Offering insurance without getting in the way

At will bank I led design for commissioned products: insurance, mobile top-up and capitalization. Insurance isn't top of mind for anyone, and even less for our customers, people in classes C and D.

Our research showed that people decide to buy protection in moments of transition and new risk. Waiting for them to find the insurance hub wasn't going to work.

I designed contextual offers: small insurance offers that show up where the customer already is, and talk about what they just did. Moved money with Pix? Protect your card. Paid the bill? Your card can be protected too.

A few rules kept them out of the way:

• The offer comes after the task, on the success screen, never as a step in the middle of it.
• One line, one tap. It never covers the screen or blocks the main button.
• The words connect to the moment, not to "insurance" in general.
• Each spot ran on its own, with its own numbers, so we could switch off the ones that didn't work.

Every offer point was an experiment. A few that shaped the rest:

• Which product on the Home minicard. Card insurance got 28% more sign-ups than account insurance, so card insurance became the main offer.
• Which argument. The test ran right at carnival, when cards and phones get stolen a lot in the crowds. So one version said: protect your card for carnival. The other offered a monthly R$10,000 prize draw. The prize draw converted 240% better. The same pull showed up again when people built their own insurance.
• Which offer after a Pix. Swapping the old offers for card insurance with Disney+ lifted conversion by 49% after a regular Pix, and revenue by 222%. After a Pix on credit, conversion dropped 46%, so that one went back to review.

• Context beat traffic. After a Pix had fewer views than the Home, and sold more than three times as much.
• Busy screens didn't sell. Credit home and account home had millions of views and converted almost nothing. They became candidates for a propensity model, showing the offer only to people more likely to want it.
• Some spots were switched off. Card details sold almost nothing, so we turned those offers off.
• Selling isn't the whole story. Offers after paying the bill sold less but kept customers longer. After a Pix sold the most, but had more cancellations. We started checking whether an offer that proactive was leading to sign-ups people didn't really mean.

The research had shown that the word "insurance" itself pushed our customers away. The next test was to call it protection instead, starting with the offer in manage cards. More offer points were lined up for pockets and CDB.