Keywords for Small Business: How to Find Local Searches You Can Actually Win

Here is an invented example to show the problem. Imagine a roofer in Warrington whose website is built around “roofing services”. Now suppose page one for that phrase is national directories and how-to guides, while “flat roof repair Warrington” shows a map of local firms. In that case the first phrase sounds bigger, but the second is the one more likely to bring calls.

That gap is what choosing keywords for small business is really about. You are not hunting for the biggest numbers. You are looking for searches where Google already wants to show a local business, in towns you actually serve, against competitors you can beat.

This guide walks through six steps: list your services, towns and modifiers; expand the list with free Google tools; check whether each results page is commercial; judge whether you can win; give each group of phrases one page; then pick and track your first targets.

Everything here uses free tools. You do not need a paid subscription to do good local keyword research.

Step 1: Build Your Services x Towns x Modifiers List

You can build a usable keyword list in about 20 minutes with a spreadsheet, before you open a single tool. The starting list should come from your business, not from software.

1. Column A: services, in your customers’ words. Read back through phone notes, enquiry forms, emails and reviews. Customers rarely say “boiler diagnostics”. They say “boiler won’t fire up”. The customer’s word often beats the trade word, because it is the one they type.
2. Column B: towns you actually serve. Only places you will travel to or take customers from. For a Widnes-based trade that might be Widnes, Runcorn, Warrington and St Helens. Leave out places you would turn work down in.
3. Column C: modifiers buyers add. Cost or price, emergency or 24 hour, repair versus replacement, quote, and credentials such as “Gas Safe”.

Start with 10 to 15 phrases. Here is how fast the combinations grow for a hypothetical Warrington plumber:

Three of each already gives you 27 three-part phrases, before counting two-part ones like “boiler repair Warrington”. You will not build a page for every combination. Step 5 shows how to group them so you avoid dozens of near-identical pages.

One note on “near me”. It is how Google reads the searcher’s location, not a phrase to stuff into your pages. Our guide to near me SEO covers how that works.

If you only do one thing today, list your top five money services in the exact words customers use on the phone.

Step 2: Expand the List With Free Google Tools

A town and service phrase that brings you real enquiries can still show as zero, or as a vague range, in a keyword tool. That is normal for local searches. So this step is about finding real wording first and treating volume figures as rough signals, not verdicts.

Autocomplete, People also ask and related searches. Type each phrase into Google and note what it suggests. Google’s help page says “autocomplete predictions reflect real searches that have been done on Google”, weighted by factors including “the location a query is coming from” (Google). The same page admits predictions “aren’t perfect”, so use them for phrasing, not proof of demand. Then check the People also ask box and the related searches at the foot of the page.

Google Keyword Planner. It is free, but you need a Google Ads account, and Google says “you must complete your account setup by entering your billing information” (Google Ads Help). Set the location to your area, because the data “is specific to the location and Search Network targeting that you’ve selected”. Average monthly searches cover “a keyword and its close variants”, and historical stats “are only shown for exact matches” (Google Ads Help).

With little or no ad spend, you may see ranges such as 100 to 1K instead of exact figures. Search Engine Journal reported this in 2016, and Google said then it had no specific spend thresholds to share, so treat it as possible rather than a fixed rule.

Google Search Console. If you already have a website, the Performance report’s Queries tab shows the searches your site already appears for, with impressions (Google). Look for phrases with impressions but no matching page, such as a joiner seeing “fitted wardrobes” plus a town name with nothing on the site about wardrobes. The tables “omit rare queries to protect user privacy” (Google), so small town searches can be missing.

Google Trends. Use it only to check a service is not fading or heavily seasonal, heat pump installs for example. It is too thin for most small-town terms.

ChatGPT is fine for brainstorming services and modifiers, but it has no search volume data. Paid tools such as Ahrefs, Semrush or Mangools add convenience. They are optional.

Step 3: Search Each Phrase and Check Whether the Results Are Commercial

Two phrases that sound almost identical can produce completely different page ones. Which one you build for decides whether the page can ever bring calls. So for every phrase on your list, answer one question: is this local search a commercial one?

Google’s own rater guidelines describe “Visit-in-person” searches, where people are “looking for a category of businesses”, and give [dentists] and [car repair] as examples, against searches like [small dog breeds]. They also say “user location is very important” for “searches related to nearby places”, and that “some queries could go either way” (Search Quality Rater Guidelines, September 2025). A commercial intent keyword, put simply, is one where the searcher wants to hire or buy, not learn.

Tools guess at this. The live results page tells you. To check:

1. Open a private browser window and search the phrase.
2. Note what appears from top to bottom: ads, a map pack, business websites, directories, guides, videos.
3. Match it against the table below.

Location affects what you see, so search as close to your customers as you can. The near me SEO guide explains how Google works out where a searcher is.

Try it with two plumbing phrases. Search “how to fix a leaking tap”, then “emergency plumber St Helens”, and compare the two pages. If the first shows guides and videos and the second shows ads, a map pack and local firms, you have your answer: one wants a guide, the other wants a business.

When there is a map pack, you have two places to win: the ordinary results and the map. Our guide on how to rank in Google Maps covers the map side.

Build for it if the results show local firms like yours. Skip it, or answer it with a guide, if page one is all guides and official information. Reading local search intent this way takes a minute per phrase and saves months on pages that can never rank.

Step 4: Judge Whether a Small Business Can Realistically Rank

You can do everything right on a page and still lose because page one is ten national brands. A commercial search is only worth chasing if page one includes businesses like yours. This step cuts your list down to keywords a small business can rank for.

Run each commercial phrase through three questions:

1. Who is on page one? If small local firm websites hold several spots, it is winnable. If national chains, large directories and comparison sites fill most of them, organic ranking will be hard.
2. How big is the area? “Roofer Liverpool” is a bigger market than “roofer Widnes”, and far more contested. A small firm usually does better making the smaller area its main target and earning the bigger one later.
3. Would ranking send enough work? First place for a phrase almost nobody searches is worth little. If a paid tool shows that the business already ranking gets very little traffic across its whole site, the prize is small. Weigh that against what a page costs you in writing, photos and links. Some towns justify a page; a neighbouring one might not.

Keyword difficulty scores in paid tools are optional. Even Ahrefs says “there’s no way a single number can capture the true difficulty of ranking” (Ahrefs on YouTube). Your own look at page one is a better guide for local terms.

Ranking in the map pack depends heavily on distance and the strength of your profile, which our guide to local SEO ranking factors covers. If you work from home or a yard without a shopfront, read our guide for service-area businesses.

Leave “best” and “cheap” phrases for later. “Best” results tend to favour firms with large review counts, and “cheap” searchers can be low-value work.

The verdict: if page one has no business your size, or ranking would send almost no one, move on to the next phrase.

Step 5: Group Phrases That Share a Results Page and Give Each Group One Page

Do “boiler repair Warrington” and “boiler engineer Warrington” need two pages? The results page answers that for you. The output of this step is a simple keyword map: each group of phrases assigned to one page.

Use the same-results test:

1. Search both phrases in a private window.
2. If most of the top results are the same, Google treats them as the same need. Group them on one page.
3. If different businesses or different types of page rank, split them.

Then match each group to a page type:

• Commercial service phrases go on a service page.
• Service plus town phrases, where the results show town-specific pages, go on a location page. Our guide to local landing pages covers how to build one properly.
• Questions and cost searches that show guides go in an FAQ or guide.
• Directory-heavy terms are better served by your profiles on those directories than by a page of your own.

Ahrefs suggests checking what the top-ranking pages target to decide whether a service needs its own page or fits a broader one. It also notes that most local businesses’ blog keywords have little local intent (Ahrefs).

Grouping also keeps you clear of a real risk. Google’s spam policy says “doorway abuse is when sites or pages are created to rank for specific, similar search queries”, and gives “pages targeted at specific regions or cities that funnel users to one page” as an example (Google Search Central). Swapping the town name across 30 copies of the same page is exactly what that describes.

Here is an illustrative map for a hypothetical North West plumber:

Build pages for groups, never for individual phrases.

Step 6: Pick Your First Targets and Check Them in Search Console

Your first target list should fit on one screen: short enough to act on, with a way to see whether it is working.

Pick roughly 5 to 10 groups that pass all three filters:

• A commercial results page (Step 3)
• Winnable for a business your size (Step 4)
• Profitable work you actually want more of

Order them with your main money service first and your nearest town first.

Once the pages are live, open Search Console’s Performance report, filter by page, and look at Queries. Impressions arriving for your grouped phrases are the first sign the page is being matched to the right searches. Repeat the Step 3 check every few months, because results pages change. Our guide to local SEO tracking covers a fuller set-up.

Expect this to take months, not weeks. Nobody can promise a ranking, and you should be wary of anyone who does.

Do it yourself if you have a few hours and a handful of services. Get help if you cover many towns or services and would rather have the mapping done for you: our keyword research service does exactly that, and you can get in touch for a quote.

FAQ

Can I do keyword research for free?

Yes. Most local businesses can do solid keyword research with free tools only: Google autocomplete, People also ask, related searches, Keyword Planner (which needs a Google Ads account with billing details), Search Console if you already have a site, and Google Trends. Paid tools add convenience but are not required. Step 2 above explains what each free tool tells you and where it falls short.

Should I target a keyword with zero search volume?

Often, yes. Tools undercount local phrasing, and Search Console leaves out rare queries for privacy, so a town and service phrase showing zero can still bring enquiries. If the results page is commercial and the work is valuable, it can be worth a page or a section of one. Check it against Steps 3 and 4 before deciding.

How many keywords should one page target?

One group of closely related phrases that bring up mostly the same results page. That might be two phrases or ten. Avoid both extremes: one page per phrase, which creates near-duplicate pages, and every keyword crammed onto your homepage. The same-results test in Step 5 shows how to build the groups.

Can I use ChatGPT to find keywords?

It is useful for brainstorming services, customer wording and modifiers you had not thought of. It has no search volume data, so treat its suggestions as ideas to test. Search each one in Google, check whether the results page shows local businesses (Step 3), and confirm the wording in autocomplete or Search Console before building anything.