Turning Website Data Into Decisions, Not Just Dashboards

Before adding anything to a dashboard, I ask what decision this number would actually change. If a metric moving up or down wouldn't change what the team does next, it's a vanity metric — worth glancing at occasionally, not worth building a report around.

This reframes reporting from "here's everything happening" to "here's what needs a decision this week," which is a much shorter, more useful list.

Across SEO, paid ads, and CRM data, the numbers that consistently drive real decisions for the accounts I run are narrower than most default dashboards suggest.

The highest-value analytics work usually isn't inside any single platform — it's connecting GA4, Google Ads, and CRM data so a lead can be traced back to the specific channel, campaign, and even keyword that generated it.

Without that connection, a channel can look expensive on cost-per-lead while actually producing the highest-value, highest-close-rate customers — a mistake that's easy to make when each platform is reviewed in isolation.

In Google Analytics 4, mark only the actions that matter to the business as key events — enquiries, calls, purchases — so reports focus on outcomes. Link GA4 to Google Ads, and make sure each conversion is counted by one system only, to avoid double counting across tools.

The routine that works: one page, five to eight numbers, each compared with last week and the same week last month, each with an owner and a decision. If a number isn't tied to a possible action, it moves to a monthly appendix.

The link between marketing data and business results is the lead record. Capture the source, campaign and landing page with every enquiry — from UTM parameters, click IDs or the form itself — and keep it on the lead as it moves through the CRM. Then 'which channel brings customers' becomes a simple report instead of a debate.

Dashboards built on messy data mislead confidently. Before building reports, make sure events are defined consistently, test conversions fire once, spam is filtered, and UTM parameters follow one naming convention. An hour spent fixing data quality saves weeks of arguing about numbers.

End every review with decisions, not observations: increase budget on the campaign with the lowest cost per qualified lead, fix the landing page with the highest bounce rate on mobile, update the service page losing impressions. If a meeting produces no decisions, the report is measuring the wrong things.

For most small and mid-sized businesses, the free stack goes a long way: GA4 for website behaviour, Search Console for organic search, the ad platforms' own reporting, and a CRM for lead outcomes. A simple dashboard tool can bring the key numbers into one view. Paid analytics tools earn their cost only when you've outgrown this and know exactly which question they'll answer.

Analytics involves personal data about visitors. Collect what you need, explain it in your privacy policy, respect consent requirements that apply to your visitors, and avoid sending personal details such as names or phone numbers into analytics tools. Good data practice builds trust — and keeps the business on the right side of data protection law.

Write down the three decisions you make most often about marketing — where to spend budget, which pages to improve, which leads to prioritise — and the one number that would change each decision. Build your first report around only those three numbers. Add more only when a new decision genuinely needs it.

Revisit the list every quarter, because as the business grows, the decisions — and therefore the numbers that matter — change too.