Negative keywords for lead generation: exclude the wrong intent

Write a service map with three groups: wanted, unsupported, and uncertain. A roofing company may want commercial leak repairs, decline retail material sales, and evaluate residential inquiries separately. That is a stronger foundation than assuming every word associated with shopping or research is waste. A person comparing prices can still be ready to buy.

Add obvious non-customer intent only when it fits your business: vacancies, training courses, wholesale supply, or do-it-yourself instructions may be irrelevant to a specific campaign. Keep a note explaining each category. When a sales team says a search produces poor leads, ask for examples of the actual requests. The cause may be an unclear ad rather than a term that should disappear entirely.

Google distinguishes negative broad, phrase, and exact matching for Search. In simplified terms, negative broad can block searches containing all the negative terms in any order; negative phrase requires the terms in the specified order; negative exact targets the specified search without additional words. Negative matching behaves differently from positive keyword matching, so review Google's examples before deploying a large list.

Use the narrowest exclusion that solves the observed problem. A broad negative for a service word can unintentionally remove a profitable combination. Inspect variations that may need separate handling rather than assuming one entry catches every related meaning. Keep campaign-specific exclusions separate from genuine account-wide exclusions so an appropriate search remains eligible in the right campaign.

Create a review sheet with the query, campaign, spend, available conversion data, proposed action, and reason. Use categories such as relevant, wrong service, job seeker, competitor support request, and insufficient evidence. Preserve an uncertain category. An unfamiliar phrase is not automatically irrelevant, and a small number of clicks without a sale is not enough to identify a persistent problem.

Link the review to call dispositions where attribution allows it. If search terms are not available at the individual lead level, be honest about that limitation and compare at a broader supported level. Look for recurring patterns across calls rather than treating one anecdote as a complete campaign diagnosis. Fix misleading copy when it invites the wrong expectation even from a relevant search.

Assign an owner to shared negative lists and record when rules change. Service expansion, new territories, or a seasonal offer can make an old exclusion inappropriate. Before applying a list to another campaign, review its rationale and look for conflicts with the receiving campaign's intended traffic. The convenience of shared lists does not remove the need for context.

After a meaningful update, monitor impressions, useful calls, and query mix alongside spend. Lower spend alone does not prove success; the list may have removed valuable demand. If volume drops unexpectedly, inspect the change log and reverse only the exclusions responsible. Keep a small set of known valuable search examples as a manual check against newly proposed rules.

Use this checklist before a list goes live and at each scheduled review. The output should be a readable decision record that explains what you are trying to prevent, where the rule applies, and how you will notice if it causes harm.