AB Testing

In this project, A/B testing was conducted to find what new hires wanted on a support page. A/B testing compares two versions of a web page to see which one performs better. The business goal is part of an onboarding initiative within the company. The product owners include Corporate Communications, Information Technology, and Human Resources. As owners, their goal is to push self-service within the workforce.

Strategy

The strategy is to place on-boarding documentation into one central area for employees and the new hires to access. The common denominator is new hires, all staff at one time or another was a new hire and resources are relevant, no matter how long a person has been with a company.

Research

Various research methods, including brainstorming techniques and feedback from surveys, required collaboration amongst departments and co-workers such as; myself as the UX Researcher/Interaction Designer, Business Analyst; IT Service Delivery Managers, and New Hires. Feedback gathered over a three-day span from various departments created documentation for qualitative research.

Employees in the company cafe and around the outer offices participated in the sessions. The feedback was candid and real. The observation was the main focus as they walked through the layout, looking for relevant content that mattered to them.

Sessions

The turnaround time for each iteration was one day. The session consisted of five employees, with each employee taking around five minutes to review the wireframes. The group tests were no more than twenty minutes long.

Feedback

Noted above, version “A” is what currently exists; version “B” is a revised version based off of feedback from surveys. Conversations revealed that version “A” was confusing to most, the employees could not relate to the defined groups on the page. Noted above, version “B” was attractive due to community support feed and the list of items to scan under “Help Center,” this approach aided with users short-term memory. The top section remained the same for familiarity factors.

After reviewing feedback from version “A” and “B”; additional iterations required more testing. Noted above, version “C” allows employees to contribute feedback to the Support Community. The functionality was present in version “B,” but the title changed to “Latest Conversations,” which invites employees to contribute ideas versus reading support posts. Being that people read from left to right, the feed moved to the left; this was intentional to make the feed more prominent for users.

Both columns expanded in width for legibility purposes. Self-service is encouraged by placing the “Contact Us” at the bottom of version”C.” Noted above, version “D” is promoting search on the top of the page; the section encourages discoverability. The middle section is limited to three groupings and recategorized. The bottom section displays favorite links; making resources readily available, with one click.

Noted above, to the left; the search remained prominent at the top, an element that stays consistent through the support section. Search is the first option for self-service. Users related to the three groupings below, restructuring of the middle part reduced cognitive load for users, it became the second choice, as their eyes move down the page.

“Popular Links” and “Latest Conversations” moved below the main groupings as the third option; creating engagement among co-workers. As the fourth option, the “Contact Us,” “Live Chat,” and “Report an Issue” are listed on the bottom of the page, making them less prominent.

The layout of the page utilizes a grid layout, which allows for an easy read amongst groupings. The interaction cost of this design, coming from the homepage, is one click. Noted above, to the right is the second page of the support site. The page focuses explicitly on new hire tasks, based on feedback from surveys.

Takeaway

The next steps recommended is to make a Proof of Concept; building the site with content and limited functionality. It would determine if the material is cohesive throughout and the interaction correlates with a user’s task. Usability tests with the POC would confirm the assumptions made above, are valid, and provide reassurance that you’ve met and exceeded user requirements.

Going through the steps of discoverability and research will reduce interaction cost, saving your company time and money.