What actually moves conversion rate (and what doesn't)

We pulled the results from forty client CRO experiments run over the past eighteen months to see what actually moved conversion rate versus what just felt like it should. Some of the results surprised even us.

What won more often than expected

Reducing form fields consistently outperformed redesigning forms. Adding specific, dated social proof (not generic star ratings) beat adding more of it. And simplifying navigation on landing pages — removing the header entirely on high-intent pages — won more often than any headline rewrite we tested.

What lost more often than expected

Urgency and scarcity messaging underperformed on considered-purchase products, even though it's a reliable lever for impulse categories. Sticky "add to cart" bars showed no significant lift outside of mobile e-commerce. And redesigning testimonials into video format didn't outperform well-written text quotes enough to justify the production cost in most cases.

What we stopped testing entirely

Button color. After a dozen tests across different accounts with no consistent signal either direction, we retired color-only tests from the roadmap completely — the effect size, if it exists at all, is too small to justify the testing slot compared to structural changes.

The pattern underneath it all

Every reliable win in our data reduced friction or added specific, verifiable trust — not persuasion tricks layered on top of an already-confusing page. If we had to compress eighteen months of testing into one rule, that would be it.