What people actually search, and how to model it
Open almost any local SEO report and look at the tracked keywords. You will usually find the same word four times: plumber, plumber near me, best plumber, plumbers in Leeds. Four rows, four positions, four line items on the invoice — and one question being asked four ways.
Meanwhile nobody is tracking is Acme Plumbing open on Sunday, or Acme Plumbing reviews, or Acme vs Riverside Plumbing. Those are questions real customers type, they are closer to the money than anything on the list, and they are answered by completely different machinery.
This chapter gives you a way to sort local queries so that a keyword set covers the market instead of covering one word. It is also the chapter that stops your rank-tracking work and your AI-visibility work from being two unrelated projects, because — as you will see — they run on the same list.
The four intents
A person searching for a local business is at one of four points in a decision. The taxonomy below is ours, not Google's: Google publishes no local query-intent classification. The four buckets are chosen because each one is answered by a different surface and fixed by a different lever, which makes them useful rather than merely tidy.
Discovery — "I need this kind of thing near me, and I don't know who."
emergency dentist, coffee near me, tyre fitting hackney. Unbranded, category-first, usually short. This is the query the entire map-pack apparatus exists to answer, and it is where relevance, distance and prominence do their work (the three forces). It is also the only intent most people ever track.
Comparison — "I have candidates, and I want to pick one."
best sushi in shoreditch, cheapest emergency locksmith glasgow, Acme vs Riverside plumbing. Evaluative or superlative. The searcher has already accepted that several businesses can do the job; they are now looking for a reason to choose.
Comparison queries are answered less by a three-result pack and more by whatever can produce a judgement — review-heavy pages, listicles, and increasingly an AI answer that names two or three businesses and says why.
Trust — "I have a name; should I use them?"
Acme Plumbing reviews, is Acme Plumbing legit, Acme Plumbing complaints. Branded and subjective. You will almost always "rank" for these — it is your name. Ranking is not the point.
What matters is what the searcher finds when they get there: your rating, your recent reviews, your replies, and whatever third-party pages carry your name. This intent is invisible to a rank tracker and decisive at the moment of purchase.
Logistics — "I have decided; I need one fact."
Acme Plumbing opening hours, Acme Plumbing phone number, does Acme deliver, parking at Acme. Branded and objective. These are answered directly out of your profile's structured fields, frequently with no click to anything.
You do not win a logistics query by ranking. You win it by having the field filled in, correctly, and by it being the same everywhere (the business entity).
Two axes that cut across all four
Two properties cut across the taxonomy and determine what you can do about a query.
Branded or unbranded. For a branded query, appearing is free — you are the answer. The work is controlling the content of the answer. For an unbranded query, appearing is the whole contest.
Objective or subjective. An objective query has a fact answer that lives in a field you control: hours, phone, address, services, attributes. A subjective query is answered from other people's text — reviews, articles, directory pages — and no amount of profile editing produces it.
Put together:
| Intent | Typical form | Branded? | Objective? | Answered by | What moves it |
|---|---|---|---|---|---|
| Discovery | category near me | No | — | Map pack, local finder | Proximity, category, prominence |
| Comparison | best category in area | Usually not | No | Reviews, listicles, AI answers | Reputation, third-party coverage |
| Trust | brand reviews | Yes | No | Your profile + third-party pages | Reviews, ratings, replies |
| Logistics | brand hours | Yes | Yes | Profile fields, often zero-click | Profile completeness and accuracy |
Only two of the four intents are rank problems at all. The other two are content and profile problems a rank tracker will never show you, which is exactly why they get neglected.
The intent predicts the surface
Chapter 1 inventoried the surfaces a local answer can appear on (what local SEO actually is). Pair it with this chapter and you can usually predict the surface from the intent, before you search:
- Discovery pulls a map result. That is what the pack is for.
- Logistics pulls the profile panel itself, and often ends there.
- Trust pulls a mixture: your profile on one side, third-party pages on the other.
- Comparison is the one that has genuinely moved. It used to pull a listicle; it now frequently pulls a generated answer that names businesses directly.
Make the prediction explicitly: it tells you which instrument to point at each keyword. A logistics query in a rank tracker produces a permanent #1 that means nothing. A comparison query tracked only by rank misses the surface that actually answers it.