After reading both
Google's usability research on federated login and
Yahoo's OpenID usability research, it is evident that there is still no best
practice for a
login and sign-up box.
One of the most popular solutions to help users pick their
OpenID provider is to use
JanRain's OpenID Selector. Here is what it looks like:
1. User types
OpenID or uses a button to select their provider.

2. The following box is displayed when the user clicks on the
drop-down button. User selects their
provider and types their ID (if required).

3. User clicks sign-in button.
What is wrong with this approach?
- Most users don't know what OpenID is. Do we even need to explain what it is? Most people will already have an account with Google, Yahoo, Microsoft or one of the other providers.
- Having a URL input box is confusing. Why should I use a website address to login?
- I would rather not link to an external javascript file on another website.
Open SelectorI recently discovered
Open Selector, an alternative to
JanRain's ID selector.
1. User selects provider from a drop down list.


2. User types
username, URL or nothing at all depending on the provider selected.

3. User clicks sign-in button.
Why is this user experience better?
- User knows that they must select a provider first. Works well for users unfamiliar to OpenID.
- Only the OpenID URL input box is displayed if it is required.
The disadvantages?
- Users familiar with OpenID will need to select "Other OpenID provider" which is an unnecessary extra step.
- There is no "http://" when "Other OpenID provider" is selected. Do I need to add this or not?
How can we improve the usability?

- No need to even mention OpenID.
- Removing unnecessary words is always good usability. This text is already in the drop down.
- I don't mind giving attribution in the HTML code, but I would rather not clutter up the login box with this text.
- The OpenID input box has a yellow background. This should be white.
I'm still not convinced a
dropdown is the best solution, but I do feel this user experience is better for those that do not understand
OpenID - according to
this study, 78% of participants have not heard of OpenID.