Poking the Bear: How Can I Engineer UX to Create A Data Advantage?

Years ago, while running at a subprime lending operation, we used to offer customers the option to receive their credit card via standard mail or pay $25 for expedited delivery. It was meant as a simple revenue boost.

Then, when we looked at the default rates for both populations, we saw that customers who chose to pay extra to expedite their card delivery were significantly higher risk. The urgency to receive the card revealed something deeper about their financial situation or behavior, and it gave us a profitable, high-signal segmentation lever. Now we were getting paid to get valuable data!

We call this method “poke the bear”. By designing experiences that invite or allow customers to reveal more information about themselves, you can gather free, valuable risk segmentation data. You might already be collecting meaningful risk data without realizing it. Consider another example, this time from a digital lender: One of our clients offered customizable loan amounts and terms through their app. Borrowers who adjusted the default settings—choosing a lower amount or shorter duration—turned out to be lower risk than those who accepted the pre-filled defaults. Customization signaled thoughtfulness and planning—strongly correlated with the likelihood to repay.

In that case, the lenders’ application was already “poking the bear”, and they just needed to put that data to work. Take a look at your current user journey. Where do customers make decisions, and what can you learn from how they behave? You can, and should, go further: intentionally engineering “bear pokes” as part of your data and modeling strategy. (Of course, poking the bear should always be paired with a solid test-and-learn plan and a disciplined modeling process that ensures all variables add value; otherwise, you might just create noise.)

At Ensemblex, we help lenders uncover their valuable data and turn it into real improvements to their bottom line. Ready to get started?