Why raw search volume hid 21 queries worth writing for.

Six weeks of Search Console data for this site surfaced 21 queries we had never targeted and never written a page for. Most of them are the kind of term a keyword tool reports as zero volume, no data, or filters out before you ever see it.

The cluster looks like this:

• automated call answering trade business
• ai phone agent for tradies
• automated invoice follow up tradies
• phone answering service for tradies
• business automation for tradies

Read them together and the intent is unmistakable. These are trade businesses with a specific and expensive operational problem — calls going unanswered while everyone is on the tools — actively looking for something that fixes it. That is a buyer, not a browser.

Read them one at a time through a volume column, and every one of them looks like noise.

Volume measures the wrong thing

Keyword tools are built on aggregated clickstream and search data, then rounded, bucketed and thresholded. Terms below a certain frequency are not reported as small. They are reported as nothing. The tool is not wrong; it is answering a question about popularity when the question you need answered is about intent and reachability.

Three kinds of opportunity fall through that gap.

Emerging categories. A term describing something that barely existed eighteen months ago has no volume history to report. "AI phone agent for tradies" is not a mature keyword. It is a category forming in real time. By the point it registers as meaningful volume, it will be contested by everyone who waited for the number.

Conversational phrasing. When someone puts a real problem to an AI assistant rather than a search box, the system decomposes it into a series of narrower questions. Each fragment is longer, more specific and individually far lower in frequency than the head term it descends from. Volume tools index the head term. The fragments are where the decision actually gets made.

Commercially narrow terms. Some queries have small audiences because the buying population is small. A term searched forty times a month by people who each represent a five-figure engagement is not a low-value keyword. It is a high-value keyword with an accurate audience size.

Why the smaller term is often the better target

Two mechanics work in your favour at the low-volume end.

Competition tracks volume. Agencies and in-house teams select targets from the same tools with the same threshold, so the terms that show a number attract the effort, and the terms that show nothing sit unclaimed. Of the 21 queries above, not one is held by either of the competitors we benchmark against.

Specificity tracks intent. "Business automation" is a research query. "Automated invoice follow up tradies" is somebody with a named problem who has already worked out roughly what the solution is called. The second query is worth more per impression, and it is cheaper to rank for.

The result is a category of target that is simultaneously higher-converting and less contested. Volume-first selection is structurally incapable of surfacing it.

How to find your own

Start with Search Console, not the keyword tool. Search Console reports what people actually typed to reach you, including terms with no reported volume anywhere else. Sort by impressions, filter to queries you have never deliberately targeted, and read what comes back. This is observed demand, not modelled demand.

Read the cluster, not the query. No single term in the example above justifies a page. Twenty-one of them pointing at one problem justifies a section. The unit of opportunity is the theme, and volume tools report at the term level, which is another reason the pattern stays invisible.

Mine your own conversations. The questions your sales team answers twice a week and the tickets your support inbox keeps receiving are the same fragments, arriving through a different channel. They are free, they are specific, and no competitor has access to them.

Check who holds it. Search the terms. If the results are generic listicles and directory pages rather than a competitor's dedicated page, the cluster is unclaimed. That is the signal worth acting on — not the number in the volume column.

What a serious operator should do

Keep the keyword tool. It remains the right instrument for sizing a market, understanding seasonality and sanity-checking a head term. Stop using it as the gate that decides what gets written.

The practical change is one of sequence. Observed demand first, from your own technical SEO and local search data and your own sales conversations. Modelled volume second, as a check rather than a filter. For businesses in trades and similar operationally specific verticals, the highest-intent demand is disproportionately the demand that no tool reports, because the vocabulary is new and the audiences are narrow.

The queries above are still unclaimed as we publish this. That is the point. By the time raw search volume confirms they matter, the window will have closed — and the businesses that got there first will have done it by reading their own data instead of a market average.