Search-Term and Negative-Keyword Hygiene: Where Google Ads Budget Actually Leaks

A Google Ads campaign can report a perfectly acceptable cost per acquisition or cost per lead and still be leaking a meaningful share of its budget — not through an inefficient bid, but through traffic that was never the intended customer in the first place. Keyword-level reporting won't show this. It shows the keyword you targeted and how it performed in aggregate, not the actual words someone typed before the ad fired.

That's the gap this article works through: how to read search-term data, decide what to do with it, and use it for more than cleanup.

Keywords Are What You Target; Search Terms Are What People Actually Typed

A keyword is the targeting mechanism — the term or phrase you've told Google Ads makes an account eligible to show an ad. A search term is the actual query someone typed that caused the ad to appear, which is not always the same thing as the keyword that matched it. This is the entire reason search-term data matters: it's the only place in the account that shows real user intent rather than the intent you assumed when the keyword was added. A keyword performing well in aggregate can still be matching to a wide spread of search terms, some genuinely relevant and some not, and the keyword-level number is just the average of all of them.

Where Search-Term Waste Comes From

Mismatch tends to fall into a handful of recognisable categories — the following are hypothetical illustrations of the pattern, not a checklist to copy: someone searching with informational intent ("how does X work") when the campaign requires commercial intent; job or career searches triggering a keyword aimed at customers rather than jobseekers; free, cheap or DIY intent matching a keyword meant for a paid or premium offer; an unrelated meaning of a term entirely (a word with two meanings, only one of which fits the business); competitor or brand queries surfacing when that wasn't the intent; geographic mismatch, where the searcher is outside the area the business can actually serve; and audience or context mismatch, where the query technically relates to the category but not to who the campaign is meant to reach. None of these categories automatically means "add a negative" — they're where to start looking, not a verdict.

Not Every Bad-Looking Search Term Is a Negative

This is the part that separates careful account management from reflexive list-building. A search term that looks irrelevant at a glance still needs to be evaluated against its relevance, the intent behind it, the quality of what it converts into, its actual business value, how much frequency and spend it represents, and whether it might represent an opportunity rather than a problem. That evaluation leads to one of four decisions, not a single default:

• Keep — the term is relevant and doing what it should. No action needed.
• Optimise — the term is relevant, but performance around it needs work — the ad, the landing page or the bid, not the targeting itself.
• Add as keyword — the term reveals real, valuable intent that wasn't being deliberately targeted yet, and deserves its own keyword rather than staying an incidental match.
• Exclude — the term is clearly irrelevant or commercially unsuitable, and continuing to pay for it doesn't serve the account.

Treating every unexpected or unfamiliar-looking search term as an automatic negative is how accounts quietly exclude their way out of real demand.

The Difference Between Relevance and Profitability

Relevance and profitability are related but not the same axis, and a decision based on only one of them tends to be wrong. A term can be relevant but expensive, relevant but low-converting, relevant and genuinely profitable, or — less intuitively — irrelevant on its face but occasionally converting anyway. None of these get resolved by looking at click volume or CPC alone. They require the same layer every accurate optimisation decision requires: whether the conversions being counted are actually being tracked correctly, and whether the resulting customer represents real business value once CAC and downstream economics are accounted for, not just a cheap click. A search term that looks wasteful on cost per click can be entirely justified once it's evaluated against what the customer is actually worth.

Negative Keywords: What They Actually Do

Conceptually, a negative keyword prevents an ad from being shown for a query pattern that's been identified as unwanted — it's a control on which traffic the account pays for, not a targeting tool in the way a regular keyword is. Used well, it protects budget from being spent repeatedly on the same identified mismatch. The specific mechanics of how different match types apply to negatives, and how they interact with close variants, are platform behaviour that changes and is best verified directly in the platform rather than treated as fixed here — the concept that matters for this article is the control itself: stopping known-unwanted queries from consuming budget again.

Negative-Keyword Hygiene Is an Ongoing Process

Negative-keyword management isn't a list built once during setup and left alone. It's a recurring loop: search terms come in, get classified by intent, get sorted into waste or opportunity, get excluded or expanded into new keywords accordingly, get monitored for how that change performed, and the cycle repeats. This is the same discipline that runs through how performance marketing actually works as a full-funnel system — an account is never "done," because the traffic feeding it keeps shifting.

What Search-Term Reports Can Reveal Beyond Wasted Spend

Read the other way, search-term data is one of the more useful sources of insight an account produces — not just a cleanup task. It shows the actual language customers use, which often differs from the language a business assumes. It surfaces new keyword ideas that were never deliberately targeted. It reveals objections and questions embedded in how people search. It shows demand patterns and unexpected pockets of intent. And it can point directly at content or messaging gaps — search terms nobody is currently writing for. Treating the search-term report only as a waste-removal exercise misses most of what it's actually telling you.

The EdTech Example

The EdTech Growth System case study — an education and lead-generation account — documents search-term and negative-keyword work as part of its diagnosis and response. At handover, the diagnosis found that "spend was leaking through search terms and keywords that weren't earning it, plus locations and times of day that didn't deserve equal budget" on the Google side of the account. The documented Google Ads actions included "search-term and keyword-level analysis, negative-keyword optimisation, location include/exclude adjustments, bid optimisation, time-of-day optimisation, and holiday-specific plus automated budget rules."

The case study's own stated learning from this work is: "Search-term and location hygiene compounds. Trimming wasted search terms, negative keywords and inefficient locations brought cost per lead down without needing new campaigns." That improvement sits alongside parallel work on Meta, the website and creative during the same engagement — the case study doesn't isolate a cost-per-lead figure attributable to search-term work alone, and no specific search terms, negative-keyword lists, or recovered-spend figures are documented, so none are claimed here. What's established is the pattern: this kind of hygiene was real, named work, not a theoretical exercise.

A Practical Search-Term Review Framework

For any search term worth a decision, work through these in order:

1. Is the intent relevant? Does it reflect what the campaign is actually trying to reach.
2. Is the user commercially relevant? A jobseeker, researcher or browser is a different person than a buyer.
3. Does it match the offer? What's being searched for and what's being sold should line up.
4. Does it match the geography? Relevant to a location the business can actually serve.
5. Does it produce meaningful conversions? Not assumed — checked.
6. Is the conversion actually valuable? A conversion and a valuable customer aren't always the same thing.
7. Is there enough evidence to act? One or two clicks rarely tell you anything reliable.
8. Should it be kept, optimised, added as a keyword, or excluded? The decision the first seven questions were building toward.

When to Add a Negative Keyword

Exclusion should follow evidence and intent, not instinct. Reasonable triggers include a clearly irrelevant query with no plausible commercial fit, a pattern of unwanted intent recurring rather than a one-off, traffic that's commercially unsuitable for what's being sold, a consistent geographic mismatch, or a recurring low-value pattern visible across enough activity to trust. None of this depends on a fixed spend or conversion threshold — what matters is whether the pattern is real and whether the evidence behind it is solid, not whether it crosses some arbitrary number.

When NOT to Add a Negative Keyword

This matters just as much as knowing when to exclude. Don't exclude because a single click didn't convert, because CPC was high on one occasion, because a term merely looks different from the keyword that matched it, because the sample size is a handful of observations, or because the query could plausibly represent a valuable new intent that just hasn't been evaluated properly yet. Premature exclusions are how accounts quietly cut off demand they never actually got to test — the negative keyword list becomes a record of impatience rather than evidence.

Search-Term Hygiene + Target CPA

This connects directly to why a Target CPA is a ceiling, not a forecast: if the traffic entering a campaign is poor, or the conversion signal being fed to Smart Bidding is wrong, adjusting the Target CPA doesn't address either problem — it just changes how aggressively the algorithm chases a number built on bad inputs. Search-term hygiene and conversion tracking accuracy are different layers of the same system. A bid strategy can only be as good as the traffic and signal underneath it, and no amount of target tuning substitutes for getting those two layers right first.

A Search-Term Hygiene Checklist

• Review actual search terms, not just keyword-level performance.
• Classify each meaningful term by intent.
• Identify recurring waste patterns, not isolated clicks.
• Distinguish clearly irrelevant terms from ones that are merely underperforming.
• Look for new keyword opportunities inside the "waste" review, not just cuts.
• Add negatives that are justified by evidence, not instinct.
• Verify conversion quality behind the numbers, not just conversion count.
• Monitor what changed after each round of edits.
• Repeat — this is a recurring process, not a one-time pass.

Google Ads optimisation isn't only about changing bids. Often the larger opportunity sits one layer below that — in actually understanding what you're paying to appear for. Good search-term hygiene protects budget from quiet, recurring waste, and at the same time surfaces exactly what customers are searching for in their own words — which makes it one of the few optimisation tasks that pays back twice.