Quantifying the impact of offering alternative signup/login methods?

Question

Nowadays it's common for websites to offer signup/signin methods that don't require making an account (email/password) for that specific site, for example 'Login with x':

• Facebook
• Google
• Twitter
• LinkedIn
• etc

By how much does offering each common alternative signup method increase the signup rate for an average website, as opposed to just offering traditional email/password signup?

Notes:

• Obviously these statistics could differ (a lot) from site to site, depending on the demographic and other factors. But I currently have very little idea of how much impact each signup method has, so any information is a massive improvement on that

• Very interested to know if any studies have been conducted, or if any companies (big or small) have released statistics.

Example

Suppose a website only offered the traditional username/password signup method and it signed up 1% of the visitors to the site, but then it got a 'Login with Google' feature, and its signup rate went to 1.1%, then we could very roughly guess the 'Login with Google' feature increased the signup rate by 10%. (figures completely made up)

Answer 1

I wouldn't have normally responded on this, but you said "any information is a massive improvement." While I can't speak to signup conversion rates, I can speak to an angle which you may not have thought of which, for me, adds a bit of trust to a name I haven't heard of, especially in a time of information/password leaks.

Security and familiarity. People know Facebook/Google/Twitter, etc. If a user signs up through those services, their benefit is two-fold. 1.) They can see or strip your access to their information at any time (using a familiar interface on that provider). 2.) You don't have any of their actual password information. If there is a data breach on your site, they don't have to change a bunch of passwords, if they are reusing them (even though they shouldn't be) in multiple places.

The case for multiple services can be made in that each one includes a larger portion of the population. Some people are anti-google/facebook, or might not have a twitter account (hard to believe, I know). It stands to reason, then, that multiple options increases the rate (though, I can't speak to how much). There is also a UX case to be made against providing too many options and overloading the user, however.

I know this doesn't answer the need-for-numbers part of your question, but I hope it helps.