From data to a page that converts

An insurance landing page pulled in traffic and converted almost none of it. The brief was the usual one, "make it convert," with no diagnosis attached, which is exactly where most redesigns go wrong.

Instead of redesigning on taste, the work started with behaviour. Heatmaps and scroll data showed attention scattered across competing elements, no clear hierarchy, and thin reassurance around a purchase that, for insurance, is really a question of trust. The single most important action on the page was getting lost.

From there it was deliberate. The page was rebuilt around one primary action, the copy was rewritten to lead with the visitor's problem rather than the product's features, and the trust signals that matter in a regulated purchase were added: proof, clarity on terms, and visible security cues. Proper event tracking went in at the same time, so the outcome would be measured rather than assumed.

Conversion improved by roughly 28% after launch, and because the changes were tracked, the lift was attributable rather than lucky. But the point of the case is the loop itself: the same person read the data, designed the fix, and measured whether it worked. Analysis that stops at a recommendation is half a job. The value is taking it all the way to the page, and back to the number.