A well-maintained negative keyword list can reduce irrelevant clicks, protect budget, and make search-term data easier to interpret. This guide provides a repeatable workflow for finding exclusions, estimating their likely impact, choosing the right scope, and reviewing decisions without accidentally blocking valuable demand.
Overview
Negative keywords tell an advertising platform which searches should not trigger an ad. They are useful when a query is clearly irrelevant, attracts the wrong type of customer, conflicts with the offer, or consumes budget without producing a meaningful business outcome. They are not simply a list of words to add whenever performance looks weak. Each exclusion changes the traffic your campaigns can receive, so it should be tied to search intent and supported by evidence where possible.
A practical negative keyword list usually contains several types of exclusions:
- Irrelevant topics: searches for products, services, or subjects unrelated to the offer.
- Audience mismatches: searches from people seeking a different location, customer type, use case, or business model.
- Unqualified intent: queries that indicate research, employment, education, repair, resale, or another intent your campaign does not serve.
- Internal conflicts: terms that belong in another campaign, product group, or landing-page path.
- Efficiency concerns: queries that have generated enough clicks or spend to justify review but have not produced an acceptable outcome.
The central question is not “Does this keyword look undesirable?” It is “What evidence shows that this search should be excluded, and at what level?” A term may be appropriate as an account-wide exclusion, a campaign-specific exclusion, or no exclusion at all.
For broader guidance on clustering and campaign structure, see the PPC keyword management guide. Negative keyword decisions work best when they are part of a wider PPC keyword strategy rather than an isolated cleanup task.
How to estimate
Use a simple estimate to prioritize which search terms deserve attention. The estimate does not predict future performance with certainty. It helps compare opportunities using the data already available.
1. Estimate potentially avoidable spend
For a candidate query or query group, calculate:
Potentially avoidable spend = clicks × average cost per click
If a group of irrelevant searches generated 42 clicks at an average cost per click of $2.10, the potentially avoidable spend is $88.20. Treat this as an estimate, not guaranteed savings. Adding a negative may reduce those clicks, but the budget could also shift to other searches rather than disappear from the account.
2. Estimate the opportunity cost
Compare the candidate group with the campaign’s relevant traffic:
Candidate share of spend = candidate spend ÷ campaign spend × 100
This shows whether the issue is material. A small number of irrelevant impressions may need no action, while a recurring query group consuming a substantial portion of a tightly constrained budget deserves faster review.
3. Check conversion and value signals
Before excluding a term, review conversions, conversion rate, cost per conversion, revenue or lead quality, and conversion delay. A query with no recorded conversion may still be early in its decision cycle, particularly when the account has limited volume or a long sales process. Conversely, a query can produce conversions but still be commercially unsuitable if those conversions are unqualified.
For a basic comparison, calculate:
Cost per conversion = spend ÷ conversions
If there are no conversions, record the spend and clicks without converting the result into an artificial benchmark. “No conversion yet” is different from “proven to be unprofitable.”
4. Assign an action
Classify every reviewed term into one of four actions:
- Exclude now: the intent is clearly irrelevant or conflicts with the offer.
- Monitor: the term is uncertain, has limited data, or could convert later.
- Move or refine: the query is relevant but belongs in another campaign, ad group, or landing-page path.
- Keep active: the query matches the intended audience and has acceptable evidence.
This action framework is more reliable than excluding every expensive query. It preserves useful learning while addressing genuine waste.
Inputs and assumptions
A dependable search term analysis workflow starts with consistent inputs. Export or review the search terms associated with the relevant date range, then capture at least the following fields:
- Search term and matched keyword
- Campaign and ad group
- Impressions, clicks, and click-through rate
- Cost and average cost per click
- Conversions and conversion value, where applicable
- Location, device, network, and other available segmentation
- Landing page and offer represented by the ad
- Decision, owner, date, and reason for the decision
Use a consistent review window, but do not assume one window fits every account. A short window may be appropriate for a high-volume campaign; a longer window may be necessary for low-volume or delayed-conversion campaigns. Keep the dates visible in your worksheet so decisions can be revisited when more data arrives.
Organize exclusions by scope:
- Account-level: use only for terms that are unsuitable across nearly every campaign, such as a consistently irrelevant topic or audience.
- Campaign-level: use when a term is inappropriate for one offer, geography, or objective but useful elsewhere.
- Ad-group or theme-level: use to separate closely related intent and reduce overlap between tightly organized groups.
Be especially careful with broad words that can appear in both relevant and irrelevant searches. Excluding a single common term may remove more valuable traffic than intended. Record the match type and scope for every exclusion, and test the logic against known valuable queries before publishing changes.
A useful exclusion library includes the term, meaning, scope, date added, evidence, owner, and review status. This prevents duplicate research and makes it easier to remove an exclusion when the offer, landing page, or target market changes.
Worked examples
Example 1: Clearly irrelevant intent
Suppose a campaign sells paid software subscriptions, but search terms repeatedly include “free download,” “crack,” and “torrent.” These terms describe an incompatible intent. Even if the account has only a modest number of clicks, the mismatch is clear. Add carefully scoped exclusions, document the reason as “offer and intent mismatch,” and check related variants in the search term report.
Example 2: Location mismatch
A local service campaign targets one service area, but searches from another region generate 18 clicks at $1.80 each. The estimated potentially avoidable spend is:
18 × $1.80 = $32.40
If the campaign cannot serve that region, a location-specific exclusion or targeting adjustment may be appropriate. Do not rely on a keyword exclusion alone if the underlying problem is campaign location configuration.
Example 3: Ambiguous research intent
A business selling implementation services receives searches containing “course,” “certification,” and “tutorial.” These words may indicate a noncommercial research need, but they can also appear in searches from prospects evaluating providers. Instead of automatically excluding every occurrence, segment the terms by the complete query, landing page, conversion behavior, and lead quality. If a recurring group has 60 clicks, an average cost per click of $3, and no qualified leads, the review priority is high because the group represents $180 in spend. The correct action could be an exclusion, a separate educational campaign, or revised messaging.
Example 4: Campaign conflict
A brand campaign receives searches that clearly belong to a product-specific campaign. The terms are relevant, so excluding them account-wide would be counterproductive. A campaign-level negative can route traffic toward the intended campaign, provided the matching structure and budgets support that decision. Review the account after implementation to confirm that impressions and conversions are appearing in the expected destination.
When exclusions change the traffic reaching an ad, also check message alignment on the landing page. The landing page message match checklist can help distinguish a keyword problem from an offer or page problem.
When to recalculate
Revisit your negative keyword list on a defined schedule and after meaningful changes. A recurring review is useful because search behavior, offers, budgets, landing pages, and campaign structures change over time.
Recalculate the estimates when:
- Budget, bids, or bidding strategy changes materially.
- A campaign launches a new product, service, location, or audience.
- Landing page messaging or eligibility requirements change.
- Search volume, cost per click, or conversion rates move enough to alter priorities.
- A new campaign creates possible overlap with existing campaigns.
- Conversion tracking, attribution, or lead-quality definitions change.
- An exclusion has been active long enough to evaluate its effect on qualified traffic.
After adding negatives, compare the next review period with the prior baseline. Look for changes in irrelevant spend, qualified clicks, conversion volume, cost per conversion, impression coverage, and search-term diversity. Do not judge success only by lower spend: an exclusion is beneficial when it improves the quality or economics of traffic without removing valuable demand.
Use this practical maintenance loop:
- Review search terms and group them by intent, not just by individual words.
- Calculate clicks, spend, and available conversion signals for each candidate group.
- Choose exclude, monitor, move, or keep.
- Apply the narrowest appropriate scope and document the reason.
- Check for accidental blocking of relevant queries.
- Revisit the outcome after enough new data has accumulated.
Keep the list as a maintained decision system rather than a one-time cleanup. Pair it with consistent campaign naming, search-term reviews, and clear intent clusters. For broader account organization, consult the PPC account structure guide; for performance and tracking context, see the cross-platform campaign tracking guide. This workflow gives you a practical way to control irrelevant demand while preserving the searches most likely to support the campaign’s actual objective.
