The search terms report is where the money went

Every Google Ads account has two lists. The keywords you chose, and the queries that actually triggered your ads. On a broad match account those diverge fast, and almost nobody looks at the second one.

Where to find it

Google Ads, then Campaigns, then Insights and reports, then Search terms. Set the date range to the last 90 days.

Now change the sort. Sort by cost, descending.

Most advice tells you to hunt through the long tail. The long tail is cheap by definition. The money is in the twenty rows at the top, and the top of a local service account is usually where the surprises are.

What to look for

Read down the cost column and ask one question about each row. Could this person ever have become a customer?

People looking for work. Queries with "jobs", "salary", "hiring", "apprentice", "careers". Somebody searching for a job in your trade is expensive and will never book.

People planning to do it themselves. "How to", "diy", "kit", "tutorial", "step by step", "yourself". They are researching, and they are researching the thing you sell, which is exactly why the match engine sends them to you.

People shopping on price when you do not compete on price. "Cheap", "free", "discount", "lowest cost". Whether these are waste depends on your position. If you are the premium option in your market, they are waste.

The wrong geography. Town and state names outside your service area. A local business bidding broad without tight location settings buys national traffic slowly and continuously.

Competitor names. Sometimes deliberate and sometimes an accident of broad match. If it is deliberate, note that bidding on a competitor's trademarked name in ad copy is a separate question with a legal answer, and it is not the same as bidding on the keyword.

Adjacent services you do not offer. The classic waste in every trade account. You do bathrooms, the query was kitchens.

What to do about it

Add the ones that could never convert as negative keywords. Campaign level for the ones specific to that campaign, shared negative list for the ones that apply everywhere.

Build the shared list once. "Jobs", "salary", "hiring", "careers", "free", "diy", "how to", "training", "course", "wholesale", "supplier". That list applies to almost every local service account and takes about five minutes.

Then look at what is left at the top of the cost column. If the expensive queries are all genuinely relevant, your waste is not here, and that is worth knowing before someone charges you to look for it.

The check that tells you whether it is working

Note the total cost of the 90 day search terms report today. Add the negatives. Come back in 30 days and look at the same report.

You are not looking for lower spend. You are looking for the same spend appearing against queries that could plausibly become customers. If cost per booked job has not moved in 60 days, the waste was somewhere else.

Where this stops

Two limits worth naming.

Google no longer shows you every query. Terms below a volume threshold are withheld for privacy reasons, so the report is a large sample and not the whole truth. You cannot negative what you cannot see.

Negatives can be overdone. An aggressive list starves the campaign of the volume the bidding needs to learn anything, and an account with too few conversions per month optimizes worse, not better. Cut the queries that could never convert. Leave the ones that merely look untidy.

An afternoon covers it. Run it before you increase a budget, not after.

• Google Ads
• local