Measuring local SEO in GA4 and Search Console

The three tools, and what each one can actually tell you

Most local SEO reports fail in the same way: they show whatever the tool exports most easily rather than what would change a decision. It helps to be clear about which question each tool can genuinely answer.

Nothing joins them up cleanly, and any report implying otherwise is guessing. Read them as three partial views rather than one funnel.

The Search Console split that matters most

The single most useful thing you can do in a local report is separate branded from non-branded queries. Branded searches — your company name — mostly reflect people who already knew about you. Non-branded searches are the ones SEO is supposed to win.

Filter queries to exclude your business name and you get a much honest picture of whether the work is finding new customers. A report showing total impressions climbing while non-branded stays flat is describing your existing reputation, not your search performance.

Watch branded volume anyway, though, for a different reason: when someone gets recommended by an AI assistant, they very often search your name next. A rise in branded impressions with no other obvious cause is one of the few visible signals that GEO work is landing.

Setting GA4 up so the numbers mean something

Out of the box, GA4 will tell you almost nothing useful about enquiries. Three things are worth doing:

• Mark real enquiries as key events. Form submissions, clicks on your phone number, clicks on your email address. Everything else is context.
• Track tel: link clicks specifically. For most local businesses the phone is the main conversion and it is invisible unless you ask for it.
• Segment by landing page, not just source. Which page someone entered on tells you which content is doing the work.

Set a baseline before you change anything on the site. Without one you cannot tell improvement from ordinary variance, and at local-business volumes the variance is large.

The measurement gaps nobody puts in the report

Three blind spots are worth knowing about, because they make numbers look worse or better than reality:

• Embedded forms. If your enquiry form is a third-party iframe, the submission happens on someone else’s domain and your analytics cannot see it. You can measure people reaching the form, not completing it. Check the provider’s own dashboard for the actual count.
• Phone calls. A click on a tel: link is measurable; someone reading the number and dialling it is not. Call tracking closes this, at the cost of introducing a second phone number you must then keep consistent everywhere.
• Deferred analytics. Many fast sites delay loading analytics until the visitor interacts or a timer fires. If that timer is long, short visits are never counted at all — which drags pageviews down while pushing average engagement time up, because only engaged visitors survive to be measured. We hit exactly this on our own site: a ten-second delay added for performance was quietly discarding bounces until we found it. It is worth checking how your analytics loads before trusting a sudden trend.

None of these are reasons to give up on measurement. They are reasons to state what a number does and does not include, which is the difference between a report and a decoration.

A monthly report worth reading

For most local businesses, five things are enough:

• Non-branded impressions and clicks from Search Console, with the trend over months.
• Positions for the handful of searches that actually bring customers — not a list of two hundred keywords.
• Calls, direction requests and website clicks from the Business Profile.
• Enquiries from the site, taken from wherever they genuinely land.
• Whether the AI assistants name you for your main questions — recorded by asking them, since no tool reports this reliably yet.

Anything beyond that is usually there to make the report look substantial.