Redesigning the Intranet

Business problem
When building a new intranet from scratch, it's impossible to predict how it will grow and how it will be used. After several years of rapid growth, it's a good idea to reevaluate the design and make adjustments as necessary. How does one perform an intranet redesign?

Identifying a problem
When we first built the intranet, there was very little content and we were all new to the organization. Lacking any better alternative, we decided to use basic content categories from another successful health care intranet built in a prior "work life." This was OK for a starting point, but as this intranet grew, we began to worry that we had outgrown the original design.

It was our sense, from anecdotal feedback and casual observations, that:

• the basic design (page layout) was still OK, and
• search worked well; but
• it was hard to find items without search.

We needed proof and some direction.

Why bother?
Corporate intranets are very expensive in terms of staff time. According to one influential study by the Nielson Norman Group, if an organization the size of ours improves the "usability" of their intranet from "good" to "best," it will save nearly $5 million annually in staff time wasted trying to find information. So it is definitely worth investing some money and effort to make the intranet as efficient as possible.

Usability testing
Based on a recommendation from a former colleague, we engaged the Bentley College Design and Usability Center to conduct usability testing of our intranet. The process was very interesting.

The goals for the study were to:

• Assess how easy or difficult the site is to navigate
• Determine whether the information architecture and navigation are intuitive and efficient to use

In preparation for the usability testing, our first assignment was to identify three user "profiles" that best represented most of our workers. We decided on these:

• Clinical
• Administrative
• Manager

We selected four participants for each of the three profiles, looking for a mix of technology literacy and organizational roles within each profile:

• Clinical: physician (x2); clinical specialist; psychologist
• Administrative: ward clerk; admin coordinator; financial analyst; nutritionist
• Manager: ambulatory; clinical finance; HR; nursing

Finally, we selected a set of tasks for participants to perform during the testing, including about a dozen common tasks for everyone and five different tasks for each role:

• Common tasks (finding information about people, a specific policy, two reports, etc.)
• Clinical (finding a specific form, who is on call for oncology, two lists, and specific patient education items)
• Manager (finding supervisory assignments, HR requests, responsible manager for a department, IT requests, and a specific list)
• Administrative (finding maintenance requests, HR requests, conference room scheduling, IT requests, and a specific list)

Over a period of three days, each participant spent an hour with one of the Bentley staff, talking about their experiences with and perceptions of the intranet, and attempting to complete the tasks for their role. Everything was videotaped with simultaneous screen capture, allowing the Bentley staff to conduct detailed analysis of exactly what each person did and said while attempting to complete each task.

Summary of findings
The conclusions from the usability study included:

• Overall, site navigation was easy.
• Of particular success was the use of search.
• The most significant problem was the information architecture itself: e.g., which of the main "tabs" would contain the desired item. Some tabs contained objects ("Directories," Resources"), while other tabs were role-related ("Staff," "Managers," etc.).

New information architecture
Now that we had identified the problem, Bentley staff recommended a "card sort" exercise with users to identify the proper information architecture and top level "buckets" (tabs).

• Bentley staff selected 75 terms from across the full range of intranet content, with a definition for each.
• We identified thirty more participants, ten from each of the same three profiles used for usability testing.
• Each participant was offered a $20 Amazon.com gift certificate upon completion of the exercise.
• The exercise was conducted on-line, from anywhere the participant chose, and required about 30-40 minutes to complete.
• Participants sorted the 75 terms into groupings that they found logical and intuitive.

After analyzing the results of the card sorting exercise, Bentley staff recommended new groupings of the intranet content. Based on these recommendations, we selected seven new top-level tabs for intranet navigation, in descending order of frequency of use:

• Tools (things used to accomplish a task)
• Staff (everything related to working here)
• Directories (lists, charts, team sites, wikis, etc.)
• Reference (policies, procedures, general reference)
• Clinical (anything useful for someone working in clinical areas)
• Reports (reorganized into clearer groupings)
• Academic (anything related to teaching programs)

We also reorganized everything within the primary buckets, creating new categories that were more consistent with the groupings from the card sorting exercise. Although it was still a lot of work, our database design (Driving with Databases) made the reorganization much easier.

Although we have not yet done another usability test to validate any improvements, anecdotal feedback indicates that people are more satisfied with the intranet's navigation and better able to find what they need. Hopefully the next usability test will confirm that the redesign moved the intranet from a "good" to "best" design based on Nielson Norman Group criteria.

Lessons learned

• Usability testing is well worth the cost in time and money.
• Investing in the best possible intranet design is justified by improved staff productivity.
• Database-driven intranets make reorganizing content much easier.

Posted 3 April 2008