
The short answer
For a local business, free sources beat paid tools, because local search volumes are too small to measure reliably anyway. Write down how twenty recent callers described their problem, check that phrasing against autocomplete and Search Console, then build the page in their words.
Twenty calls, written down word for word
The best keyword source a local business has is the first sentence of the last twenty calls it took. Write them down the way they were said, not cleaned up.
"My AC is running but it is blowing warm air." "There is water coming through the ceiling under the bathroom." "My tooth broke and I do not have a dentist." No keyword tool returns those sentences. Tools return "HVAC repair services" and "emergency dental care," which are the phrases other agency websites already published, which is precisely why the tool knows them.
Those verbatim sentences become your headings and your opening lines. They match how people search now that most queries are typed as full sentences or spoken to a phone, and they read like a person instead of a category.
Autocomplete, People Also Ask, and the bottom of the page
Open an incognito window, set the search location to the town you care about, and start typing your service. The suggestions are real queries in aggregate, not guesses. Type the service and stop. Then type it followed by "cost," "near me," "at night," "emergency," and each letter of the alphabet if you have the patience for it.
Click a People Also Ask entry and the box expands with further questions related to the one you opened. Four or five clicks gives you a map of what people are unsure about, and a map of uncertainty is a content plan. The related searches at the foot of the results page do the same work in different phrasing.
Two cautions. Your own history and location skew the suggestions, which is what incognito mode and the location setting are for. And autocomplete proves a phrase exists somewhere, not that anybody in your county types it.
Search Console shows what you almost rank for
This is the strongest free source and hardly anyone uses it properly. In the Performance report, set the date range to the last three months, switch on average position, and look at queries sitting between roughly position six and twenty with real impressions and very few clicks.
Those are terms Google already associates with you and is nearly willing to rank. Moving a page from position eleven to position five is a far shorter job than trying to rank for something you appear nowhere for. Filter by page as well as by query, so you can see which existing page is the one worth strengthening rather than writing a new one.
The only cost is verifying the site, which takes about ten minutes. If somebody built your site and never set this up, that fact tells you something on its own. The same report is where any serious organic search work starts.
Keyword Planner without a spending account
Google's Keyword Planner lives inside a Google Ads account and you can reach it without running ads, although Google has steadily narrowed what it shows to accounts with no spend. Expect bucketed ranges rather than precise figures. Ten to a hundred. A hundred to a thousand.
Take those as order of magnitude and nothing finer. They are averaged across wide geographies and grouped across close variants, so the number attached to your term is often the total for a whole family of phrasings across an entire state.
What search volume is actually worth locally
Here is the honest limit, and it is the reason paid tools matter less than people assume for local work. A term showing ten to a hundred searches a month statewide might be four searches a month in your town. Four people with water coming through a ceiling is a good month.
So do not sort your list by volume. Sort it by job value and by how close the searcher is to paying somebody. A low-volume phrase that means an active emergency is worth many times a popular research phrase, and the popular phrase is exactly where every competitor is already spending.
Where free stops being enough
Three situations justify paying. Comparing your rankings against several named competitors across many towns. A product catalog large enough that you need to see gaps at scale rather than one page at a time. And any project where somebody has to defend a content investment to a partner or a lender with numbers attached to it.
Even then, a permanent subscription is rarely the answer. Buy one month, export everything, cancel. Keyword research does not go stale in a quarter, and the sentences from your phone log go stale even more slowly than that.
Turning the list into pages
Group everything into the handful of things people actually want, not into individual keywords. Most local businesses land somewhere between five and twelve real intents, and each one deserves a single page.
Put the customer's own phrasing in the heading and answer it in the first two sentences underneath. Do not build a page per keyword variant. Twelve pages about the same job with the words in a different order compete with each other, split whatever authority you have, and none of them wins.
Follow-up questions
Usually not on an ongoing basis. Your call log, Search Console and the search results themselves cover most of what a local site needs. Rent a tool for a month when you need competitor comparison at scale, then cancel it.
Treat them as order of magnitude only. They are grouped across close variants and averaged over wide areas, so a statewide range can mean a handful of searches in your town. Rank terms by intent and job value instead.
Fewer than most people expect. Five to twelve genuine intents with one page each beats fifty thin pages built around variants. If two terms would be answered by the same page, they are one target, not two.
Let’s build something.
Twenty minutes on the phone and you’ll know whether we’re the right fit. Either way you’ll leave with something useful.