The Google Ads search terms report shows the real queries that triggered your ads, not only the keywords you selected. A weekly review turns that data into three practical decisions: which searches to exclude, which themes deserve new Google Ads keywords, and which queries need closer measurement before you change the campaign. This workflow provides a repeatable classification method, cost estimates, escalation rules, and a review log you can use as query volume changes.
Overview
Search term analysis is one of the most useful forms of PPC query mining because it connects account structure with actual user language. A keyword describes the targeting instruction in your account; a search term is the query a person entered that matched that instruction. The two may be related without being identical.
Open the relevant campaign or ad group in Google Ads, locate the search terms report, and set a date range that contains enough activity to interpret. Review clicks, cost, conversions, conversion value where available, and the search term itself. Do not judge a query on wording alone. A term that looks broad may produce qualified leads, while a term that appears commercially relevant may spend without producing a useful outcome.
Classify every meaningful query into one of five groups:
- Keep and monitor: Relevant to the offer, but with insufficient evidence for a structural change.
- Promote: Relevant and promising enough to add as a deliberately managed Google Ads keyword or to build into a tighter ad group.
- Exclude: Clearly irrelevant, unsuitable, or outside the product, service, location, or customer definition.
- Investigate: Ambiguous wording, unusual intent, or a conversion that needs validation in analytics or the CRM.
- Separate: A valuable theme that should have its own budget, ad copy, landing page, or audience treatment.
This classification prevents a common error: adding every converting query as a keyword or blocking every query that has not converted yet.
How to estimate
Use a simple decision model before making an account change. For each query, record:
- Relevance: Does the query describe the product, use case, and customer you want?
- Intent: Is the person researching, comparing, looking for instructions, or trying to buy?
- Efficiency: How much has the query spent relative to your acceptable cost per conversion?
- Evidence: Does the conversion represent a meaningful business outcome, or only a low-value interaction?
- Control: Would a dedicated keyword, ad, landing page, or budget improve your ability to manage it?
A useful estimate is expected conversion cost:
Estimated cost per conversion = query cost ÷ query conversions
For queries with no conversions, use a conditional estimate rather than treating the result as proof of failure. For example, if a query has received 12 clicks at an average cost of $3, its current cost is $36. If your working conversion-rate assumption is 5%, the estimated conversion count is 0.6 and the implied cost per conversion is $60. This is a planning estimate, not a reported result. Label the assumption clearly and avoid making aggressive changes from a small sample.
You can also estimate potential waste:
Potential waste = cost of excluded or clearly misaligned queries
Use this figure to prioritize work, not to claim that every dollar would have been saved. A negative keyword may affect future matching beyond the exact query you observed, so check the intended scope before applying it.
For promotion decisions, compare the query with your existing keyword structure. Promote a search when its intent is clear, the landing page matches it, and separating it would give you better control over bids, budget, ad relevance, or reporting. If several queries express the same intent, use a keyword clustering approach rather than creating a separate ad group for every wording variation. The PPC keyword clustering guide provides a framework for grouping terms by theme and intent.
Inputs and assumptions
Set these inputs before reviewing the report so the analysis does not change from query to query:
- Review window: Use a consistent weekly period, while extending the range when a campaign has low volume.
- Primary conversion: Define which action counts as a business result. Separate primary outcomes from soft actions such as page views or partial form starts.
- Target cost per conversion: Use the account’s approved planning threshold or unit economics, not a generic benchmark.
- Minimum evidence rule: Decide how much spend, traffic, or conversion evidence is needed before promotion or exclusion.
- Relevance rules: Write down prohibited products, audiences, locations, employment or education intent, free-only intent, and other exclusions that matter to the business.
- Attribution setting: Keep the conversion window and attribution approach consistent during comparison.
Use a spreadsheet or dashboard with columns for date range, campaign, ad group, search term, clicks, impressions if available, cost, conversions, conversion value, classification, action, owner, and review date. Add a notes field for ambiguous terms. If tracking is inconsistent, resolve measurement before treating a query as a budget problem. Campaign UTM naming conventions can help connect paid traffic with downstream analytics; see the UTM naming framework.
| Query evidence | Recommended decision |
|---|---|
| Clearly irrelevant wording | Add the appropriate negative keyword and record its scope. |
| Relevant, high-intent query with validated conversion | Consider promoting it into a controlled keyword or ad group. |
| Relevant query with clicks but no conversion | Keep monitoring unless it exceeds the agreed cost or relevance threshold. |
| Ambiguous term with material spend | Investigate the landing page, conversion quality, and search context. |
| Repeated related queries with different wording | Cluster the theme and assess whether dedicated messaging is justified. |
Worked examples
Example 1: Budget-wasting query
A software campaign records a search for a free version of a paid product. It receives 9 clicks at $2.50 each and produces no meaningful conversion. The cost is $22.50, and the intent conflicts with the offer. Classify it as exclude, add a negative keyword at the narrowest suitable level, and check whether related free, template, tutorial, or download searches are also appearing. Do not automatically exclude every query containing one word if that word can also occur in a qualified search.
Example 2: New keyword opportunity
A campaign targeting a broad product category produces a query that names a specific use case. It has 14 clicks, one qualified lead, and a cost below the working target. The query is relevant, the landing page addresses the use case, and the intent is commercial. Classify it as promote. Add a controlled keyword only after checking for duplication, then consider a dedicated ad message and landing-page alignment. This is where search-term data can improve both targeting and creative.
Example 3: Do not overreact
A relevant query has four clicks and no conversion. Its cost is modest and below the escalation threshold. Classify it as keep and monitor. A zero-conversion result at very low volume is not enough to establish that the query is inefficient. Review it again after the next agreed data interval.
Record the action and expected outcome for every change. For example: “Exclude free-version terms from this campaign; review related queries next week,” or “Promote the use-case query; compare conversion quality after sufficient traffic.” A clear log prevents repeated debates and makes Google Ads optimization cumulative.
When to recalculate
Run the workflow weekly when a campaign has enough new search activity to produce decisions. For lower-volume campaigns, use a longer review window rather than forcing a weekly conclusion. Recalculate estimates after meaningful changes to bids, budgets, match settings, landing pages, conversion tracking, geographic targeting, or ad messaging. Revisit the report sooner when spend rises unexpectedly, conversion quality changes, a new product or service launches, or the business changes its definition of a qualified customer.
Use this practical checklist:
- Confirm the date range, conversion action, and campaign scope.
- Sort for cost, conversions, and unusual or newly recurring queries.
- Classify each material query as keep, promote, exclude, investigate, or separate.
- Estimate cost per conversion and potential waste using clearly labeled assumptions.
- Apply only the smallest appropriate keyword or campaign change.
- Record the change, expected effect, and next review date.
- Compare the next period with the same definitions and attribution settings.
Maintain the negative keyword work separately from keyword expansion, and review both for conflicts. The negative keyword maintenance framework can support that process. When a search-term theme needs new messaging, pair the change with a structured ad copy testing framework rather than changing targeting and creative without a record. Repeating this review with consistent inputs turns the Google Ads search terms report into an operating routine: less reactive cleanup, clearer keyword opportunities, and better control over where the budget goes.
