On Thanksgiving, Pinterest's co-founder Ben Silbermann sent an email to his entire user base saying thanks. It was fitting, as Pinterest was born two years ago on Thanksgiving day 2009. Ben had been working on a website with a few friends, and his girlfriend came up with the name while they were watching TV. Pinterest officially launched to the world 4 months later.
Some startups go crazy with hype and users right after launch. And some don t. I don't know the founders, but I thought I d take apart Pinterest's story to discuss growth and virality in consumer web startups. Pinterest was not an overnight success. On the contrary, its growth was surprisingly modest after Turkey Day 2009.
Take a look at Pinterest's one-year traffic on Compete from Oct 2010 to Oct 2011, which is the picture in this post, and shows Pinterest rising from 40,000 to 3.2 million monthly unique visitors. I took both ends of this chart and estimated monthly compounded growth over Pinterest's lifetime, then interpolated the curve using constant growth and put the results in this Google Spreadsheet.
Backing out of Compete's numbers, we see Pinterest grew about 50% month over month from a base of zero since its inception (on average, smoothing the curve). Today growth is catching fire, as evidenced by the near doubling of traffic last month, and Pinterest's page views growing 20X since June, according to comScore.
Note these numbers are approximations and also do not count the significant traffic the service sees from mobile (Pinterest's app currently takes the #6 social spot in the iTunes store). Also my guess is that a lot of its unique visitors arrive out of network (from Facebook / Twitter), and many of these uniques leave Pinterest without registering (more on this below) so it's tough to know their exact user numbers.
But let's play pretend and use the data we have to do some projections on where Pinterest could be a year from today. Its recent VCs certainly did this, and decided to give the startup a $200M+ valuation. Ron Conway said recently that Pinterest is growing like Facebook did in 2006. Facebook actually grew from 14 million uniques to 26 million uniques from May 2006 to May 2007, then a year and half later they had rocketed to 140 million uniques, and were growing at about 20 million uniques per month. So monthly growth early on for Facebook was around 10-15%.
Can Pinterest really sustain its wild 45% monthly growth? No, unless it's destined to be the fastest growing startup in history. However, we can be pretty sure Pinterest has hit a tipping point their page view numbers are simply insane. If they were to grow 20% month/month over the next year, Pinterest would be at 30 million uniques a year from today. And with 25% month/month growth, they d be at 50 million. These are pure guesses, but Ron's statements and last month's growth make this look possible, so let's examine virality and Pinterest's underlying fundamentals.
Virality Does My Startup Really Need It?
Viral sharing is typically emphasized in products or services where the cost to acquire a customer needs to be low, and you can t afford to spend money on paid channels, like Mint.com did, or like a Living Social / Groupon. Having a free method, ideally a user-driven method, is critical to consumer web startups.
But there's a lot of confusion around virality. The reality is that you can build a sustainable business without going viral and this point is not understood well among techies or investors. The connotation going viral typically means having a viral growth coefficient of greater than 1. For every user that comes on your platform, he or she refers 1 additional user. This ensures a service will hockey stick .
But what if every user only refers, say 0.2 new users? Contrary to popular belief, this can also lead to a sustainable business. However, the lack of pure virality implies that you absolutely must retain existing users to grow. This is why daily active users and monthly active users are such important metrics, and are tracked maniacally by CEOs and investors. Churn is your ultimate enemy. Sure, it will take longer to grow if each user brings in fewer new people, but as long as most users don't leave, you re all good.
Viral cycle times also factor in the shorter the referrer period, the faster the virality takes hold for example, if a user invites 1 new user every month, that's better than if she invites 1 per year. Social games like Zynga historically have extremely short viral cycles; however they must, because churn is extremely high (user lifetime is often measured in hours or days, because users get bored with games quickly and move on).
Obviously, the more viral a service, the more sustainable it is, but it's really in the details. And overnight success is not a guarantee for sustainability. Many startups are pushing way too prematurely on press before they ve demonstrated real sustainability. I see this all the time and Eric Ries covered this in his discussion on the danger of vanity metrics.
The truth is that startups often draft off of artificial success that originates simply from hype within the tech echo-chamber. Some sites go from 0 to 100K+ visitors in weeks and people high five one other, then look around and say what's next, how do we keep these users happy?
Pinterest's Virality and Sharing Examined
Pinterest's story is much different they didn t have the same early hype spikes as many other startups do, standing at only 40K uniques 8 months after launch! It took Pinterest quite a while for a network-effect to take hold. Clearly every startup should hope for early virality. But if it doesn't exist initially, you must work to perfect a soft onboarding of virality that's based on high engagement, and create a product that people love and will come back to, while layering viral techniques on top of that.
Today, Pinterest is clearly insanely addictive among its user base, and they are sharing. But one reason I see Pinterest as a valuable a case study is precisely because they didn t experience the early adopter hype spike 18 months ago. Like many tech startups, I am sure some content was seeded within the tech community by the founders. But their normal user was a housewife in the Midwest, not a techie reading TechCrunch or Hacker News. And I'm sure Pinterest made a bunch of product tweaks early on to iterate around sharing and engagement, as virality took hold.
Let's compare Pinterest to Instagram for a minute, since both base off of photo content as their primary unit. Instagram has actually done nearly all its user acquisition virally out of network (sharing to Facebook and Twitter) and via word of mouth, with limited sharing in the network. This is remarkable when you think about it. By that I mean that there's no real sharing inside the network other than liked photos surfacing due to popularity, and manual discovery via hashtags. There is no regram if you will. You also can t follow people or like a photo from Instagram without an iPhone (even if it's tweeted or posted to the web!) unless you re a power user on an API mashup (see Inkstagram). All this reveals just how impressive Instagram's growth has been, as they went from about 1 million to 10 million users over the past year.
Being realistic, Pinterest could have much higher growth than Instagram based on the fact that it's unconstrained as a platform it works from web, mobile web, in-app, and has easier baked-in virality around sharing. Pinning has this built in because many initial pins start as repins of other people's content. In this way, existing content will often be the seed for a new user's stream. The pin unit is genius.
Users can also visit the Pinterest site and participate (i.e. browse endlessly) before signing up. This allows full consumptive access before registering and is a secret weapon if done right. Fred Wilson discussed this recently in his post about the logged out user, giving an example of his mom checking his tweet stream without logging into Twitter she gets value out of the service while ironically bringing down Twitter's monthly active user metric. Similar value can be granted to consumptive users who visit the site but don't initially register.
Then when a user is ready to pin content, they create an account and go wild Pinterest leverages web content from Tumblr like no site that has ever existed, thus riding on top of its network-effect while not requiring user generated content like many services. They ve also perfected in-network virality (pin, repin, like) in addition to out of network sharing (Facebook, Twitter) to grow virally. For these reasons Pinterest could conceivably grow as fast as any consumer service we have seen in recent memory. It's fun to speculate on all this when you factor in Ron Conway's statements comparing growth to Facebook's early heydays.
And perhaps most notably, though it will surely take a while, Pinterest is already threatening to monetize, as those Midwest housewives are literally using it for shopping discovery, which Pinterest can profit off of by taking attribution for purchases that originate off its platform. I know several friends who ve purchased stuff spontaneously via random discovery on the site. I expect Pinterest to be thriving a year from now (my guess is 30 million users next Thanksgiving) and also spawn hundreds of copycat startups in other verticals ( Pinterest for that ). Sadly, many of these will arrive on TechCrunch and spike in hype, then fail to nail any true virality before they are slowly forgotten… After all, this is the cycle of consumer startups.