Information Architecture from Mental Models
In 2003, Knowledge Pathways, the software division of Global Knowledge needed a complete redesign of their corporate website. Given the opportunity, I quickly took lead of the design strategy, formed a team, and set up plan to rework the site's information architecture based on the mental models of the target users.
Step 1: Identified the Users
The team brought together representatives from every department that had contact with the customer base. We worked with the representatives to brainstorm the target users of the site and identified 11 distinct user types, or personas.
Step 2: Identified User Tasks
With budgetary and time constraints ruling out contextual inquiry with the target users, I instead went to contacts who interfaced with the users on a regular basis. I drafted questionnaires to illicit responses on tasks and goals and asked our contacts to respond to them from the users' point of view.
Step 3: Analyzed the Responses and Built the Mental Model
Once the responses were in, I extracted task, goal, and concern data elements and transferred them to cards. The team gathered and we determined common categories. We then culled the categories into mental spaces.
Step 4: Did a Gap Analysis
I did a gap analysis of the new model as compared to a content inventory of the existing site. There were clearly defined areas where content was missing in the existing site as well as content that was not important to the target users. The team agreed that the site needed to shift to the new model.
Step 5: Derived the Site Structure
With the new mental model as a foundation for site navigation, I worked with the team to validate content terminology and integrate content necessary for business requirements.
Step 6: Developed, Tested, and Launched the Site
With content established, I designed and developed the site. After testing, I launched it in time to meet the debut of new site marketing collateral at the 2003 PeopleSoft Conference. Global Knowledge representatives at the conference received very positive feedback.
