How Adding One Button Increased Our Signups by 3x
We’ve run quite a few experiments at Product Hunt in 2015 — however one of them stood out. Here’s how we tested the idea before implementing it.
At Product Hunt, I’ve been focusing on prototyping and running experiments, adopting a lightweight, iterative approach to improve the experience for our community.
One particular recurring internal discussion has been when to add a login option besides Twitter. For us Twitter plays an important role — our community loves Twitter and uses it to continue many of the conversations that begin on Product Hunt.
We’ve also built in-depth integrations, like ShareMeow:
Focusing on Twitter as the first platform made sense for us given our community’s use of Twitter and it would have taken a lot of effort to get it right on two platforms.
In December 2015 we decided to open our doors to the 1.6 Billion Facebook Users by launching Facebook Login, but not without a test.
Measuring Before You Build
First, in order to establish whether we should invest the time implementing Facebook Login into our website and iOS app, or not, we launched an experiment.
Everything we build requires tradeoffs. We wanted to be sure a significant number of people in the Product Hunt community actually wanted to login with Facebook, before investing the time to build it.
After all, if too few people wanted it, our time would be better spent elsewhere. We could ask the community — and we did — but as we all know, actions speak louder than words.
Quickly we learned, that nearly 60% of new users preferred Facebook.
Based on this success, we decided to sit down and plan the implementation.
As you can see below, the iOS App Submission was a big factor in getting this out before the end of the year — but Rado on our team got it done in record time.
The Devil is in The Details: Usernames
One of the more difficult problems to solve, was how we handle usernames on Product Hunt going forward.
Since the beginning we mirrored Twitter usernames with Product Hunt usernames. Adding Facebook to the mix complicates this slightly.
After taking a close look at how other sites, like Medium, Pinterest & Meerkat solve this, we’ve decided: New Facebook users will get a username based on their name, and they can change it during the signup process.
For the time being, and for an easier to understand overall user experience, username changes are limited to the initial signup flow.
This has worked surprisingly well, especially in light of all the other little details we also had to take care of.
The Success
It worked. We launched Facebook Login on December 18th, and since then it has clearly shown to have been the right choice, in both signups, as well as long-term retained users. [1]
We could have probably done it sooner, but then again the technical effort was non-trivial. The data has certainly helped us make the right decision.
If interested, please send me your questions below and if you’d like to check out the Facebook Login yourself, click here. ✨
😺🎉
P.S. Special thanks to the whole Facebook Partnerships team, and especially Konstantinos Papamiltiadis for supporting us & giving us hands-on feedback.
[1] It’s always important to measure a secondary, opposing metric (retained users, # of followers for new users, etc) to the primary one you’re trying to increase (signups). This ensures that your secondary metric does not go down as the other one increases.
