Usability testing Medversation.com
Usability testing, eye-tracking,
UX strategy and design recommendations
Centocor Ortho Biotech (Centocor) developed Medversation.com
as a website sales tool to allow healthcare providers (HCPs) the ability to compare the benefits and risks of various biologic therapies. But the company had a problem – web site hits were way down, and worse, the web site wasn’t being adopted by HCPs. Centocor’s mission was clear – refine the site or scrap it altogether.
Centocor and k-nekt approached us to help develop a method to identify and recommend enhancements to Medversation.com to increase adoption and usage of the web site by physicians (Rheumatologists, Endocrinologists and Dermatologists), nurses (LPNs, RNs, ANPs), Centocor reps and, potentially, consumers.
Working with k-nekt, we kicked off the project with a mini-workshop bringing together marketing, brand managers, the web site’s development team and market research to: vet project goals; analyze the original Information Architecture (IA) and site design; identify participant segments, specialties and locations; diagram established user flows and page hits, brainstorm opportunities for growth and increased value; and identify success metrics for the testing.
Immediately following the meeting, we interviewed a large group of HCPs to understand their baseline interests, attitudes and opinions (IAOs) regarding content and usability of a web site of this nature. As an independent third-party, the user interviews provided us an opportunity to better understand the chaotic nature of the physicians office where HCPs search for therapeutic information. The HCP interviews also provided insights into the vernacular they use in their role and setting. The direct output of this investigation led to the creation of 9 archetypical users, or personas. The personas we developed segmented the vast number of HCPs into 9 user groups, each with their own unique set of needs and level of online experience. During the actual online testing of the website, these personas allowed us to effectively tailor usability testing to each group’s demands and navigation behaviors.
With users brought to life through the personas and project goals firmly in mind we worked with Centocor’s market research group to identify and screen participants. Due to the difficulty of recruiting highly specialized HCPs (and an expedited time frame – Christmas was right around the corner) we identified 4 geographic markets in which to conduct user testing, Philadelphia, Chicago, Atlanta and San Francisco.
We then developed a usability protocol outlining the needs-driven exercises, scenarios and role-playing we would use for each persona during testing. Once approved we then drafted all testing materials and data gathering devices to maintain a uniform standard for testing and ensure all elements were administered consistently across all markets and users.
Usability test identified how HCPs did in five areas of performance:
1. Learnability How easy was it for users to accomplish basic tasks the first time they encountered the design?
2. Efficiency Once users learned the design, how long did it take them to complete basic tasks?
3. Accuracy How many errors did users make, how severe were their errors, and how easily did they recover from the errors?
4. Recall How easily did users establish proficiency? How much did the user remember post-testing?
5. Emotional Response How did the person feel about the tasks completed? Was the person confident, stressed?
The test was administered in three parts: 1. a participant ‘pre-test’ to gather demographic information, qualify areas of expertise and baseline web usage of the web, etc.; 2. an online test documenting participants attempts to complete the protocol exercises on the live web site; and, 3. a post-test survey (administered immediately following the online portion of the test) to document the users experiences on the web site.
The user experience and usability data was documented in a number of ways. To determine the user’s emotional experience, each participant was videotaped using real-time picture-in-picture. Recording users during the test allowed us to match their emotional state with the successes or failures they were experiencing on-screen. Serving as moderator during testing also allowed us to probe deeper and query users about their experience and emotional state during testing, and of course all conversations were recorded.
Understanding the emotional state of users is important to good usability, but much like a physician, it is critical to differentiate between signs (what we see) and symptoms (what we hear). Hearing what the ‘patient’ has to say allows us to zero in on the problem, but actually seeing what the ‘patient’ is doing informs us as to what is working, and what is not. While a participant is saying they are doing one thing, in reality them may be actually be doing something quite different. To deliver actionable recommendations it is important to recognize the ‘symptoms’ from testing, but it is more important to identify the ‘signs’ of ineffective usability.
To unequivocally identify recurring and systemic problems with the usability of Medversation.com we augmented user videos with eye-tracking technology. For the first time ever, Centocor could now ‘see’ where and what their users were doing on the web site. The results of the eye-tracking were documented and presented to Centocor in two ways. The first visualization method, gaze plotting, was highly insightful and provided a second-by-second diagram of where participants looked and/or clicked while navigating the site. The second visualization technique, heat mapping, was utilized to identify those page elements which held the most interest for users.
The results of the eye-tracking technology were immensely informative and allowed Centocor and its web development partner to refine navigational problems and web site inefficiencies based on real-world actions and insights. One key finding from the eye-tracking was the different ways that physicians and nurses searched for information. Busy doctors completely ‘scanned’ a page and its content before making a decision to click on a link. Nurses on the other hand were amazingly methodical in their search and could quickly zero in on the information they were searching, and were not distracted by other page elements or content.
Upon the conclusion of the usability study we delivered a comprehensive presentation (PowerPoint deck and printed report) documenting our methodology, research, insights, usability protocol, as well as strategic recommendations and tactical refinements.
Below are a few of the ‘findings’ slides we delivered to the Centocor and Medversation.com teams.