Most analytics is watched, not used
Traffic dashboards get checked weekly and change nothing, because the numbers on them don’t map to a decision anyone can make. The test for any metric is what you would do differently if it moved.
1. Enquiries, and where they came from
The count of real enquiries, attributed to the page and source that produced them. This is the number the site exists for, and everything else is a leading indicator of it.
2. Form starts versus completions
The gap between people who began and people who finished is the most actionable number on the site, because the causes — too many fields, a broken validation, a missing reassurance — are all fixable. It is where you lose them, quantified.
If a number moving wouldn’t change what you do next week, stop reporting it.
3. Which pages precede an enquiry
Not most-visited, but most-visited-by-people-who-converted. These are usually unglamorous pages — pricing, a specific service, the about page — and they deserve the investment that the homepage tends to absorb.
4. Real-user performance, at the slow end
Field measurements from actual devices rather than lab tests, watched at the slower end of the distribution — the visitors having the worst experience are the ones you are losing invisibly.
5. AI appearance rate
How often you appear for a fixed set of buying-intent prompts, and how you are described. Analytics cannot see this at all, which is why it needs its own deliberate measurement.
Review monthly, not daily
Weekly wobble on a marketing site is noise, and reacting to it produces churn rather than improvement. A monthly review against these five is enough to run the site well.