Documentation
Search Term Insights gives you a structured view of what Google Ads is actually matching your ads to. It surfaces where budget is leaking to non-converting queries, flags search terms that are semantically misaligned with the keywords that triggered them, and shows which queries are splitting performance signals across multiple campaigns or ad groups. Use it to prioritize exclusions, discover keyword expansion opportunities, and strengthen account structure before you take action.
- Best for
- Search term cleanup, intent analysis, and account structure review
- Core modules
- Non-converting vs. converting spend, Query Keyword Similarity, Double Matched Queries
- Primary pivots
- Search term, ad group, campaign, and account
- Similarity score range
- A relative score where lower means weaker query-keyword alignment
- Inline action
- Exclude a reviewed search term directly from campaign or ad group context
#What Search Term Insights does
Search Term Insights organizes raw Google Ads search-term data into three focused modules. Each module answers a different question about where your traffic comes from, whether it belongs, and how your account structure is handling it. The page is for analysis and prioritization. Use it to build a clear picture before applying exclusions or requesting keyword additions.
The three modules at a glance
| Module | Core question it answers | Typical outcome |
|---|---|---|
| Non-Converting vs. Converting Search Terms | Where is spend going to queries that never produce a conversion, and how large is that gap? | Identify waste targets, understand budget efficiency, and prioritize cleanup by campaign or ad group. |
| Query Keyword Similarity | How closely does each search term match the keyword that triggered it semantically? | Spot mismatched intent for potential exclusion or discover new keyword clusters worth adding. |
| Double Matched Queries | Which queries trigger across more than one campaign or ad group, splitting learning and attribution? | Assign clear query ownership, add negatives where needed, and reduce internal competition. |
#Who benefits most
Search Term Insights is most valuable for account managers and campaign owners who want to go beyond automated smart bidding and keep a close eye on what Google Ads is actually matching. The three modules cover different seniority levels of the same problem: from quick waste identification to deeper structural diagnosis.
Common use cases
Regular spend audits
Review non-converting spend weekly or monthly to catch budget leaks before they compound. Sort by absolute non-converting spend to find the highest-priority exclusion targets first.
Post-launch quality checks
After a new campaign goes live, use similarity scores to confirm that broad and phrase match keywords are attracting the right intent, not adjacent noise.
Account structure reviews
Use Double Matched Queries to find campaigns and ad groups competing for the same queries. This is especially common in large accounts after multiple rounds of restructuring.
Keyword expansion planning
Converting queries with low similarity scores often reveal high-performing intent clusters you have not explicitly modeled yet. These are candidates for new keywords or ad groups.
#Reading non-converting vs. converting spend
The non-converting vs. converting view separates search-term spend by whether any conversion was recorded on that query. It shows absolute spend and a percentage split for each search term, ad group, and campaign. The spend visualization at the top of the module gives you an instant read on how much of your budget is going to queries that have not produced a result.
Start with absolute non-converting spend to find the campaigns and ad groups where cleanup has the most budget impact. Then look at the non-converting percentage to understand concentration. A campaign with a very high non-converting percentage may be a priority even if total spend is lower, because it means nearly every query it triggers is wasted.
Interpreting spend patterns
| Pattern you see | What it usually means | How to respond |
|---|---|---|
| High absolute non-converting spend with few or no conversions | A meaningful share of budget is going to queries with no measurable return. | Check search intent, the matched keyword, and the landing page before excluding or adjusting bids. |
| High non-converting percentage but low total spend | The signal may be early-stage or low volume. | Monitor rather than act unless the query is clearly irrelevant. Wait for more data. |
| Mixed converting and non-converting spend on the same term | The query may convert in some contexts but not others. | Pivot by campaign or ad group to find where it performs and where it does not before excluding. |
| Strong converting spend with low similarity | The query performs despite being semantically distant from the matched keyword. | Do not exclude. Consider adding it as a managed keyword with its own ad group and landing page. |
| High non-converting spend across many campaigns for the same term | The term may be irrelevant to your business entirely, not just to one campaign. | Consider a shared negative list or account-level exclusion after reviewing across all pivots. |
#Understanding Query Keyword Similarity
Query Keyword Similarity assigns each search term a relative score that reflects how closely it aligns semantically with the keyword that matched it. A low score means the two are very different in meaning. A high score means the query is textually and semantically close to the keyword. The score is generated by semantic analysis, not by Google Ads match type logic, so it captures intent distance rather than surface-level keyword overlap.
The similarity breakdown also shows where the most low-similarity queries are happening by account and ad group, so you can pinpoint which parts of your structure are generating the most mismatched traffic before you dig into the individual query table.
Similarity score interpretation guide
Always read similarity alongside performance metrics. The score explains alignment. It does not replace conversion data, CPA, ROAS, or business judgment.
| Similarity pattern | How to interpret it | Recommended action |
|---|---|---|
| Low similarity, spend present, no conversions | The query is semantically distant from the keyword and is not converting. Strong signal of mismatched intent. | Review for exclusion at the correct level. Confirm the keyword match type is as tight as needed. |
| Low similarity, spend present, conversions present | The query performs despite being different. It may reveal an intent cluster you have not yet modeled. | Do not exclude. Evaluate whether a new keyword, ad group, or landing page would serve this intent better. |
| High similarity, weak performance | The query matches the keyword well, but the economics are poor. Landing page, bid, or audience may be the issue. | Look at CPA, ROAS, conversion value, and volume. Act on performance, not similarity. |
| High similarity, strong performance | The query is a close, high-value match for the keyword. | Add it as a managed keyword if it is not already. Protect it from broad exclusions. |
#Finding and resolving double matched queries
Double Matched Queries shows search terms that triggered ads in more than one campaign or ad group. When the same query fires across multiple parts of an account, it fragments performance data, creates internal competition between your own ads, and makes attribution harder to trust.
For each repeated query, the module shows where it triggered, how often, and how it performed in each location, including spend, conversions, and other key metrics. The goal is to decide which campaign or ad group should own the query and use negatives to prevent it from triggering in the wrong places.
How to resolve duplicate triggered queries
| Situation | What to do | Why it matters |
|---|---|---|
| One placement converts clearly better than the others | Keep the query in the best-performing campaign or ad group. Add a negative for it in the weaker placements. | Consolidates conversion signals into one part of the account and stops internal CPC competition. |
| Performance is similar across placements | Choose the campaign with the most relevant ad and landing page as the owner. Exclude it elsewhere. | Even when performance looks similar, duplicate triggering fragments Smart Bidding's learning. |
| The query is irrelevant in all placements it appears | Exclude it from all campaigns where it triggered. Consider a shared negative list if the term is globally irrelevant. | Stops waste from a term that should not be in the account at any level. |
| Duplicates are widespread across many terms | Review match types, keyword grouping, and campaign boundaries before adding individual negatives. | Widespread duplication usually signals a structural problem that individual negatives will not fully solve. |
#How to use filters and pivots effectively
Each module in Search Term Insights includes a dynamic pivot table that you can reorient by search term, ad group, campaign, or account. The right pivot depends on the question you are trying to answer. Use account-level pivots to identify patterns and prioritize. Switch to campaign and ad group pivots when you are ready to act, so you exclude at the correct level.
When to use each pivot
| Pivot level | Best used for | Watch out for |
|---|---|---|
| Account | Spotting broad waste patterns, recurring intent gaps, and duplicate-query themes that affect the whole account. | Account averages can hide campaign-specific winners buried in a weak overall picture. |
| Campaign | Prioritizing budget cleanup by comparing absolute non-converting spend and conversion value across campaigns. | A high non-converting spend campaign may also be your strongest converter. Always compare both sides. |
| Ad group | Understanding whether a query belongs in one ad group and not another, and checking keyword-level fit. | Matched keyword, similarity score, and conversion metrics need to be reviewed together at this level. |
| Search term | Reviewing the exact query, its similarity score, both spend columns, and all performance metrics. | Do not act on a single metric in isolation. Review the full row before deciding. |
Sort by non-converting spend when the goal is finding waste. Sort by conversions or conversion value when looking for expansion opportunities. Sort by similarity score when hunting for intent mismatches. Apply campaign and ad group filters before triggering an exclusion so the action targets the correct level.
#Recommended review workflow
- Set the scope and date range Start with the account or campaign you are responsible for. Choose a date range that has enough recent spend and conversion data to be meaningful, typically at least 30 days. Use a longer window for lower-volume accounts.
- Identify the largest non-converting spend Sort by absolute non-converting spend to find where cleanup has the most budget impact. Note the top campaigns and ad groups before looking at individual queries.
- Investigate with similarity scores Switch to the Query Keyword Similarity module and pivot by the campaigns flagged above. Low-similarity, high-spend, zero-conversion rows are the clearest exclusion candidates, so start there.
- Check for duplicate triggering Open Double Matched Queries and look for terms that appear in your priority campaigns and in one or more other places. Confirm where each query performs best.
- Decide the action for each term For each flagged query, choose one of four outcomes: exclude it (directly from the Insights table or via Search Terms Optimization), add it as a keyword, restructure ownership using negatives, or flag it for monitoring if the data is insufficient.
- Confirm the right exclusion level Use the ad group and campaign pivot to verify where the negative should land before applying it. A term that is irrelevant in one ad group may be useful in another.
- Re-check after changes land Return to the same pivots 2 to 4 weeks after applying exclusions or structural changes. Confirm that non-converting spend and duplicate triggering moved in the right direction.
#Search Term Insights vs. Search Terms Optimization
Search Term Insights and Search Terms Optimization both work with search-term data, but they serve different parts of the workflow. Insights is the diagnostic layer: it helps you understand what is happening and build your action list. Optimization is the execution layer: it applies rule-based logic, logs decisions, and supports automation.
Choosing the right tool
| Workflow | Purpose | Use it when |
|---|---|---|
| Search Term Insights | Analyze converting and non-converting spend, similarity scores, and duplicate triggered queries before deciding what to do. | You are diagnosing patterns, investigating structure, or building a prioritized action list. |
| Inline Exclude (from Insights table) | Exclude a single reviewed term from the exact campaign or ad group shown in the table. | The term is clearly irrelevant, the exclusion level is obvious, and you are acting on one term at a time. |
| Search Terms Optimization | Run rule-based exclusion workflows with configurable thresholds, preview modes, logs, email summaries, and automation controls. | You want a repeatable, auditable process for handling inefficient search terms at scale. |
#Limits and common pitfalls
Search Term Insights reflects the data available in your Google Ads account for the selected date range and scope. Some important caveats apply to how the data should be read and acted on.
Known limits and how to work around them
| Limit | Why it matters | How to handle it |
|---|---|---|
| Short date ranges may show incomplete conversion data | Conversion lag means some queries that will eventually convert look non-converting in a short window. | Use at least 30 days of data, and longer windows for products or services with multi-day buying cycles. |
| Low-volume queries can show extreme percentages | A query that spent €2 and converted once looks like 100% converting. That is not a reliable signal. | Filter or sort by absolute spend before acting. Minimum volume thresholds matter. |
| Similarity score reflects semantic distance, not commercial fit | A semantically distant query can still be commercially valuable. Similarity is a signal, not a verdict. | Always cross-reference with conversion data and business context before excluding based on score alone. |
| Duplicate triggering may reflect intentional structure | Some accounts run the same query in a branded and a generic campaign by design. | Review the campaign strategy before adding negatives. Not all overlap is accidental. |
#Troubleshooting
Common questions
Why are some search terms showing zero conversions even though I know those queries perform?
What does it mean when a query has a similarity score of 0?
I see the same query in the Double Matched Queries table dozens of times. Do I need to add a negative for each row?
Can I use Search Term Insights to find new keywords to add?
The non-converting spend percentage for a campaign looks very high, but the campaign manager says it performs well. What is happening?
How is an Exclude action in Insights different from adding a negative in Google Ads directly?
#Best practices
These practices help you get consistent value from Search Term Insights reviews without over-excluding or acting on unreliable signals.
- Always review absolute non-converting spend before percentage split. Volume context prevents false urgency.
- Use a minimum of 30 days of data; extend to 60 to 90 days for low-volume campaigns or long purchase cycles.
- Never exclude a query based only on a low similarity score. Always check conversion data first.
- Pivot to ad group level before applying an exclusion to confirm the action targets the right scope.
- Treat converting low-similarity queries as keyword expansion signals, not waste.
- Review Double Matched Queries at least quarterly in large or frequently restructured accounts.
- Use inline Exclude for single confirmed terms; use Search Terms Optimization for repeatable patterns.
- Keep a running list of strong converting queries identified in Insights to feed into Keyword Additions or Keyword Opportunities reviews.
- After applying exclusions, return to the same pivots 2 to 4 weeks later to verify the intended impact.