up vote 2 down vote favorite
share [g+] share [fb]

Been having issues with stackexchange creating weird usernames for me and grouping them with my account, it appears to be a known issue, and emailed the stackexchange team... which leads me to my questions, I'm looking for usability studies of user authentication methods and interfaces, do you know of any?

link|improve this question
98% accept rate
feedback

1 Answer

up vote 2 down vote accepted

I can tell you from our experience that OpenID is a giant impediment to signups. I don't have hard numbers to back that up, only anecdotal advice from

  • complaints about openid on message boards any time a tangential reference to stackoverflow.com was mentioned
  • customer's insistence when stackexchange 1.0 was first released that a username/password auth method was critical
  • personal experience

At the Business of Software 2010 conference, one of the attendees had done an A/B test to guage the adoption of OpenID vs. google, facebook, and username/password. The only statistic I remember is that of the 1500 data points for signups, zero were from OpenID.

link|improve this answer
@Michael_Pryor: Stackexchange's OpenID implementation appears twisted beyond need -- but, I've also never used now, or coded for it either. Main thing so far with me is that I can't delete my OpenID ... AND it's not secure, meaning everything I look at my profile my OpenID is listed and profile views are done over HTTP; maybe I'm missing something, but that's what I see. – blunders Oct 11 '10 at 20:09
@Michael_Pryor: Anyway, while adoption is part of usability, I'm looking more for real usability studies on user authentication methods and interfaces. For example, CAPTCHAs are an user authentication method for sorting humans from bots, and likely one of the most common CAPTCHA interfaces being reCAPTCHA's; which is owned by Google. captcha.net/images/recaptcha-example.gif – blunders Oct 11 '10 at 20:14
@Michael_Pryor: Also, truly wonder if on the A/B that brand name trust wasn't the largest factor in people not selecting OpenID. Since in thinking about it, OpenID sort of sounds unsecure, as in OpenWallet... :-) – blunders Oct 11 '10 at 20:19
@Michael_Pryor: Guess the only way to remove brand from the test would be not to list logos, etc. – blunders Oct 11 '10 at 20:20
I don't get it... Isn't allowing someone to sign into a site using their Google/Facebook account an implementation of OpenID? I really don't see the point of that A/B test. – Lèse majesté Oct 12 '10 at 20:43
show 1 more comment
feedback

Your Answer

 
or
required, but never shown
discard

Not the answer you're looking for? Browse other questions tagged or ask your own question.