Project: Recommendation Engine
Opportunity
The vast majority of travel insurance companies file their benefits as bundled packages because it’s quicker and easier to do so and that is how they have been successful selling their products in the past. However, this can lead to paying too much because the bundle contains benefits not relevant to your travel situation or not getting the right coverage. After all, it’s easy to assume that the bundled plan contains all of the important benefits.
At battleface, however, we filed each benefit separately meaning we could mix and match any combination of benefits to create the perfect package for each customer’s specific needs. This allowed customers to get better coverage at a lower price. But this also introduced a problem, the market wasn’t used to selecting from a large menu of benefit options. We needed a unique way to get the right people the right benefits as quickly and easily as possible.
Solution
As this was a direct-to-consumer e-commerce site, optimizing for every step in the conversion funnel was imperative to driving successful business results. Using a Drop Off Value calculation, I recognized that our Product Select portion of the funnel had the highest drop-off value, and therefore, the most opportunity for increased revenue if improved.
Example of Drop Off Value calculation for each step in the conversion funnel
The data indicated that a large number of customers were indicating that they were interested in customizing their plans, however, they would then just click around and leave the site as they seemingly didn’t know which benefits to choose.
We came up with a screen we called the Recommendation Engine that would allow us to collect some additional details about the customer’s trip that would allow us to recommend specific benefit packages.
Early wireframes for the Recommendation Engine experience
Our main concern was that if we introduced additional questions into the quote flow, then it may hurt conversion. We released the Recommendation Engine as an A/B test to 50% of our traffic over the course of a week.
Results
At the end of the A/B test period, we observed that the Recommendation Engine harmed conversion as well as the average order value. Because of this, we decided to turn it off moving forward and come up with a different solution to test.
I believe that product teams must be ruthlessly objective when analyzing impact results. In this case, the wise thing to do was to remove the Recommendation Engine. In the end, we were able to successfully test a solution that led to an increase of over $8,000 in profit per week.
The final solutionn involved taking customer's straight to the customization screen but including the ability to filter the benefits based on trip aspects
