SEO keyword research for a local service business: services, suburbs and intent

Why search volume tools say so little about suburb searches

Keyword tools report averages that are rounded and grouped, which suits a national advertiser and hides most of what a local trade needs to see. A search like "switchboard upgrade Narangba" usually shows up as a zero, a dash or a tiny number, even when real people make it every few weeks.

Google's own Keyword Planner explains why. Its average monthly searches figure covers a keyword and its close variants, is averaged over 12 months by default, and is rounded. Google also says that keywords with very low search volumes aren't discoverable or forecastable in the planner. That's the description at the time of writing (October 2026), and Google changes the tool from time to time.

Put those three facts together and the problem is plain. A suburb term that gets a few searches in winter and none in summer averages out to almost nothing. A dozen ways of saying the same job collapse into one figure. And the rare wordings, which are often the most specific and the most likely to turn into a booked job, drop below the line where the planner reports anything at all.

Third-party tools build their own estimates on top of this kind of data, so their numbers won't always match Keyword Planner or each other. None of that makes the tools useless. They're good for comparing service terms against each other and for spotting wording you hadn't thought of. They're poor at telling a trade whether one suburb deserves its own page. The table below reads the common outputs for a suburb-level search.

Start from your jobs: building the service and suburb list

The keyword list for a local business starts in its own records, not in a tool. Your invoices already say which services people pay for and where they live, and your phone calls say how they describe the problem before they know the trade term for it.

Pull the last 12 months of jobs and group them two ways: by service, using the words a customer would use, and by suburb. You may find the work clusters in fewer suburbs than you advertise, and that a few services bring in most of the revenue. Those clusters are where keyword research earns its time. Then widen the list with the sources below, in roughly this order, because each one checks the one before it.

• Job records and invoices: every service and suburb pair you've been paid for in the last year, with a rough count of jobs. This is the only source that tells you a term is worth money.
• Call and enquiry notes: the exact phrases people used, such as "half the house has no power" or "the safety switch keeps tripping". These become question headings, FAQ wording and guide topics.
• Search Console queries: the searches your site already appears for, filtered by page. Our Google Search Console guide shows where to find them and how to export the list.
• Google autocomplete: type each service plus a suburb into Google and note the predictions. Google says autocomplete predictions reflect real searches that have been done on Google, and that the location a query comes from and trending interest both shape them, so run the test from the area you serve.
• People Also Ask boxes: the questions Google shows under a service search. They feed informational guides and FAQ blocks, not suburb pages.
• Competitor service pages: only for the names they give services. A competitor calling it "switchboard replacement" while you say "upgrade" is worth knowing. Their copy isn't.
• Keyword tools last: to compare service terms against each other and to catch wording none of the above produced.

Sort every term by intent before you count anything

Intent is what the person wants when they search, and for a local trade it decides which page should answer, so sort by intent before you look at a single volume figure. Two searches with the same service and suburb can want completely different things.

Four intents cover almost every search a service business cares about. Emergency searches want someone today and will call the first business that looks able to come. Quote searches want a price range, availability and some reason to trust you before they ask. Informational searches want an explanation and may never hire anyone. Brand searches already know your name and want your number, hours or reviews.

The wordings in the table are illustrations for a hypothetical electrician, not measured searches. What matters is the last column: each intent points to a different kind of page, and a page that tries to serve all four usually serves none of them well.

Why a "zero volume" suburb term can still bring calls

A suburb term with no reported volume can still bring calls because the tools can't see small, scattered demand, and because one booked job is worth more to a trade than hundreds of visits from people who'll never hire. Zero in a tool means "too small to report", not "nobody searches this".

Search Console has a blind spot of its own here. Google's Search Central team explains that anonymised queries, those not issued by more than a few dozen users over two to three months, are left out of the query tables to protect privacy, though they still count in the chart totals. Suburb searches often sit below that line. So the query list can look thin even when the page is being found and the phone is ringing.

Location matters too. Google's Business Profile help says distance from the searcher is one of the main inputs for local results, and that when a person doesn't share a location, Google relies on what it knows about where they are. Someone in Narangba who searches "electrician" with no suburb at all can still be shown local businesses. That's why a service page with clear service-area wording supports suburb demand the tools never count.

None of this is a reason to build a page for every suburb on the map. It's a reason to judge suburb terms on evidence a tool can't see:

• You've done paid work there in the last year, and more than once.
• Enquiries name the suburb, or a landmark or estate inside it.
• Search Console shows the suburb name in queries for your service page, even at low impression counts.
• Autocomplete offers the service plus suburb wording when you search from the area.
• You can say something true about the work there that wouldn't be true one suburb over: older housing stock, a new estate, access, travel time from your depot.

Tool volume against Search Console: what 214 keywords on our own site showed

On our own site, Search Console impressions came to about 3% of keyword-tool volume at the median, so the two numbers measure different things and shouldn't be read as the same thing. Tool volume estimates how often a phrase is searched. Impressions count how often your page was shown for it.

The method, so you can repeat it. In October 2026 we matched DataForSEO's Australian monthly volume estimates against 90 days of Search Console data for smart-local.com.au, covering 27 July to 24 October 2026. 214 keywords had both a tool volume and at least one impression. For each one we divided the 90-day impressions by three to get a monthly figure, then divided that by the tool's monthly volume. Google defines an impression as a time a user has seen, or potentially seen, a link to your site in Search, Discover or News, so impressions only start once your page is in the results at all.

The spread matters more than the median. A quarter of the keywords drew about 1% of tool volume or less, while a tenth drew three quarters or more, and 16 drew more impressions than the tool said existed. Even among the 38 keywords where our average position was 10 or better, the median share was still about 3%. The disagreement runs both ways, so neither number corrects the other.

What the comparison can and can't tell you

The comparison shows that tool volume is a poor stand-in for the visibility a page will get, and it can't tell you what the ratio will be on your own site. Read the limits before you borrow the 3% figure for anything.

Here's how we use the two numbers now. Tool volume ranks ideas against each other: if one service term carries ten times the volume of another, it probably has more demand, even if both figures are rough. Search Console tells you what's actually happening for pages that exist. When they disagree about a term you already rank for, trust Search Console for that page and keep the tool figure for planning new ones. The limits:

• It's one site, and that site is a marketing agency, not a trade. A plumber's or electrician's ratios could look quite different.
• The median average position was about 32, so many of these keywords were seen rarely because the page sat deep in the results. Low impressions there reflect ranking as much as demand.
• Search Console's query rows leave out anonymised queries, so low-volume wording is under-counted on that side.
• The Search Console pull wasn't split by country, while the tool figures are Australian estimates.
• Ninety days of impressions were compared with a tool figure that averages a whole year, so seasonal terms can swing either way.
• DataForSEO's volume is itself a third-party estimate, not a count from Google.
• 214 keywords is a small sample, weighted towards marketing terms our site happens to appear for.

Worked example and the keyword-to-page map

The keyword-to-page map is a sheet with one row per query group and exactly one page that owns it, and it's the step that stops two of your pages competing for the same search. The worked example shows how one hypothetical business gets from a grid of combinations to a short list of pages.

Take a hypothetical electrician based in Morayfield, north of Brisbane. Six services: emergency call-outs, switchboard upgrades, smoke alarms, EV charger installation, ceiling fans and lighting. Twelve suburbs: Morayfield, Caboolture, Burpengary, Narangba, Deception Bay, North Lakes, Mango Hill, Kallangur, Petrie, Bellmere, Elimbah and Beerburrum. That's 72 service and suburb combinations, and a keyword tool would report nothing useful for most of them.

Run the records first. Say the job sheets show switchboard and EV charger work clustered in the newer estates around North Lakes and Mango Hill, emergency jobs spread evenly, and smoke alarm jobs mostly for rental property managers across the whole area. Then the map writes itself. Six service pages own the service terms. Two suburb pages earn their place, each with things to say about the housing there. Everything else is handled by the service pages' area wording, a Business Profile service area and a couple of guides. Our guide to service area pages covers the test a suburb has to pass, and suburb SEO explains how we build the ones that do.

Google's spam policies name pages targeted at specific regions or cities that funnel users to one page as doorway abuse, and that's what 72 near-identical pages would be. A map with fewer, stronger pages is easier to keep honest. If two pages already chase the same query, our guide on when to merge overlapping pages shows how to pick the keeper, and prioritising your next pages scores what to build next. Build the map in this order:

1. Group wordings that want the same thing into one query group, such as "switchboard upgrade", "switchboard replacement" and "new switchboard".
2. Give each group one owner page and write that URL in the sheet.
3. List the pages that should link to the owner, using wording close to the query.
4. Check Search Console every quarter: if a different page keeps getting the impressions for a group, change the owner or merge the pages.
5. Add a new row only when the records, enquiries or Search Console give you a new group, not because a tool suggested a variation.