Conversion rate is the metric most ecommerce teams optimise for, and it's the wrong north star on its own. It's possible to lift conversion rate and lose revenue — a discount banner or a simplified product range can both increase the percentage of visitors who buy while decreasing what your business actually earns.
A more honest single number
Revenue per visitor — total revenue divided by total sessions — captures conversion rate and average order value in one figure. It's harder to accidentally improve through a change that quietly damages the business, because it's measuring the outcome you actually care about: money, not a proxy for it.
How this changes what we test
Every test in our programmes is evaluated against revenue per visitor as the primary metric, with conversion rate and AOV reviewed alongside it to understand the mechanism behind a result. A test can win on conversion rate and still be rejected if it doesn't hold up on revenue per visitor.
It's a small shift in what gets measured, and it consistently protects clients from optimisations that would have looked good in a monthly report and cost them money in the bank account.