Create trust and reassurance for Compare the Market

Generated millions in revenue

Strategic qualitative study

product marketing Design and copywriting

Designed multivariate tests


Compare the market knew through analysing their purchasing funnel where users were churning out of the purchasing flow. So they asked us, “Why and how do customers made decisions about insurance, and what were the attitudes toward the purchasing process?” I hypothesized that by explaining the product offering more clearly, users would be reassured as they progressed through their journey to purchase, thus increasing sales and clickthrough rates for car and home insurance policies.

Compare the market wanted to work quickly, and they were keen for us to test our hypotheses with real users, live, on their site.

At the onset of the project I conducted stakeholder interviews and user interviews asking various questions about the overall experience of buying car or home insurance, and how customers felt generally about using comparison websites and comparative products to make big purchase decisions. The goal was to discover the main influences on decision-making behaviours generally. Then we had customers share how they found their insurance policy before purchasing. I shaped a hypothesis based on investigative research, and interpret that into a series of testable design solutions.


From findings to solutions


From the homepage to the end of the journey, we needed to reassure users enough along the way so that when it was time to make the decision to buy, they would feel better about doing so. By interpreting feedback from users, we knew just where to target: key places in the flow where users felt unsure, frustrated, uneasy, or overwhelmed.


“The Research showed where to target—key places in the flow where users felt unsure, frustrated, uneasy, or overwhelmed.”


Existing design to a design variant

I learned some of the limitations of multivariate testing in determining product strategy, as product hypotheses are not always fully encapsulated in a single variant. To craft a useful test, you need to limit the number of variants and really design a test so the team can be sure why it succeeds or fails. In this case the only feedback we got was that the variants we’d designed and deployed had almost all produced positive results, and that we had generated millions in revenue for the insurance comparison giant.

My only real regret is not somehow taking home a life-size meerkat of my own.

Meg Porter