it, right? Like forget the incision. And then that becomes a replacement for thinking. It becomes a replacement for thinking about how things are exact over a product. It becomes a replacement for even thinking about are the things you’re testing even that important? Right. I think the 43 Shades of Blue, the classic Google story is a great example of it. Like, I just that’s just broken. And I think that’s a place where you naturally end up in. And if you’re not aware of the limitations of it, the other more practical issues around a pure data driven development is that the data you’re you’re naturally they’re going to be biased to building things that you can test very quickly. And not all things that deliver user value are of that class. Right. But you’re going to be biased against from doing those things that actually deliver value. And the other part is you’re also really biased towards building things that you can see the results of very quickly. And not everything that you build, you can see the results of quickly, right? So I think then Silberman said of the Lean Startup, but it’s like if I read this book when I was building Pinterest, Pinterest wouldn’t exist because I would have gave up, right? Because all the data says it wasn’t going to work, but it took time for it to work. And these are biases that I think are rather subtle. That’s if you’re not aware of it, then you end up with a company that’s purely data driven and end up with a bad product. And I don’t know if you’ve seen this, but there is an extremely high correlation between very data driven companies and products and very ugly products.
Bronson: Really never really thought.
Siqi: About it. But yeah, I mean, like, I’m not going to, I don’t want to name any products companies, but you know, the ones that seem like are talk about data and maybe testing everything are also the ones that just look like crap.
Bronson: That’s that’s great. That’s awesome to know. Then what role do you think metrics should play? Because you said, you know, serious business was very metrics driven in a lot of ways. You’re not saying metrics don’t have a role. You may be saying they have they’re not the main thing. They’re not the only thing. So what’s the right role for metrics? How should a company incorporate?
Siqi: Metrics are a tool. And so I think I just believe in balancing many, many things, whether it’s team management or process. And where I think metrics falls down is if there is no balance, right? It’s all about the data. All decisions are in a purely metric driven. Yeah, I am a huge fan of Facebook as a team and how they articulated a role of data. They didn’t call it everyone calls it data driven development, right? They call it data informed development and design, which I really think is accurate. Right. Metrics are important. You want to measure as much as you can, but it should not be the driving factor behind design and product decisions. But it’s in time. If you are making design decisions, it is very, very, very useful to respect and be aware of the data that you are.
Bronson: Yeah. Do you think that overall this newfound excitement around metrics is a net win or a net loss for startups? Do you think we’ve gained overall because of all this emphasis or not?
Siqi: I think I think it’s a net. Gain overall. If I would rather be slightly too data driven than having no data at all and just rely on instinct. I think that’s the state in which startups most targets were before metric. So lean startup. Yeah. So I think it’s in that one.
Bronson: Yeah, that’s right. So you talked about, you know, being data informed. So what are some of the tools that you personally use to inform yourself with the data? Maybe it’s apps you use, maybe it’s in-house things you build that you consistently do with different products. But what are some of the tools in your kind of toolkit?
Siqi: So at serious business is purely serious infection and raw cycle calls. And Zynga has their own internal analog systems you track, which is a very, very sophisticated and powerful. And I mean, it’s actually human powered, which is really interesting.
Bronson: What do you mean by that?
Siqi: So a team of analysts who write reports, customizable reports or any questions you want. So it’s you know, instead of having building this tool for our engineers, you can just do custom one off queries by throwing it over the analyst’s wall and these team of amazing, amazing people, all predecessor and partners for you, and they’ll save it and everyone else can use it. It’s a really interesting system.
Bronson: That more efficient than having an engineer designing something. Is that why they did it that way?
Siqi: So there is, you know, costs and benefits. So the costs are sort of, you know, there’s obviously the human cost of actually having it ultimately seamless. And also there’s some inefficiencies in that, you know, sometimes reports will be given twice, but the benefits are, I think, well worth it in at least in this case, given the size of the company, because any question that you have, you can it reduces the cost of getting the answer to it right. Because it just an engineering sort of driven metrics tool. Then you have the prioritizing, oh, is this question worth asking? Should we build it? But you don’t have to answer questions that you just like. I just want to know. And once you build it once, then everyone can just use it. Yeah. So it’s sort of brute force, but it works really, really well, which is sort of Zynga’s culture. So, yeah, I think it works pretty good.
Bronson: Yeah. Any other tools besides the ones we’ve listed there that you go to?
Siqi: So then when you start of the main one is probably Mixpanel. Yeah, really easy to integrate the tools work well, but eventually we’re going to have to roll our own. I mean I think any large consumer start up, it really gets to a point where you have to roll your own.
Bronson: Yeah. What do you where do you think that line ends? Like where does Mixpanel in and roll your own begin. What do you need that you think you’re going to push the limits of Mixpanel just to help our audience, because they may be in the same situation where they’re hitting the limit of certain things.
Siqi: Um, I think the function of scale and functionality so has to say a million users, it can just become cheaper to build your own.
Bronson: Because charging by the event calls. So that could become very pricey if you’re actually building something with with true growth.
Siqi: That’s right. Yeah. And if you want to do more sophisticated queries and just like really be testing or demographics or an overall funnel, if you want to get on raw user insights through writing custom queries, then you kind of have to roll your own.
Bronson: You know, that makes a lot sense. Now, one of your favorite quotes is from Steve Jobs. I think we all have a favorite Steve Jobs quote, somewhere in a few years. His design isn’t about how things look, it’s about how things work. And I think that kind of speaks to the things that you’ve been talking about here. But what does that mean to you?
Siqi: Well, the first part is that it just addresses a common misconception about what design is. Right. I think people see design as a surface level veneer. You put on products and they look at the iPhone or the iPod as, oh, this is an amazing skin on a music player. But I think what it hides for me is I’ve been getting into sort of like Zen Buddhism lately, and it sort of forms how I think about product and data to. It also suggests that great things, great products are an integral whole that can’t just be created and tested. Know aspect by aspect aspect. So how things work is, is not just the surface level. It’s everything about this product is how it all fits together is so fitting finish. It’s so sportsmanship is everything that you don’t do. And the third part is, it also suggests that design isn’t just graphics. It’s it’s problem solving. Mm hmm. Which. I enjoy. So I just like that quote a lot.
Bronson: Yeah. No, that’s a great quote. I’m a designer myself, so I love looking at that to kind of put things in perspective. Now, after Zynga, you’ve been working on your own stealth start up and it’s at Hey Dot CEO, which is a great domain name, by the way. I’m sure you already know that. And the name of the company is Heyday, is that correct?
Siqi: We’re now hailing and we’re making a product called Heyday.
Bronson: Okay. So, hey, Inc is making Heyday. So I know you haven’t released much. You know, I went on the Web side, I read the fake you. I looked at kind of what you had up there. I can maybe make some guesses as to what you guys are going to do or try to do. But what can you tell us about Heyday? Is there anything you can release about this app and what it’s going to do?
Siqi: Um, there is not much I can say about what the app is supposed to do. I can tell you a little about my, the genesis of this idea as it relates to growth hacking. I think it’s actually really, really appropriate. So after I left Zynga or even I was there, I was thinking about what I wanted to do for the next company. And I was looking at things that are successful. Right. And there are exactly three companies that have over $1,000,000,000 are market cap that were started in last ten years. It’s Facebook, Twitter and LinkedIn. That’s it. Wow. And all of these are have network effects. So it’s thinking about things that work and things that didn’t. So when I think about things that don’t work as well, I think about maybe some of the games we in a serious business where it was highly successful for a time, but eventually it has a life cycle and it dies and things like this all seem to have this attribute of being the acquire user and keep them forever. So the only criteria for our new idea, our new startup, is it had to be a product that we believe was going to be around forever and such that we can acquire user and have a pretty good idea that this person will use the product forever. Yeah. And so we were thinking about what are some ideas that could get us there. And one of the frameworks for coming once we had ideas like what are some really foundational human needs that are projects that can be improved by technology, right? Facebook is like social. Like everyone needs to interact with other people and make it better and make it more efficient to share. So for us, one of those foundational human needs is memory. So we’re trying to solve that with new technology.
Bronson: I’m very interested. When do you guys think you might be launching in any ideas.
Siqi: Within the year? Within that, we’re pretty far along. We have a small beta testing list and the feedback so far has been pretty strong.
Bronson: Yeah, that’s great. Well, we’re all waiting with bated breath to see your heyday. Now, you also make investments in startups from time to time. When you go to make an investment, I’m sure you’re looking at the startup to see if it has growth potential. I mean, before doing Heyday, you were thinking, okay, which, you know, what are the largest businesses ever in technology? And let me try to do something that can have that kind of market cap. What do you look at within a startup to kind of, you know, size them up to see if they have growth potential?
Siqi: So I switched to as a more formal process for me. But for me, it’s pretty intuitive. It’s a combination of product and team. I’m not a much of a market driven thesis kind of person. So the first thing is team like I’m going to Skype with them and see if there is like the spark, you know, like are they is their founder market founder product fit which I find pretty important. So in one case I invested in this great team called shirt me and they’re, they’re, they’re doing social dating on Facebook doing really well and the thing that that made me investors they were extremely creative growth hackers right and that’s sort of what you need to succeed in that kind of business. In this other case, I missed that in coming on flash soft and I just like the technology is just like fit into how I thought the current state of the art could be improved. So it’s fairly intuitive, but so far I got my money back at least, and the rest, you know, we’ll see.
Bronson: Well, that’s that’s more than most people can say who play the the investment game. So that’s good. You know, as you interact with startups because you interact with them all the time, what do you think that many startups get wrong about growth? Are there any themes that you see that consistently just off?
Siqi: There is. And the obvious one is, you know, there is this recent the open graph span system companies that like blown up. And then sort of just went nowhere. So this idea of growth at all costs and gross adds without thinking about the sustainability of it, seems to be a large waste of resources. So back to like my one of my other favorite quotes is retention is the best growth hack, right? Like, you know, any, any company can like bring a few attackers in from engineering to that growth hacking and figure out how to grow for a time. Right. That’s not a difficult part. I think that’s like knowledge is being increasingly commoditized as growth hacking becomes more popular but still difficult. And what skill is required to build a company of sustainable value is how do you make a product that users will use for a very, very long time? And I think that startups the mistake that a lot of startups make is they focus on growth over retention first. And I think growth, hacking and morality, it’s a force multiplier and you have to have something strong to multiply first.
Bronson: So it’s a great quote and that might be the quote of the episode that retention is the greatest growth hack. I love that one. This has been an awesome interview. Suki, is there any final advice that you can give to any startup is trying to grow? Any parting words of wisdom?
Siqi: I think I said that retention is about.
Bronson: There you go. I say we end on the high. No, that’s awesome. Well, Siki, thank you so much for for coming on the program and for taking time out your busy schedule to share your immense knowledge.
Siqi: Thank you. It’s been a pleasure.
Get the strategies, motivation, and in-depth interview with all the details every week!