How negative keywords for local service ads improve lead quality.
Negative keywords can improve local service campaigns by preventing ads from serving on clearly irrelevant searches. Google’s current negative-keyword documentation explains that Search campaigns support negative broad, phrase, and exact match, and that negative keywords do not behave exactly like positive keywords or automatically cover every close variant.
The strategic risk runs both ways. Too few exclusions waste budget and pollute conversion data. Overbroad exclusions block legitimate customers. Every negative should be traceable to a business rule, observed query, or documented experiment.
Decision: keep the query, route it to a more specific campaign, test it with controls, or exclude it.
“Looks bad” is not a durable rule. Record the reason.
Use a shared decision taxonomy
A master query-control system starts with categories that can be applied consistently, then adds industry and campaign exceptions. The purpose is to make the business rule visible—not to accumulate an unreviewed list of words.
| Decision category | Question to answer | Default treatment |
|---|---|---|
| Service scope | Does the company perform this exact work? | Keep, route to a dedicated campaign, or remove based on the service matrix |
| Property and customer fit | Does the company serve this property type and customer relationship? | Separate when economics or intake differ; remove only under a firm rule |
| Territory | Can the company serve the location named or implied? | Use location controls first, then query controls where intent is unmistakable |
| Project stage | Is the search for repair, replacement, maintenance, inspection, or information? | Route commercial stages; evaluate research stages with outcome data |
| Customer lifecycle | Is this new-customer acquisition, existing-customer support, or another workflow? | Separate workflows rather than making a broad account-wide rule |
| Business economics | Can this intent become profitable work under current capacity and pricing? | Use qualified-lead and revenue evidence before changing reach |
A shared list should contain only rules that are true across every campaign receiving it. Put service-specific, territory-specific, or offer-specific controls closer to the campaign where the context is known.
Choose negative match type deliberately
| Negative type | General behavior in Search | Use case |
|---|---|---|
| Broad | Blocks searches containing all negative terms, even in a different order | Concepts that are consistently irrelevant when all words appear |
| Phrase | Blocks searches containing the negative terms in the same order, with possible extra words | Specific phrases that signal a known bad intent |
| Exact | Blocks the exact negative phrase without extra words | One observed query that should be removed with minimal reach impact |
Negative keywords require variant awareness. Review singular and plural forms, synonyms, and alternate phrasings where the underlying exclusion remains valid. Misspellings and casing have different treatment than semantic variants; use Google’s current rules rather than assuming positive-keyword behavior.
Before applying an account-wide phrase, search for legitimate queries that contain it. A plumber excluding “commercial” globally could erase profitable commercial work. A roofing company excluding “repair” could destroy its highest-intent emergency traffic if it also performs repairs.
Model the company’s service boundaries
Write a service-acceptance matrix with “perform,” “do not perform,” “case by case,” and “partner” states. Negative decisions should reference that matrix.
| Query pattern | Possible interpretation | Better action |
|---|---|---|
| Service + repair | Urgent service need | Keep when repair is offered; isolate when economics differ |
| Service + parts | DIY product search or a customer describing a broken component | Review actual terms before broad exclusion |
| Service + cost | Commercial research, not necessarily low quality | Usually keep and answer with a strong landing page |
| Service + jobs | Employment intent | Usually exclude in customer campaigns |
| Adjacent trade | Wrong provider or a bundled project | Exclude only if the business never performs or partners on it |
Apply the same logic by property and customer type. For example, “commercial,” “apartment,” “mobile home,” or “rental” should not become default negatives merely because their close rates differ. First decide whether the company serves them, whether they need separate campaigns, and whether intake can qualify them efficiently.
Use location controls and query exclusions together
Campaign location targeting controls who may be eligible based on platform location signals. Search queries can still contain place names that reveal different intent. Review both.
- Maintain an approved territory list by service.
- Separate physical serviceability from brand or office location.
- Review queries containing neighboring cities, states, and national terms.
- Exclude unmistakably unsupported place intent when observed.
- Do not paste every outside city into every campaign without assessing scale and maintenance.
- Track “out of area” as a CRM disqualification reason to validate the advertising rule.
Territory decisions should also connect to the landing page. A page that promises statewide service while operations accept only a few ZIP codes creates bad leads even if the keyword list is perfect.
Turn the search-terms report into a decision loop
Collect
Review search terms at a cadence matched to spend and query volume. Include campaign, ad group, keyword, match type, cost, conversions, and available lead outcomes.
Classify
Assign intent, service, customer type, territory, and business-fit labels. Separate obvious irrelevance from ambiguous or strategically useful terms.
Decide
Keep, add as a positive keyword, isolate into a new structure, change the landing page, tighten the offer, or exclude at the correct level.
Validate
Check affected queries before applying the negative, then monitor volume, qualified leads, and conversion mix after the change.
Waste reviewed = spend on classified non-fit search terms
Negative precision = exclusions with documented business rules ÷ total new exclusions
Query outcome coverage = search-term leads with qualification outcomes ÷ attributable search-term leads
Use CRM outcomes from the attribution system. Without qualification and revenue data, a search-term review tends to overreact to words rather than economics.
Govern negatives like production changes
- Name shared lists by purpose, not by the person who created them.
- Document owner, scope, reason, match type, source query, and date.
- Require review for account-level and high-impact terms.
- Maintain an exception log for campaigns that should not inherit a list.
- Audit deleted campaigns and inherited lists after restructures.
- Compare traffic and qualified-lead mix before and after major additions.
- Remove exclusions when the business adds a service or expands territory.
A negative-keyword list is a living model of what the business refuses, cannot serve, or should route elsewhere.
Build the starter list from business evidence
The safest starter list is derived from the company’s own service rules and observed search terms. Begin with a small, documented set; review affected traffic; and expand only when the reason remains clear.
Write the service matrix
For every promoted service, label the adjacent needs the company performs, does not perform, handles case by case, or refers to a partner.
Write the customer and property rules
Document residential, commercial, owner, tenant, property-manager, building-type, project-size, and scheduling constraints that truly affect acceptance.
Write the territory rules
Map serviceable places by service and capacity. Distinguish a temporary staffing limit from a permanent market boundary.
Review real queries
Classify observed search terms against those rules, connect tracked leads to CRM outcomes, and apply the narrowest control that solves the confirmed mismatch.
| Required record | Example evidence | Why it matters |
|---|---|---|
| Rule | Written service, customer, territory, or capacity decision | Prevents subjective edits |
| Source | Observed query, intake pattern, or operating policy | Shows why the control exists |
| Scope | Account, shared list, campaign, or ad group | Limits unintended reach loss |
| Review date | Scheduled check after business or campaign changes | Keeps the system current |
| Outcome | Traffic, qualified leads, bookings, and revenue after the change | Validates whether the rule improved economics |
Pair this system with the local service landing-page framework and the CPL economics guide. Query controls, page qualification, and downstream lead scoring should describe the same definition of a valuable customer.
Are irrelevant searches training the wrong outcome?
We’ll connect search terms, campaign controls, landing pages, lead quality, and sold revenue—then show which rules are safe and which could erase demand.
Request a growth assessment