Analytics tools show hundreds of figures and almost none of them will change a decision. Here are the six that will, and what to do about each.
1. Enquiries, and where they came from
Start here, because everything else is only interesting in relation to it.
Set up conversion tracking so that a submitted contact form, a completed booking or a click on your phone number is recorded as an event. Without this you are measuring attention rather than outcomes.
Then look at which channel produced them. Search, direct, social, referral, paid. It is common to find that a channel absorbing most of the marketing effort produces almost nothing, and one nobody thinks about produces most of the work.
2. Which pages produce enquiries
Not the pages with the most visits — the pages people were on before they converted.
These are usually not the homepage. They are a specific service page, a case study, or an article answering a narrow question. Once you know which, you know what to write more of and which pages deserve investment.
It is also common to discover that a page nobody had touched in three years is doing the heavy lifting.
3. What people searched for to find you
This lives in Search Console rather than Analytics, and it is the most directly actionable data available to a small business.
Look at queries where you appear in results but rank around positions eight to twenty. Those are pages that are nearly working. A modest improvement — a better title, a clearer answer, a section that addresses the question directly — often moves them onto the first page.
Also look for queries you are receiving impressions for and had never considered. Customers describe problems in their own words, and those words are frequently better than the ones in your marketing.
4. Where people leave
Exit rates by page, particularly within a journey that should continue.
People leaving from a blog article having read it is fine. People leaving from step two of a four-step enquiry form is a problem, and a specific one you can go and look at.
Pair this with the pages people arrive on. A page with high entries and high exits is either answering the question completely — which is good — or failing to offer an obvious next step, which is fixable.
5. Phone versus desktop, compared properly
Do not just look at the traffic split. Compare the conversion rate on each.
If sixty per cent of visits are from phones but eighty per cent of enquiries come from desktop, the phone experience is losing you business. That is a specific finding with a specific remedy, and it is invisible if you only look at totals.
The usual culprits are covered in mobile-first design: awkward forms, small tap targets, untappable phone numbers.
6. Speed, from real visitors
Search Console reports Core Web Vitals from actual visits rather than a laboratory test. That is the version that reflects your customers' phones on mobile data, rather than your laptop on office broadband.
Watch it quarterly. Sites drift slower as content accumulates, and the drift is gradual enough that nobody notices until it is significant. See Core Web Vitals explained.
If a number would not change what you do next, it is decoration. Most analytics dashboards are ninety per cent decoration.
The metrics that mislead
Total visits. Traffic can double while enquiries fall, if the new traffic is the wrong audience. Always paired with a conversion figure, never alone.
Bounce rate, as traditionally understood. A visitor who arrives, reads your opening hours and leaves satisfied is recorded identically to one who arrived and recoiled. It measures single-page sessions, not dissatisfaction.
Average time on page. Distorted by the tab somebody left open for three hours. Medians would be more useful and are rarely shown.
Social followers. Unrelated to revenue for the great majority of businesses, and easy to grow in ways that produce nothing.
A monthly routine that takes twenty minutes
- Enquiries this month against last month and the same month last year.
- Top five pages that led to enquiries — has anything changed?
- Search Console queries between positions eight and twenty — pick one to improve.
- Phone conversion rate against desktop.
- Any page whose traffic dropped sharply, which usually means a technical problem rather than a ranking one.
Twenty minutes a month beats an hour once a year, because it catches problems while they are still small — a broken form, a page accidentally deindexed, a redirect that stopped working.
Getting the setup right
Before any of this is trustworthy: exclude your own office traffic, make sure conversion events actually fire, verify Search Console, and check that analytics is present on every page rather than just the templates somebody remembered.
It is worth an hour to confirm, because decisions made on broken tracking are worse than decisions made on none.
Our consultancy team sets up measurement that answers business questions rather than filling a dashboard. Start a conversation if you have analytics installed and have never once used it to make a decision.
