Why most CRO stops at the click — and what closing the loop actually means

Open almost any CRO case study and you'll find the same shape: a hypothesis, a variant, a conversion lift, a victory lap. What you'll rarely find is what happened after the click. Did anyone actually buy anything?

For a SaaS trial or an ecommerce cart, that omission is forgivable: the click and the sale sit close enough in time and place that a conversion-rate lift is a reasonable stand-in for revenue. For a car, a home, a loan or a degree, it isn't. The decision starts online. The purchase closes somewhere else (a showroom, a branch, a call with a counsellor) days or weeks later, through a completely different system that your analytics tool has never heard of. Most CRO stops measuring exactly where that gap opens up.

The proxy metric problem

Conversion rate, leads, form completions, add-to-cart: these are all proxies. They're useful because they're cheap to measure and they arrive fast. But a proxy is only as good as its correlation with the thing you actually care about, and for considered purchases, almost nobody checks that correlation. It's assumed, not measured. A form that's easier to fill in produces more form-fills almost by definition. Whether it produces more sales is a separate question entirely, and it's the one that usually gets skipped.

A "winning" test is just a hypothesis about revenue, until you check.

That's not a small distinction. We've seen changes that lifted online engagement while quietly attracting a worse-fit audience: more enquiries, but enquiries from people who were never going to buy. Online, the test looks like a win. At the sale, it's neutral at best. Nobody notices, because nobody was set up to look.

What closing the loop actually requires

• A way to match the same person across systems: a phone number, an enquiry ID, a reference the sales team can key against.
• An attribution window that matches how long the real decision takes: days for a loan, weeks for a car, longer again for a home.
• The discipline to check back on a schedule, since outcome data always lags online data, sometimes by a lot.

None of this is exotic. It's mostly a data-matching problem and a patience problem, not a technology problem. Most teams already have both halves of the data: a CRM or DMS on one side, web analytics on the other. The two have simply never been introduced.

What this looks like in practice

On our homepage we show an illustrative scorecard: a configurator CTA reframe that lifted online conversion 9.4%, with test-drives attributed back at 6.1%. The gap between those two numbers isn't noise. It's the whole story. Stop measuring at the click and you'd report a 9.4% win. The number that actually pays salaries was a third smaller, and it's worth knowing precisely rather than assuming it tracks.

The same scorecard shows a form that got shorter: completions up 5.2%, no lift at the sale. That's not a failure to report. That's the report. A closed loop tells you the truth even when the truth is unflattering, and that's exactly when it's most useful. It's the only way you find out a "quick win" wasn't one, before you've built three more features on top of it.

Start closer to the sale, not further from it

You don't need a perfect data warehouse to start. You need one journey, one matching key, and the willingness to look at a number you might not like. If you want a blunt read on where your own funnel currently stands, our free two-minute assessment will tell you, no email required.

The alternative is what most of the industry already does: optimise the part of the funnel you can see, and hope the part you can't is coming along for the ride. Sometimes it is. Increasingly, for considered purchases, it isn't. And the only way to know which is to actually look.