CUA were approaching re-development of their online lead generation systems to combat unknown usability issues and unexplained drop off rates. The executive had assumptions about how the design could be upgraded based on the variety of designs employed elsewhere in the health insurance market. My team at Symplicit were engaged to conduct research and design of the customer acquisitions experience. The objective was to increase conversion rates by metricizing CUA's user experience and evaluating competing acquisition flows to decide on the best path forward for new development.
Design recommendations defensible with usability metrics
Mean usability score of 76 with a 95% CI 63.5-87.5
Stakeholder design decision support
User and stakeholder research
Low fidelity design
CUA contained a lot of internal knowledge about health customers across stakeholder, call center and branch staff but no primary research to validate the accuracy of their assumptions. We identified in stakeholder interviews that customers were calling back after their policy purchase with a complaint that their first bill came out higher than they expected. Staff blamed the online flow, however there was no primary research to validate these assumptions.
We recruited sixteen participants that were screened through criteria such as recent policy purchase, device preference and target demographics. These participants were briefed and invited to CUA to attempt to retrieve a quote for health insurance using online apps. We tested CUA against BUPA's acquisition flow in alternating order to control bias because our clients were interested in validating stakeholder assumptions about competing patterns. User task time, task completion and system usability score were recorded as metrics for analysis.
The mean usability score for CUA was 76 with a 95% confidence interval of 63.5 to 87.5. This was a good score and at nine points higher than BUPA there was only a 20% likelihood that the difference between the two was due to chance. However, not all CUA users understood that their purchase price included a 4% discount, conditional on owning a CUA transaction account. To control bias, we did not reveal their mistake for discussion until after the usability score was captured. At this point, all participants told us that the inclusive discount was confusing.
The average time for completing the CUA flow was 4 minutes 27 seconds with a 95% confidence interval of between 2m52s and 6m55s, more than a minute longer than users took with BUPA. The briefer time spent on BUPA could be explained by frustration with encountering information overload during the comparison stage. Participants were more successful completing a comparative workflow using the CUA site. This workflow involved more frequent opening of multiple PDF summaries which may explain the slower task times.
Design in low fidelity
We conducted a competitor analysis to search for appropriate solutions to the identified usability issues. I sketched the best combination of approaches and tasked my visual designer with digitizing the wireframes in the lowest fidelity required to communicate patterns effectively. The wireframes would be a presentation of our research findings, the key was to clearly express the problem in a way that the development team could thoroughly understand and implement a best fit solution within the constraints of their technology.
CUA gained clarity from our research recommendations and were able to pitch an acquisitions flow to the executive that was research based and defensible by usability metrics. The report revealed a robust and challenging evaluation of the user experience. The most significant usability issue identified by users was the obscurity encountered throughout the acquisitions flow around pricing. The business’s bet on price obscurity as a customer acquisition strategy was in fact depressing the usability of the flow leading to user frustration and cognitive dissonance. The sketch recommendations gave users visibility and control in order to provide a more transparent and successful user flow.
Visual design and research assistance credit: Simran Assan