Google quietly hands every business owner a list of the exact phrases people typed before their profile appeared. Most have never looked. Of the ones who have, most looked once, saw a list of words, and closed the tab.
That list is the most under-used asset in local marketing, and the reason is not laziness. It is that raw data is not a decision. A list of two hundred search terms tells you nothing until somebody splits it, trends it, and compares it against what your profile actually says.
Brand searches versus discovery searches
This one split changes how you read everything else.
Brand searches are people typing your name. They already knew about you; your profile just confirmed where you are and whether you are open. Discovery searches are people typing what they need - "emergency plumber", "gluten free bakery", "24 hour gym" - and finding you among the options.
Discovery is where growth lives. A profile that is ninety percent brand searches is serving existing customers well and winning almost no new ones. That is a completely different problem from a low rating, and it needs a completely different fix. Without the split, both look like "we get a decent number of views".
Rising and fading terms
Demand moves before revenue does. A term climbing month over month is a signal to write about it now, while the interest is building. A term fading is worth knowing before you plan a season around it.
This has to be compared per location. A phrase rising sharply in one neighbourhood can be completely flat three miles away, and a chain-wide average hides both.
The content gap: what you rank for but never mention
This is the report that actually changes behaviour.
Take the searches bringing views to your profile. Take the words present in your business description, your services list and your recent posts. The terms in the first set and not the second are your content gaps: real demand, arriving despite you rather than because of you.
Covering them is the cheapest visibility work available, because you already have evidence the phrase brings people. You are not guessing at keywords. You are writing down something customers have already demonstrated they search for, and which your profile is currently silent about.
A bakery discovers "gluten free" brings a steady stream of views while the word appears nowhere on the profile. A garage finds "pre purchase inspection" is climbing and is not in their services list. Neither needed a strategy session. They needed the gap surfaced.
How you compare with the businesses nearby
A 4.3 rating with 90 reviews is excellent in one neighbourhood and mid-table in another. Absolute numbers are meaningless without the local field, which is why generic advice like "aim for fifty reviews" is worth so little.
The useful version benchmarks your rating and review count against businesses in your category near each location, then converts the gap into something you can act on. Not "you are behind" but "the businesses ahead of you average 140 reviews and you have 60, which is eight review requests a week for six months".
That is the difference between a dashboard and a plan.
Where AI genuinely helps, and where it does not
AI in this field is oversold in one direction and underrated in another.
What it is good at
Drafting review replies. The reply structure that works is well understood - thank, acknowledge the specific thing, say what changes, move offline. That is a drafting job, and a model reading the actual review text does it in seconds. A person who would otherwise leave a review unanswered for three weeks now sends a good reply in one minute.
Drafting posts from your own data. Not "write a post about coffee" but a draft grounded in your real categories, services and the specific gap term you are trying to cover. If your shop is marked as a regulated business, drafts can be steered away from the claims that get posts rejected.
Turning a checklist into a plan. An action plan that says post twice a week about these three topics, ask for reviews at this rate, and fix these two profile fields first, is more useful than a score on its own.
What it should never do
Publish unattended. A system that posts replies to Google without human approval will eventually post something you would never have said, under your name, in public, permanently. The economics of that are terrible: you save two minutes a day and risk the review section your business runs on.
In GBP Manager, AI drafts. A person edits. A person sends. There is no setting that changes that, deliberately.
What this looks like assembled
Data on its own is a tab nobody opens. The loop that works is short:
- Read the monthly search terms per location, split brand from discovery.
- Look at the gaps - the terms bringing views that your profile never mentions.
- Draft the description change or the post that covers one, edit it, publish it.
- Check where you sit against businesses nearby, and set the review and posting targets that close the gap.
- Work the weakest-first list, and let next month's data tell you whether it moved.
All five steps live in one place in Search Insights and the Visibility Score, per location, for every profile you manage. The point is not the dashboard. The point is that the distance between "people search for this" and "our profile now says this" collapses from a project to a few minutes.