Search queries and keywords answer different questions
A keyword is a targeting instruction you add to an account. A search term is the query someone actually enters. Broad match, phrase match and close variants can connect a keyword with searches that use different wording. The report exposes those connections so you can judge the visitor's likely need against the offer on the landing page.
A query is not waste simply because it differs from the keyword. Someone searching for a replacement after a breakdown may be a better prospect than someone repeating your category name. Read the complete phrase, the campaign promise and the destination together before deciding whether to exclude it.
Where to find the Google Ads search terms report
In Google Ads, open Campaigns, expand Insights and reports, then select Search terms. Choose the account, campaign or ad group scope and set a date range. Interface labels can change, so use the account's navigation search for Search terms if the menu differs. Remove old filters before treating the visible rows as the whole report.
The standard table is most useful for keyword based Search campaigns. Shopping can also provide search term reporting, although it does not use positive keywords in the same way. Performance Max reporting surfaces and available detail differ. Check the campaign type before expecting the same columns or exclusion workflow everywhere.
Columns that explain relevance and spend
Add campaign and ad group before exporting. The same query can be waste in one campaign and useful in another. Include the matched keyword where the report provides it, and keep the network context so search partner traffic does not silently blend into your interpretation.
| Column | What it tells you | How to use it |
|---|---|---|
| Search term | The reported query wording | Read the complete intent before choosing an exclusion |
| Campaign and ad group | Where the query entered the account | Decide whether the problem is local or shared |
| Keyword | The targeting keyword where applicable | Find broad targeting or routing problems |
| Match type | How the query relates to the keyword | Do not treat this as a proposed negative match type |
| Impressions | How often the ad appeared | Understand exposure without confusing it with spend |
| Clicks and cost | Visits and their recorded cost | Prioritize expensive mismatches |
| Conversions | Actions counted by the selected setup | Check the action definition and reporting delay |
| Conversion value | Value assigned to those actions | Distinguish revenue or qualified value from raw volume |
| Cost per conversion | Spend relative to counted conversions | Compare relevant traffic with your economic target |
| Added or excluded | Existing targeting or exclusion status | Avoid repeating work without checking scope |
A lead form, an imported qualified lead and a completed purchase are different outcomes. Check which actions appear in Conversions and which appear only in All conversions. A report built around minor engagement actions can make irrelevant searches look successful. For reporting context beyond this table, use the Google Ads reporting guide.
Sort by cost before judging efficiency
Sort cost from highest to lowest, then read the largest spend rows in their campaign context. Start with obvious offer mismatches such as recruitment queries reaching a patient acquisition campaign. Cost provides an order for review, not proof that the query deserves a negative.
Separate decisions into exclude, retain and investigate. Exclude when the requested service, product or audience clearly falls outside the offer. Retain relevant demand even when its wording is unfamiliar. Investigate when the problem could be a weak landing page, incomplete measurement or a sales cycle that has not matured.
Zero conversion queries need a second check
Filter for cost greater than zero and conversions equal to zero to find candidates. Avoid a universal spend cutoff. The appropriate observation period depends on your conversion delay, lead quality process, expected acquisition cost and available evidence. A fresh query with a relevant offer can remain valuable before a conversion appears.
Compare recent activity with an older date range that allows conversions to arrive. Check tracking health and offline imports before excluding a whole theme. For lead generation, inspect qualified outcomes where available. A query with form fills but no serviceable leads can need different treatment from a query with no recorded form fill.
Use an n-gram view to find repeated intent
An n-gram is a word or sequence of words extracted from queries. Grouping phrases such as "job application" or "repair manual" reveals repeated intent that scattered rows can hide. Aggregate their cost and conversions, then open the original queries before proposing a phrase negative.
An n-gram total is not an additional spend bucket. One query can contribute to several overlapping groups, so adding every group's cost together double counts spend. Inspect converting queries within the group and keep word order visible. The negative keyword match types guide explains why an n-gram and a blocking rule are different things.
What other search terms hides
Google does not expose every individual query. Privacy thresholds and reporting criteria can leave searches out of the row level table. Other search terms represents aggregated activity for searches without individually reported rows in that view. Its cost can matter even when the actual wording is unavailable.
You cannot recover hidden queries by subtracting visible rows or exporting the same table again. Insights may provide broader categories, but a category is not the missing query text. Compare totals within the same date range, network and scope, and check active filters before attributing a difference entirely to privacy thresholds.
Treat the visible report as partial evidence. Document how much activity is reported individually and avoid claiming complete query coverage. NegativeKeywords reviews the search term data the platforms expose. A tool connection does not remove platform privacy limits or reveal hidden searches.
Review around changes and conversion delay
Check more frequently after launching a campaign, changing the offer, expanding match types or introducing a new product feed. New matching patterns can surface quickly. Stable campaigns still need regular review because seasonal language, inventory and user intent change. Schedule a deeper recurring review alongside alerts for expensive new mismatches.
Keep recent relevance checks separate from mature performance decisions. An unmistakable job query does not need a completed conversion cycle to reveal its intent. A commercially relevant comparison query does. Record the date range and reason for each decision so a later reviewer can understand the evidence.