OpenID: The Web’s Most Successful Failure
First 37Signals announced it would drop support for OpenID. Then Microsoft’s Dare Obasanjo called OpenID a failure (along with XML and AtomPub). Former Facebooker Yishan Wong’s scathing (and sometimes wrong) rant calling OpenID a failure is one of the more popular answers on Quora.
But if OpenID is a failure, it’s one of the web’s most successful failures.
OpenID is available on more than 50,000 websites. There are over a billion OpenID enabled URLs on the web thanks to providers like Google, Yahoo and AOL. Yet, for most people, trying to log in to every website using OpenID remains a difficult task, which means that while thousands of websites support it, hardly anyone uses OpenID.
OpenID promised to solve two problems. First, it would offer an easy way to log in to any website without needing to create a new account. And, second, it would enable you to have a consistant identity across the entire web. This worked well with the limited audience of bloggers and tech-savvy users that were part of the original vision.
But then as the vision of OpenID grew to encompass, well, everything, it became bogged down in the details. Despite widespread support, there is no uniform user experience. Every site that supports OpenID does it slightly differently, which only further confuses the majority of people.
The main reason no one uses OpenID is because Facebook Connect does the same thing and does it better. Everyone knows what Facebook is and it’s much easier to understand that Facebook is handling your identity than some vague, unrecognized thing called OpenID. That’s why, despite the impressive sounding billion URLs and 50,000 sites supporting OpenID, it pales next to Facebook Connect. Facebook Connect has been around less than half the time of OpenID and yet it’s been adopted by some 250,000 websites, is available to the hundreds of millions of Facebook users and has the advantage of Facebook’s brand familiarity.
Facebook also added a key ingredient that helped drive other sites to adopt Facebook Connect — sharing user data. One of the reasons more sites support Facebook Connect is that they get a piece of the user pie.
Web publishers never warmed to OpenID since it allows a user to log in to a website and leave a comment on a story, a blog post or a photo while essentially remaining anonymous to the publisher. That anonymous aspect has made OpenID less attractive to publishers who want to collect more data about their readers or interact with them — whether that means following them on Twitter, connecting with them on Facebook or sending them e-mail.
The OpenID Connect proposal aims to solve this shortcoming by using OAuth to allow publishers to request more information from a user when they log in using OpenID. But so far there has been very little support for OpenID Connect. Facebook Connect is still far more popular.
However, not everyone wants to tie their website’s login structure to a single company like Facebook. If 37Signals is the poster child for OpenID failure, Stack Overflow is the poster child for its success. The popular programming Q&A site abandoned traditional username/password based accounts in favor of OpenID and declared the experience a resounding success.
Government sites are also looking to use OpenID rather than tie themselves to Facebook. And the Obama administration has announced plans for an Internet identity system that sounds a lot like OpenID, though the exact details have yet to be revealed.
Eventually OpenID will likely disappear from the web, not because it was a failure, but because identity will be managed in other ways. Mozilla is hard at work putting identity in the browser. It’s not hard to envision Firefox managing your OpenID credentials for you, just as it does today with your passwords. In that sense OpenID may end up like RSS (another tool routinely declared dead), invisibly powering features behind the scenes, essential, but unnoticed. Eventually online identity may even come full circle and move back into the real world — chips in your phone, tokens that generate random codes or biometric devices.
The legacy of OpenID may well be that it was ahead of its time, but that hardly makes it a failure.