Stop redesigning the button before you talk to the user

A founder noticed visitors to their pricing page rarely clicked the call-to-action button, so they prepared to redesign it for better visibility (Sadaf M. from Learn Why We Trust). Before making changes, they watched people navigate the page and asked them to explain the offer. These users understood the button's function but expressed uncertainty about pricing, support costs, and cancellation terms. The founder shifted from a visual redesign to clarifying these specific details on the page, which allowed them to measure the impact on conversion rates directly.

This investigative approach mirrors the success of Steven Cramton and Zackary Stephen, two American amateurs who defeated a team led by grandmaster Vladimir Dobrov in a 2005 chess tournament (Growth Mindset). While the grandmasters relied on their experience, Cramton and Stephen organized their attention to use computer calculations for evaluating specific possibilities rather than letting the machine dictate their entire strategy. They won the competition and a $10,000 prize by using technology to support their own judgment instead of replacing it.

Xinran Ma observes that AI tools frequently pull designs toward the average because they are trained on common data, which fails to serve users with unique needs like a nurse checking a chart one-handed (Alex Pawlowski from The Strategy Stack). A specific point of view remains essential because generated work often lacks the context required to solve non-common problems. Investigating failure provides the necessary clues to build products that people actually value, as seen when the creators of Faved moved from failed ideas to a successful brand-deal platform after interviewing their users (Produced Newsletter).