Work/Concept testing
Card Dial
Ninety percent liked it. Forty percent believed it.
Card Dial had a rewards-optimizer concept ready to build. Before writing a line of code, they needed to know if people would trust it enough to use it.
The idea tested well. The trust did not.
Card Dial's plan was a rewards-optimizer app: enter your cards, and it tells you which one to use at checkout. Appealing on paper. Before committing engineering time to a manual-entry version, they wanted a real read on whether people would trust a recommendation built on data they typed in themselves.
We ran a synthetic panel of 1,000 respondents across thirteen segments. About nine in ten liked the concept, and two in three said they would try it for free or at low cost. But only around four in ten fully trusted the recommendations once they understood the model ran on manually entered card data. The gap was not privacy. It was confidence in the math.
We recommended narrowing the launch: an invite-only beta focused on high-frequency spend categories, preloaded card logic instead of manual entry, and a visible reason behind every recommendation, since the clearest driver of trust was showing the reasoning, not hiding it.
What the panel found
First reaction
What actually builds trust, out of 10
Start here
Have something like this
on your desk?
A short call is usually enough to know whether we are the right shop for it. If we are not, we will tell you who is.
Start a conversation ↗