Is That Search Demand Yours? A 3-Step Check

An interior design firm in Singapore calls itself an ID. So does half the trade. Type "id singapore" into a keyword tool and a large number comes back, and it looks like the biggest opportunity in the plan.

Most of that search demand is not yours. Of the 50 suggestions our keyword tool returned for that seed on 18 October 2026, Singapore time, 6 were about interior design. Twenty-nine were Google matching "id" to the word "is" — "is singapore water safe to drink", "what is singapore fried rice", "how is singapore airlines". Five were about a Volkswagen electric van, 5 about passport and pass photos, and 1 about a hair styler sold in Indonesia.

An agency that quotes you the headline number for that term is quoting you demand for noodles, tap water and a van.

The problem is not bad tools. It is a word with more than one owner.

Keyword tools do exactly what they are built to do: they take the letters you typed and return what people type that starts the same way. They have no idea what you sell. Autocomplete is a prediction engine, not a market research panel, and Google says so plainly: its help page notes that predictions "reflect real searches that have been done on Google" but that autocomplete "doesn't simply display the most common queries on a given topic" (How Google autocomplete predictions work).

So the tool is honest and the output is still wrong for you, which is the hardest kind of error to catch. A fabricated number looks suspicious. A real number measuring the wrong thing looks like evidence.

Singapore makes this worse than most markets, for three reasons. Our trades love abbreviations: ID for interior design, TCM for traditional Chinese medicine, PT for personal trainer, reno for renovation, aircon for air conditioning. Our place names are shared with other countries. And our search market is small enough that a term with a global second meaning will be dominated by the global meaning.

What we found when we looked, twice, today

Two seeds, run in the same session, both from industries that buy marketing.

"id singapore", 50 suggestions. Six were recognisably about interior design: "best id singapore", "best id firms singapore", "best renovation id singapore", "top id singapore", "good id singapore" and the Reddit version of the first. Twenty-nine were the engine reading "id" as "is", "how is" or "what is". Five named a Volkswagen ID model. Five wanted an identity-card photo booth, three of those with "near me" attached. One was a skin clinic's price list and one was a hair styler.

There is a detail in that pull worth more than the counts. To put a scale on the biggest row, our tool anchored it against a known query from its own data, and the query it chose was "volkswagen id buzz singapore price" — an electric van. The number an interior designer would have been shown was calibrated against a car. We publish none of those figures, because the real ones in our tool derive from the Search Console data of businesses we work for, and that is their data, not ours to hand out.

"tcm for back pain", 56 suggestions. Three were not Chinese medicine at all: "tcm replacement cost", "tcm module cost" and "tcm test review". TCM is also the abbreviation for a transmission control module in a car's gearbox, which is not the trade the TCM Practitioners Board registers. A third meaning turned up in the results rather than the suggestions: when we searched "tcm singapore price" the same morning, 1 of the 15 slots went to a market report on Singapore's transcutaneous monitoring systems.

Three meanings, one acronym, and a clinic owner reading a keyword report has no way to see which is which from the totals.

We are not the first to trip over the place-name version of this. In our study of 109 local pages across five Singapore regions, 31 of the 120 results returned for five Woodlands searches were about somewhere else called Woodlands, mostly in Texas. For "plumber woodlands", just two results were Singapore plumbers.

Why the results page is the better witness

When we searched "best tcm in singapore for back pain" in our own metasearch that morning, 8 of the 20 results were not about Singapore TCM in any sense: two Amazon heating pads, a PubMed abstract on lumbar discectomy, a Harvard back pain test, a news tag page, a Pastebin post and a PC Gamer article about the settings for a video game called Back 4 Blood.

Some of that is our tool's fault, and we would rather say so than use it as evidence. Our metasearch had most of its engines erroring in that window and one weak engine returned obvious noise. Google would not serve that page.

But the lesson survives the caveat, and in our view it is the most useful thing on this page. A search engine's results are a decision about what a phrase means. If the first page for your term is holding products, academic papers and another industry, the engine has already concluded the phrase is not about your trade. No amount of on-page work changes that conclusion. You are not competing for position; you are arguing about the definition of a word, and the engine has the last say.

The three-step check, before the term enters a plan or a quote

This takes about twenty minutes and it is the check we would want run on us.

1. Read the rows, not the total. Open the actual suggestion list and ask one question of every line: could a person who wants to buy what I sell have typed this? Count the yeses. In today's "id singapore" pull that count was 6 out of 50. The total was never the number.

2. Search the bare term and look at who holds the page. Not your rank, just the ten results. If they include a different industry, a product listing, an encyclopedia entry or another country, the engine has settled the meaning against you. Our guide to reading a trending keyword asks whether a spike is real; this step asks whether the words are yours in the first place.

3. Add one disambiguating word and pull again. For the designers, "interior design singapore" instead of "id singapore". For the clinics, "tcm clinic singapore" or the symptom: that clinic-seed pull returned six suggestions in the shape "best tcm in singapore for back pain", "for knee pain", "for cough". If the demand survives the narrower seed, it was always yours. If it mostly disappears, the first number was borrowed from somebody else's market.

Step three is the one people skip, because it always shrinks the plan.

What this costs you

It costs the pleasure of a big number. A plan built on the disambiguated term looks smaller, slower and less exciting than one built on the bare acronym, and if you are comparing two agency proposals, the honest one will usually show the lower figure. That is a genuine commercial disadvantage for the honest proposal, and it is worth naming.

It also costs some real demand. A handful of people genuinely do type the bare acronym looking for your trade, and the broad term is where some brand discovery happens. The answer is not to abandon it. It is to stop pricing it as if all of it were yours, and to stop making it the keyword your homepage title is built around.

The deeper trade-off is this: once you know the real size of the demand, you may find it is too small to justify the spend. That is a useful thing to learn before you sign, not after. We have set out how we read that judgment for small markets in few buyers, big orders: is SEO worth buying?, and if you have not started the business yet, checking whether anyone is searching comes before any of this.

What we would do this week

• Pick the three terms your current plan leans on hardest and run the three steps above on each. Write down the yes-count out of the total, so the honest figure is on paper.
• Rebuild your page titles around the disambiguated phrase, not the acronym. "Interior design", not "ID". The acronym can live in the body copy where your customers already use it.
• Keep a page for the ambiguous term only if the results page gives you room. If the first page belongs to another industry, spend the hours elsewhere. On whether a close-variant phrase needs its own page at all, see one question, nine phrasings.
• Ask whoever is quoting you to show the suggestion rows, not a total. If they cannot or will not, the number has not been read by a human.

A closing concession, since the examples here are our own tools misfiring. Our keyword tool produced the 44 irrelevant rows. Our metasearch produced the heating pads. We publish the counts anyway because the alternative is a report that looks cleaner than the thing it describes, and we would rather you distrusted the numbers for the right reasons.

One next step

Send us the two or three keywords your current plan or your current agency's proposal is built on. We will run the three steps on each and send you back the yes-count out of the total, plus what the results page says about who owns the phrase. That is free, through our contact page.

If the honest answer is that the demand is smaller than the retainer makes sense for, we will tell you that and you should not hire us. Our SEO, AEO and GEO programme starts from S$500 a month with a three-month minimum, and it is the wrong purchase for a market that cannot carry it.

Figures describe keyword and metasearch pulls run on 18 October 2026, Singapore time. No business, agency or person is named.

FAQ

How do I know if a keyword's search volume is really for my industry?

Read the suggestion list instead of the total, and count the rows a buyer of your service could plausibly have typed. Then search the bare term and look at who holds the first page. If the results are another industry or another country, the volume is not yours, however large it looks.

Is this a problem with keyword tools or with Google?

Neither is malfunctioning. Tools return what people type that starts with your letters, and Google's help page says autocomplete does not simply show the most common queries on a topic. The mismatch is between a string of characters and a market, and only a human who knows the trade can close it.

Should I still target my industry's abbreviation?

Usually in the body copy, rarely as the main target of a page title. Six of the 50 rows in our "id singapore" pull were about interior design, so the term is not worthless, just much smaller than it appears. Build the page on the full phrase and let the abbreviation appear naturally.

What about place names shared with other countries?

Same problem, same fix: add the country. In our five-region study, 31 of 120 results for five Woodlands searches were about a Woodlands elsewhere, mostly in Texas. Writing "Woodlands, Singapore" in the opening paragraph, the title and the address block costs nothing.

Does a smaller, honest keyword number mean SEO is not worth it?

Sometimes, and that is worth knowing before you sign. A small market can still be worth serving if each customer is valuable. What it cannot support is a plan priced against demand that belongs to a van, a gearbox part or a plate of noodles.