Photos with pro teams and data-driven campaigns.

On productions with pro riders and teams such as Visma and Lidl-Trek, I’m in charge of photography. I also handle performance campaigns, landing pages and social media content.

Creatives used to be tested without a system. When I sorted an archive of hundreds of older files into a new taxonomy, it turned out that only a fraction of the 85 possible message combinations had been tested. We knew nothing about the rest, and there was no way to tell why a creative won.

So I built systematic creative testing. Every creative has its place in a matrix along three axes: angle (the message), perspective (the topic it speaks about) and the rider type a cyclist should identify with.

For evaluation I built an app in Lovable, plus an AI system for long-term testing of Meta campaigns. It goes through the matrix, suggests messaging and visual direction for unexplored cells based on collected data, and logs launched campaigns back into the matrix. When I need to know which influencer, format, message angle, visual or anything else works best right now or in a past period, I have the answer in a few clicks. The result is a living database of creative and messaging performance that shows where the budget makes sense.

Decisions about creatives are made differently now. New briefs don’t come from gut feeling but from which cell in the matrix is missing and what the data says.