We used a User Centred Design approach to define the process by which we designed and built Single Sign On. The resulting methodology follwed three phases namely the Understanding phase, the Concept Development phase and the Design and Build phase.
During the Understanding phase, research was done into the usability and feature sets of similar services on competitor sites, such as MSN's Passport and Yahoo's ID system. This ascertained some of the requirements for our system, and pointers to good and bad practice.
Research was also conducted into the current user base using BBC sites and systems requiring authentication/registration. Combined with market research, interviews and strategic directives, we had enough data on which to base user Personas.
These Personas were 'walked through' the existing systems user interfaces to which SSO would be applied, to expose weaknesses and potential traps for the user.
Because we had two primary personas, the decision was taken to design two seperate interfaces: One for adults and one for children.
Paper prototypes were drafted, tested and iterated to form the foundations of the system. These were worked up from sketches drawn in facilitated team workgroups through to higher fidelity computer based versions. Personas were used heavily to determine the right implementation. Flow diagrams and use cases were completed to flush out other permutations of the system. At this point, the technical build and final user interface design could begin.
As the initial stages of the technical build commenced, the design team worked towards defining the brand stragegy. The result of this work was a set of brand attributes that was to permeate through language, interaction and graphic treatment. There was no logo as such, because our research showed that registration and authentication were already obstructing the pathway to features users wanted to use. Adding to the already complex visual brand hierarchies within the website would not lend any benefit to users. Instead a 'System Persona' was developed to help define the voice of the system.
User testing was carried out on the visually designed prototypes and tweaked before being integrated with the back end system. The final product was then integrated with each application that required it, one by one. Roll out continues now.