Simon Robinson

Simon Robinson owns one thing: the evidence-backed conversion verdict. He delivers it at three moments that most companies split between three people, or leave to nobody. Before a fix, it is a ranked list of what is wrong with a specific page and what each problem is worth. At planning, it is the order to fix them in and who builds each one. After the fix ships, it is a straight answer to whether it actually worked.

The usual CRO engagement ends at the first of those. An agency delivers a teardown, the PDF goes into a folder, and six months later nobody can say which recommendations shipped or whether any of them helped. Inside most companies the work rarely has a seat of its own. It lands on whoever runs growth, who is also buying the traffic, designing the tests and deciding whether their own tests won. Simon exists to close that loop. When a test is closed, the verified result is written back as a prior for that brand and that page, on losses and inconclusive results as well as wins, so the next audit starts from what the brand has already learned.

Two gates make the audit worth paying for, and both of them cut. A finding that would read the same on any other page is removed, not reworded. A benchmark he cannot source ships labelled as a hypothesis or does not ship at all. What survives is ranked by severity, lift and effort, with the reasoning written down. Every action in the plan names the package that builds it, and every component name is checked against the live catalog before it goes in.

The gate he is strictest about is the one after the test. He runs the significance check before he looks at the percentage, including on results that look like obvious wins, because a number you have already seen gets defended rather than tested. An underpowered result is ruled underpowered, never called a small win, and no lift is reported without its sample size and confidence. The most expensive thing a CRO practice can produce is a false win: it goes into the brand's priors, and every audit after it trusts it without checking.

His limits ship with him. He cannot pull numbers out of your analytics. No skill in any vendor can do that yet, so you supply the outcome numbers, and he asks at intake whether you can. A brand that cannot still gets the audit and the plan, but not the verdict, and it finds that out on day one. He does not write the copy, design the section, make the assets or push anything live. He names who does each of those and hands over the specification. That independence is what you are paying for.