The Google Ads search terms report shows the actual searches that triggered your ads. It is the evidence you need to decide whether a query should become a negative keyword, a new target, a landing-page improvement, or no change at all.
The common mistake is to open the report, sort by cost, and exclude every row without a conversion. That can cut waste, but it can also block relevant demand, distort a Smart Bidding campaign, and hide problems in conversion tracking or landing-page routing.
A useful search-term audit separates query intent from short-term performance. The 30-minute workflow below does both.
Search term vs keyword: the essential distinction
A search term is what a person typed into Google. A keyword is the targeting input an advertiser added to an ad group.
For example:
- Keyword:
emergency plumber - Search term:
24 hour frozen pipe repair near me
Those phrases can be different because keyword matching evaluates more than literal text. The report’s Keyword column identifies which keyword matched the search. Its Match type column describes how closely the search term related to that keyword; it is not necessarily the same as the match type configured on the keyword.
Google explains this distinction in its search terms report documentation.
How to open the Google Ads search terms report
In the current Google Ads interface:
- Select Campaigns in the left navigation.
- Open Insights and reports.
- Select Search terms.
- Set the account, campaign, ad group, and date range you want to inspect.
If you use the Report editor’s predefined search-term report instead, Google notes that you cannot add keywords or negative keywords directly from that view.
For Performance Max, use the dropdown for Search terms and landing pages for Performance Max. That version adds landing-page and ad-format context. Our Performance Max negative-keyword guide covers the campaign-specific controls and risks.
Why report totals do not show every query
The search terms report contains terms used by a significant number of people. Some lower-volume searches are omitted to meet Google’s data-privacy standards.
Search terms insights handle this differently: they group queries into themes and subthemes and can aggregate low-volume demand into a category or “other queries” without exposing every phrase. Insights are useful for identifying demand patterns. The report is more useful for deciding what to do with a visible, individual query.
Do not force the visible rows to reconcile to every click or impression. Treat the report as actionable evidence, not a complete transcript of all search activity.
Build the right report before judging terms
Use a date range long enough to capture conversion lag and enough traffic for the account. Then add or confirm these fields:
| Field | Question it answers |
|---|---|
| Search term | What did the user actually ask for? |
| Keyword | Which target caused the match? |
| Campaign and ad group | Where did Google route the query? |
| Match type | How closely did the term match the keyword? |
| Added/Excluded | Has someone already acted on it? |
| Impressions and clicks | How often did it appear and earn traffic? |
| Cost | How much evidence have we paid for? |
| Conversions and conversion value | Did it produce the outcomes used for bidding? |
| Cost per conversion or value/cost | Is performance economically viable? |
| Landing page, where available | Did the query reach the right destination? |
Segment by conversion action when mixed goals could hide the truth. A query that generated three low-value page views and no qualified leads should not be treated as a three-conversion success.
Use five decisions, not just “negative or keep”
Every meaningful row should end in one of five queues.
1. Exclude
The intent is outside the offer, location, customer type, or acceptable context.
Examples:
plumbing certification classesfor a residential repair companywholesale plumbing supplyfor a service-only businessemergency plumber buffalo nyfor a Toronto-only campaign
These can justify negatives even before they accumulate many clicks because the business mismatch is explicit.
2. Add or strengthen targeting
The search is relevant and reveals demand worth treating deliberately.
Before adding it as a new keyword, ask whether the existing keyword and ad group already cover the intent well. Adding every converting query creates unnecessary duplication. A new keyword is useful when it enables better ad copy, landing-page routing, bidding, reporting, or budget control.
3. Fix the destination or message
The intent is valuable, but the ad or landing page does not answer it.
A search for same day tankless water heater repair that lands on a generic plumbing homepage may need a better route and offer—not an exclusion.
4. Fix measurement or operations
The query appears valid, but the recorded outcome cannot be trusted.
Check call tracking, form deduplication, primary conversion settings, offline qualified-lead imports, revenue values, spam, and the team’s response process. Paid-search optimization cannot repair a broken definition of success.
5. Observe
The query is relevant but has too little evidence for a confident decision. Leave it active and give the row a specific review trigger, such as another full conversion-lag window or a spend threshold based on the campaign’s economics.
“Observe” should be a queue with a next review date, not a permanent place to avoid decisions.
A decision matrix for negative keywords
Use intent first, then economics.
| Intent | Performance evidence | Default action |
|---|---|---|
| Clearly irrelevant | Any amount | Exclude at the narrowest safe scope |
| Relevant | Profitable or promising | Keep; consider stronger targeting or creative |
| Relevant | Weak | Check tracking, conversion lag, landing page, offer, and operations |
| Ambiguous | Low volume | Observe or use a narrow exact negative only when the downside is clear |
| Ambiguous | Material spend | Review the actual SERP intent, calls, forms, and downstream lead quality |
Do not build a “waste” filter using only Conversions = 0. Recent clicks may still convert, and offline outcomes may not have returned to Google yet.
Calculate waste in business terms
There is no universal number of clicks after which a query becomes bad. Use the account’s allowable acquisition cost and downstream rates.
For lead generation:
Allowable cost per raw lead = target cost per qualified lead × qualified-lead rate
If the target cost per qualified lead is $200 and 40% of raw leads qualify:
$200 × 0.40 = $80 allowable cost per raw lead
That gives the audit a meaningful reference point. It still does not turn every relevant term above $80 into an automatic negative. First check sample size, conversion lag, close-rate differences, and whether the term produced qualified offline outcomes that tracking missed.
For ecommerce, compare query-level contribution margin or conversion value with cost. Revenue alone can make low-margin or heavily returned products look healthier than they are.
Choose what text to block
The entire search term is often not the best negative keyword.
Search term: free plumbing certification classes online
Possible negatives:
[free plumbing certification classes online]blocks only that exact query."certification classes"blocks searches containing that ordered phrase.certification classesas negative broad blocks queries containing both words in any order.freeblocks a much larger category and could also remove a valuable query such asfree estimate for emergency plumber.
Extract the smallest phrase that captures the unwanted intent without overlapping a profitable offer.
Google says negatives added directly from the search terms report default to negative exact match. You can change the match type with symbols before saving. Its negative-keyword ideas instructions also stress that negative match types work differently from positive match types and do not expand to variants in the same way.
Negative broad, phrase, and exact match
Use the negative plumbing jobs as an example.
| Match type | Syntax | plumbing jobs near me | jobs in plumbing | plumbing job openings |
|---|---|---|---|---|
| Negative broad | plumbing jobs | Blocked | Blocked | May show |
| Negative phrase | "plumbing jobs" | Blocked | May show | May show |
| Negative exact | [plumbing jobs] | May show | May show | May show |
Negative broad requires every term, in any order. Negative phrase requires the ordered phrase. Negative exact blocks the exact term.
Negative keywords are literal controls, not semantic instructions. Google documents that they do not automatically cover all variants and gives house versus houses as an example in its negative-keyword troubleshooting guide. Revisit the report after applying a negative to see what still appears.
If the report labels queries as close variants or an apparently related negative fails to block them, use the Google Ads close-variants diagnostic to separate positive keyword expansion from negative matching behavior.
Choose the right level
The destination is as important as the text.
Ad group
Use an ad-group negative when a query belongs elsewhere inside the same campaign. This can reduce internal routing conflicts without excluding demand from the entire campaign.
Campaign
Use a campaign negative when the intent is invalid for that campaign but could be valid elsewhere in the account.
Negative keyword list
Use a shared list when a defined policy should apply to several selected campaigns. Name lists by purpose—such as Employment intent or Unsupported services—and document an owner. Avoid one giant list called Master negatives with no explanation.
Account
Use account-level negatives only for terms that should be blocked across all relevant Search and Shopping inventory in the account. Google says the account-level list can affect Search, Performance Max, Shopping, App, Smart, and Local campaigns. One overlooked campaign can turn a convenient account-wide exclusion into a costly mistake.
The 30-minute search-term audit
Minutes 0–5: define the window and economics
- Set the date range with conversion lag in mind.
- Write down the primary conversion, target acquisition cost, and relevant value measure.
- Confirm which campaign or account is in scope.
Minutes 5–10: validate the report
- Add keyword, campaign, ad group, cost, conversion, value, and Added/Excluded fields.
- Check whether conversion tracking changed during the period.
- Separate brand and non-brand demand if their economics differ.
Minutes 10–20: classify terms
- Start with high-cost rows, but judge intent before performance.
- Tag each meaningful row Exclude, Target, Destination, Measurement, or Observe.
- Record a reason; do not rely on color alone.
Minutes 20–25: design exclusions
- Extract the smallest safe negative phrase.
- Choose exact, phrase, or broad match deliberately.
- Choose ad group, campaign, list, or account scope.
- Check for overlap with profitable queries and active keywords.
Minutes 25–30: approve and log
- Have an accountable person review material changes.
- Apply the approved negatives.
- Record text, match type, destination, source query, reason, date, and approver.
- Set the next audit or follow-up date.
At the end, you should have a reviewable decision log—not just a longer negative-keyword list.
How often should you review search terms?
Set cadence based on how quickly the account buys new evidence.
- High-spend or rapidly changing campaigns may need daily or several-times-weekly triage.
- Stable accounts with moderate query volume may need a weekly review.
- Low-volume campaigns can use a longer window, provided urgent brand-safety and wrong-market terms are handled promptly.
The correct cadence is the interval before preventable waste becomes material, balanced against the need for enough data to judge valid intent. Use the time since the last successful audit as the next window so weekends, holidays, and skipped review days do not create gaps.
Common audit failures
- Excluding every non-converter: This confuses limited performance evidence with irrelevant intent.
- Copying a universal negative list: Words such as
free,cheap,used, orjobscan have different value by business and campaign. - Ignoring match type: The right idea at the wrong breadth can block qualified traffic.
- Using account scope by default: Local or campaign-specific mismatches become global exclusions.
- Ignoring close wording: A literal negative may not cover the singular, plural, synonym, or reordered version you expected.
- Never revisiting lists: Offers, locations, inventory, and campaign structures change.
- Automating the push without approval: A classification error becomes a live traffic loss.
- Measuring success as lower spend: The goal is better profitable demand, not simply fewer clicks.
Turn the audit into a controlled system
Spreadsheets work for a small account, but recurring audits become difficult across many campaigns and clients. The scalable part to automate is evidence collection and recommendation preparation.
LeadUp’s Google Ads negative keyword tool reviews available search terms with campaign structure, targeting, active keywords, and existing negatives. It flags irrelevant intent, explains the reason, and prepares a suggested negative. You choose which recommendations to accept, edit the text or match type, select the destination, and approve the push.
Audits can run on selected days and pick up where the last successful review ended. That creates a repeatable search-term review without giving an automated classifier unchecked authority over the account.
If you want to test the workflow on your own data, run a free search term audit.
Bottom line
The search terms report is not a list of mistakes. It is a record of real demand and how Google routed it.
Classify intent first. Use performance and business economics second. Fix measurement or landing pages when the query is valid, and add a negative only when the demand truly should not enter that scope. A short, well-documented exclusion list built from actual searches is safer and more useful than hundreds of generic negatives copied from another account.