Formspring - A Postmortem
This Sunday, after nearly three and a half years, Formspring is shutting down. While it’s a sad moment for me to see a startup I worked on and loved close its doors, I can’t help but also look back and examine ways we could have steered the product toward a more successful outcome. What follows are a few points that I’ve been reflecting on ever since Ade announced Formspring’s closure that I hope will be helpful for current and future entrepreneurs. To be clear, my critique is of product decisions we all made together, as a company (while I was there). As we would have shared in its success, so too shall we share and learn from its end.
We protected anonymous content to a fault
Formspring’s initial success was, in large part, due to giving our users the ability to ask each other questions anonymously (even without a Formspring account). In under a year, we skyrocketed to our first billion questions answered and showed few signs of slowing down. Yet even as we celebrated these milestones, we were all discussing how anonymity would or wouldn’t play a part in the future of our product. On the one hand, anonymity was a really popular feature (duh). On the other hand, we saw a lot of bad and abusive content come through that channel (double duh). A fact that we wound up being pretty infamous for.
But man was it hard to let go of anonymity as a core feature. We tried workaround after workaround. We prompted for sign-up after asking an anonymous question. We started pushing privacy settings for users into our on-boarding (which they never changed, of course). We started setting up elaborate filters to catch bad or abusive questions and put them behind a “Flagged Questions” link in users’ inboxes.
We spent a lot of time on anonymity. It was our sacred cow. Looking back, we should have spent that time finding ways to gracefully degrade that feature instead of finding ways to keep it alive. When you find yourself constantly giving a feature CPR, you should stop and consider whether or not it’s worth saving (or even possible to save).
Our opaque follow-model shot us in the foot
In a way, this lines up with our stance on anonymity. Following on Formspring was, for years completely anonymous. You couldn’t see who followed you and others couldn’t see that you were following them. This meant that we gave people a microphone and they kind of had to hope people heard what they were saying. And until we eventually launched our Smiles feature (akin to Facebook Likes), there was no way to know that your content was being consumed. We debated this a lot internally and came to the conclusion that the Twitter public-follow model was broken in that it put unnecessary social pressure on users to follow back. We felt we could build social features on top of the content (like Smiles) that let our users receive feedback and let their followers out themselves purposefully.
Formspring eventually allowed public following (not as a default, and after I left), but it was too little too late. My takeaway from this has been to always double check to make sure you’re not designing toward your own biases instead of what’s best for your product and users. Formspring had clearly struck a chord with people aching to share more about themselves with their friends. And instead of making it apparent that they were achieving their goal, we put an artificial barrier in place and prevented them from knowing if Formspring was working for them or not.
We skated toward the hockeystick
The biggest sin of them all from a product perspective, but also the hardest to avoid (and one that I see companies make over and over again).
Our initial graphs at Formspring, as you probably know, all hockeysticked up and to the right. Nearly straight up. That part was totally awesome! We were super popular! We could be the next [insert gigantic company name here]!
Oh wait, the graph has peaked and is starting to slowly (very slowly) trend downward. What do we do? Make big bets, right? Try to recapture that crazy growth!
And so we tried. The first big project we worked on was a Formspring button that sites could embed at the end of blog posts or other content. We had millions of users, so we figured it wasn’t a stretch to imagine they browsed other web sites and would gladly click a Formspring button at the end of a post (which asked “What did you think?” and allowed them to post a response to their Formspring page). This was just as the Facebook Share and Twitter “Tweet This” buttons were appearing, so we figured it made perfect sense to follow who we viewed as our closest competitors at the time.
We literally spent months on that system. We had to make sure our servers could handle a potentially huge influx of traffic (we based our estimations on our main site’s traffic, which was honestly insane), had to design and implement the feature, make sure the implementation was easy for publishers, make deals with publishers, etc. We bet huge. On someone else’s (Facebook and Twitter’s) plan.
How about photos? Instagram had just come out and was killing it! Our users would surely be more engaged if they could represent their questions and answers visually, right? I worked on those designs for at least a month and a couple more were spent iterating on and actually implementing that work.
I actually don’t know exactly how that launch went (because I had moved on to Amazon right before it went out the door), but now it seems obvious that photos weren’t the killer feature we’d all hoped they would be. Oof.
Entrepreneurs: build your product, not someone else’s. The most successful products execute on a vision that aligns with their product’s and users’ goals. It’s hard to put blinders on when your stats are slowly coming down and you see other startups skyrocketing around you with various tactics and strategies. For the love of god, put them on. It’s the only way to build what you should instead of chasing others’ ideas.
Also, print out Paul Graham’s Startup Curve and put it on every wall in your office. It will be you. Buckle up.
Even in the face of Formspring’s shutdown, I’m grateful for having been a part of it. I learned so much from both the team we built and the product that I’ll carry with me the rest of my career. I learned that small changes can frequently (and surprisingly) have the biggest impacts. I learned that gigantic, sweeping changes make it hard to track what’s going right and what’s going wrong. I learned just how hard it is to keep a huge site up and operational (also what happens when all of EC2-East goes down). I learned that I love designing for people I can communicate with every day within the product.
But most of all, I learned to focus. To put the product vision first. To not try keeping up with what’s trending. To build your product for your users.
Thank you, Formspring. It was a wild ride.