Michael Seibel argues viral growth isn't magic but a measurable system: turn user actions into share opportunities and place sharing as early as possible in your funnel. Socialcam used Mixpanel to A/B test every variable, from how many items to show to share button design, transforming random viral spikes into predictable growth. The best founders admit what they don't know and let data decide.
Chapters
Viral marketing is an extremely shitty distribution mechanism. Yeah. It's really bad. Socialcam was successful at viral marketing for approximately 4 months. Had we not been bought, we would have been in a very, very bad position. There are a couple reasons why viral marketing is kind of shitty. Especially nowadays, a lot of viral marketing is based on the cooperation of a third party, whether it's Google for SEO, whether it's Facebook or Twitter for distribution. And those third parties constantly are changing the rules. Stuff that you share into Facebook right now will have a different click-through rate and a different number of impressions than something you shared into Facebook a year ago. And at Socialcam, the only reason why we got viral marketing up and running was because we actually could measure that, which very few companies do. And Facebook didn't really like that we could. And so with Facebook's full knowledge, basically, I would argue, semi-fighting against Facebook, we were able to get viral for 4 months. So with that being said, what do you do, right? You have to grow your product. Growing a product is really freaking hard. And I think that when I think about growing a product, one thing that I think about is, how does someone using my product turn into someone sharing my product? And when you think about growth, I think what most people don't realize is that the share, where the share is in the flow, So let's say to use my product, I want you to click 10 times. Well, every time you ask me to click, you're gonna lose some number of people. And so if that share step is on the 10th step, you're not getting a lot of shares. If that share step's on the first step, you're getting the most shares you can possibly get. And so a lot of the times I hear people talk about their product, let's say their product is like Someone pitched me they wanted to make a— a lot of people would make a custom recipe book. And they were like, this would be awesome. You take your family recipes, you make a custom recipe book, and then you share it. And I'm like, OK, well, how is that gonna get growth? And they're like, well, every time you make a recipe book, you share it. And I'm like, OK, well, how much time does it take to make a recipe book? Well, like a couple hours. How many recipe books are people going to make? One. So it's like, wait, it takes a couple hours for you to get a share, and every unique user can only be one share? How is that going to grow? Whereas if you conceptualize it a little bit differently, maybe instead of the first experience about your recipe book app is not making a recipe book, it's not even recipe book at all. It's browsing recipes. Right? Well, someone can browse 100 recipes or 20 recipes in a sitting, and they can do that every day. And then what happens if in the process of browsing a recipe they share it? Okay, well, now you've taken an activity that can happen a lot of times on your website, and let's be clear, one person can trigger this activity many times, and you turn that into a share. So that's a very, very important distinction. Look at actions in your product that one person can do many times over the course of one day. Those are the types of actions that should be distributing your product. Stop looking at the things that one person can do every week, or that's a culmination of tons of effort. Those things don't work for sharing. That's the— That's the first best insight I can give you on virality. I think the second best insight where everyone seems to screw up, where we screwed up with Socialcam at the beginning, is designing a viral flow is exactly like designing a product. You're going to fail at first. The only way to get a product from bad to good is through measuring. The only way to get a viral flow from bad to good is through measuring. And so— When we launched Socialcam, from day one, we integrated Mixpanel. How many of you guys are aware of Mixpanel? Beautiful. Love it. So basically, what Mixpanel allowed us to do first was instrument every single step of our funnel. The second thing it allowed us to do was A/B test various steps. So basically, imagine once again like a recipe book. If I show you 20 recipes on the page, are you more likely to share one of them than if I show you 10 than if I show you 40? If I can get some preferences from you, if I show you a personalized list of recipes, is that better than showing you a public list of the recipes that are popular across the entire group? These are all things that you want to be able to A/B test. These are all things that you can A/B test with Mixpanel. And so basically, like, if you build those analytics into your product from day one, And if every question and change that you anticipate or that you think up, you measure and you run it as an A/B test, you will not only make your product better as you respond to user wishes, but you'll also improve your growth because you'll optimize for shares. And the more shares you've got going out, the more you're going to grow. And then what's cool is the second you've instilled measuring in your product, and your team, it will take on all kinds of crazy dimensions. You know, it turns out if you publish a photo to Facebook in the context of a share— so you know those share stories on Facebook? That's a square photo and then like a block of text with a gray background, right? So it turns out if you put a play icon in the middle of that photo, you increase click-through rate by 25%. Now, it turns out that if you test 5 different play icons, you get results anywhere from negative to positive click-through rates. If you just copy, if you just see your competitor putting a play button on their shares and you just copy, God forbid you're watching their A/B test and you put in a share that gets you negative, but you have no idea why they chose that one. And you have no idea what the impact is. Once you instill data and measurement in your product and A/B testing, these things will happen naturally. You will naturally look at your share and say, how do I make it better? Let's try adding a play button. And you will see the results. And that will permeate every part of your product. And I have to say, as you guys grow your businesses, one of the things you have to make sure of as founders is how you deal with disputes. In companies that are run by what I call Steve Jobses, they're fundamentally either authoritarian or very confrontational companies. Basically, if there's some guy in your company who thinks that they know the answers to every question, the only thing gets built— the only things that get built are things that guy wants or things other guys argue with that guy and convince that guy. That is a really, really bad way to run a company. I like companies that are run by people who assume they don't know the answer. If you assume you don't know what your product looks like, if you assume you don't know what your users want, that is the beginning, that is the starting point for a very good product, because then the first thing you do is you start asking users what they want. The second thing you do is after you ask them, you measure your users to see if they actually do what they said they would do when you talked to them. The third thing is you implement one of their suggested changes and you measure in an A/B test, did they actually— when you give them what they want, did they result in doing more of what you wanted? All of that comes from not knowing the answer. If you're the kind of person who knows the answer— Steve Jobs was not successful for a very long time, let's be clear. He did not realize his vision for a while and took a lot of hits. I would get ready for a lot of hits if you think you know the answer. If you don't know the answer, you can be successful a lot faster.
