When we were initially developing Clay.io, we chose to only use Facebook, Twitter, and Email as login options... leaving out Google Federated Login. Google Login is something I don't see anywhere near as much as I do Facebook and Twitter, which is why I assumed it could be a lower priority. It wasn't until we implemented it that I realized we were wrong.
I thought it would be helpful if I posted some statistics of the signups we're getting per login/signup method.
Here is how the signups break down for our 1,058 users at the time I collected data:
We didn't implement Google Federated Login until we were already at 200 users. Below are the stats for our most recent 835 users.
As percentages: 24.7% Facebook, 10.5% Twitter, 22.3% Google and 42.5% Email.
The main thing I would say to take from this is: using Google as a third option for login is certainly worthwhile. If you are hoping to get the most people signing up, I would use all 3 (+email). If you need to cut one out, give the ax to Twitter. Of course, there is still a lot of value in Twitter (for example if you want to look at their followers and find people within your site to recommend).
Implementing Google as a login option was relatively painless.
The most popular option is still email, and there are sites that don't allow that as an option... that's not very smart.
I will mention that a lot of our audience has been from Hacker News, Reddit, and other early adopters. For a more broad population, it's hard to say how it would be different. Even fewer Twitter users I'd imagine, and perhaps a lower percentage of Google as well.
Some extra stats:
A graph of activity (basically how much a given user plays the games on Clay.io) per signup method. With such a small sample size, a few really active users can throw this off, so don't read to much into it:
Users on Clay.io can 'upgrade' their account to developer status. It basically just lets us know they're interested in our API and the development side of things. Here's the same analysis applied to all 143 developer accounts:
It's a small sample size, but it's clear the developer types prefer to use email much more than the average user. The percentages are: 12.6% Facebook (half as frequent as the typical user), 9.1% Twitter, 17.5% Google and 60.1% email.
For reference, this is what our signup form looks like. We don't put too much emphasis on the 'social' options, but they are there. In fact, more of our new users come from those 3 options than through email (and we aren't asking for a lot of information with the email option).
It would be interesting to see how these statistics would look on a much larger site... Any takers on HN?