How to Use the Google Ads Search Terms Report to Cut Wasted Spend
The search terms report is the highest-ROI 20 minutes in Google Ads. Here is the weekly mining routine - n-gram thinking, when to harvest vs negate, and what match-type signals actually tell you.
The Search Terms Report Is the Only Report That Shows You Reality
Your keyword list is a hypothesis about what people search. The search terms report is what they actually typed. In every account that has been running for more than a month, the gap between those two things is where the wasted spend lives.
This guide covers the weekly routine: where to find the report, how to read it in a way that scales, when to harvest a term as a keyword versus block it as a negative, and what the match type column is quietly telling you about your account structure.
Where to Find It and How to Set It Up Properly
Go to Campaigns > Insights and reports > Search terms. You can also open it filtered to a single campaign or ad group by selecting rows in the Keywords view first, which is how you should usually work.
Before you read a single row, set the view up:
- Date range: last 30 days for a weekly routine, last 90 days when you are doing a deep clean. Shorter than 30 days and you get noise; longer and you are acting on stale data.
- Columns: Search term, Match type, Added/Excluded, Impressions, Clicks, Cost, Conversions, Cost/conv., Conv. rate. Drop everything else.
- Sort: by Cost, descending. Always. The report can have thousands of rows and you do not have time to read them all - but 80% of the money is in the top 50.
One limitation: Google only shows terms that meet a privacy threshold of activity, so a chunk of your spend appears as "other search terms." That bucket is normal and grows with broad match. You cannot mine it directly, which is a reason to keep some tighter match types in the account rather than going 100% broad.
The Weekly Routine: 20 Minutes, Three Passes
Do this once a week. It is the highest-return recurring task in Google Ads and it does not require expertise, just discipline.
Pass 1: The Money Pass (spend, no conversions)
Sort by cost descending and read the top 30-50 terms. For each one ask a single question: would I be happy to pay for this click again?
The rows you are hunting are the ones with real spend and zero conversions over a meaningful window. On a $60 CPC, five clicks with no conversion is $300 - worth investigating. On a $2 CPC, five clicks means nothing yet. Scale your threshold to your CPC: a useful rule is that a term deserves action once it has spent roughly your target cost-per-acquisition with nothing to show.
Sort those into three buckets:
| Verdict | What it looks like | Action |
|---|---|---|
| Irrelevant | Wrong product, wrong intent, wrong audience entirely | Negative keyword, exact or phrase |
| Relevant but unprofitable | Right intent, wrong economics (too cheap a job, wrong region) | Negative, or lower bid via a dedicated ad group |
| Relevant and untested | Right intent, just not enough data yet | Leave it, note it, recheck next week |
The third bucket matters. Negating too aggressively is a real failure mode - you starve Smart Bidding of the exploratory data it needs, and volume quietly collapses over a few months.
Pass 2: The N-Gram Pass (patterns, not queries)
This is the pass most advertisers skip, and it is the one that scales.
Instead of judging queries one at a time, look for the recurring words across bad queries. That is n-gram thinking: a 1-gram is a single word, a 2-gram is a two-word phrase. If you see "roofing jobs," "roofing careers," "roofer salary," and "roofing apprenticeship" in your report, you do not need four exact-match negatives. You need three phrase negatives - "jobs," "salary," "career" - and you have blocked an entire category of future waste you have not paid for yet.
The recurring offenders across almost every account:
- Job seekers: jobs, careers, salary, hiring, apprenticeship, resume
- Free and DIY intent: free, diy, how to, tutorial, template, homemade
- Research intent: what is, meaning, definition, wikipedia, reddit, forum
- Bargain intent: cheap, discount, coupon, used, second hand
- Wrong audience: for sale, wholesale, supplier, manufacturer (if you sell retail)
- Competitor and comparison terms where you have decided not to compete
To do this properly, export the report to a spreadsheet, split terms into individual words, and pivot cost by word. Fifteen minutes of that on a mature account usually surfaces two or three words that are eating a real percentage of budget. If you would rather have the pattern-finding done for you, the negative keyword generator scans your own search terms and suggests the recurring waste words to block.
Pass 3: The Harvest Pass (winners)
Now flip the sort to Conversions descending. You are looking for terms that convert well but are being served by a loose match on a keyword that does not describe them precisely.
Harvest a term as its own keyword when all of these are true:
- It has converted at least 2-3 times at or below your target CPA.
- Its intent is meaningfully different from the keyword currently serving it - different enough to deserve its own ad copy.
- You can write a landing page or ad that speaks specifically to it.
If those are not all true, leave it alone. Adding a converting search term as an exact keyword inside the same ad group with the same ad changes essentially nothing except your account's complexity. This is the busywork trap: hundreds of harvested exact-match keywords, all pointing at the same ad, all splitting conversion data into piles too small for Smart Bidding to learn from.
The genuine reason to harvest is control - separate ad copy, a separate landing page, or a separate budget for a high-value theme.
Negative Keyword Match Types: Get These Right
Negative keywords do not behave like positive keywords, and the difference causes real damage when misunderstood.
| Negative match type | Blocks | Use it for |
|---|---|---|
| Negative broad | Queries containing all the words, in any order | Blocking a concept, e.g. "free trial" |
| Negative phrase | Queries containing the exact phrase in order | Most n-gram blocks, e.g. "jobs" |
| Negative exact | Only that exact query | One-off terms you must block precisely |
Two rules that save accounts: negative keywords do not match close variants, so you must add plurals and common misspellings separately. And a negative broad match with several words only blocks queries containing *every* one of those words - "cheap roof repair" as a negative broad will not block "cheap roofing."
Place them at the right level too. Universal blocks (jobs, free, porn, reddit) belong in an account-level negative list applied to all campaigns. Campaign-level negatives handle service or product mismatches. Ad group negatives are mainly for steering traffic between ad groups, not for waste.
What the Match Type Column Tells You About Structure
The Match type column shows how the query matched - "Exact match," "Phrase match," "Broad match," or "Exact match (close variant)."
Read it as an audit of your structure:
- Lots of spend arriving via broad match on terms you would never bid on: your negative list is too thin for the amount of broad match you are running. Broad match is only safe with a well-maintained negative list and conversion tracking Smart Bidding can trust.
- "Exact match (close variant)" showing genuinely different intent: Google's close-variant matching is loose now. If "insurance for contractors" is triggering on your exact keyword for "contractor insurance quotes," that is fine. If it is triggering on a different service line, negate it.
- A converting term matching against a keyword in the wrong ad group: your ad groups are not tightly themed, which drags ad relevance and Quality Score down.
Why This Has to Be Weekly
Search terms are not a fixed set. New queries appear constantly as language, products, seasons, and competitors change. An account cleaned thoroughly today will have accumulated new waste within a month, and broad match plus Smart Bidding accelerates that because the system keeps exploring.
The compounding effect is real. Cutting 15% waste this month means next month's budget buys more relevant clicks, which feeds better conversion data to Smart Bidding, which improves targeting further. Skipping the routine for a quarter works the same way in reverse.
If a weekly 20-minute session is not going to happen in practice - and for most business owners it does not - this is exactly the work AdTurbo does daily. It mines your search terms every day, surfaces waste patterns rather than one-off queries, proposes negatives you approve with one click or auto-apply, and logs each change with the reasoning behind it. It is $149/month flat with a 7-day free trial and you can cancel anytime. You can also run the negative keyword generator against your account first to see what it finds before deciding.
Related Guides
- The Google Ads Search Terms Report Explained
- How to Find Negative Keywords
- Negative Keywords: Examples and Lists
- Google Ads Keyword Match Types
Frequently Asked Questions
How often should I check the search terms report?
Weekly for most accounts, and twice a week if you run broad match on a budget above roughly $10,000 a month. Twenty minutes sorted by cost descending catches the majority of new waste before it compounds. A full 90-day deep clean with n-gram analysis is worth doing quarterly.
Why do some search terms not appear in the report?
Google only shows terms that meet a minimum privacy threshold of activity, so low-volume queries are grouped as "other search terms." That bucket tends to grow with broad match usage. You cannot negate from it directly, which is one practical argument for keeping some phrase and exact match keywords in the account.
Should I add every converting search term as a keyword?
No. Only harvest a term when it has converted 2-3 times at or below target CPA and deserves its own ad copy, landing page, or budget. Otherwise you split conversion data across more keywords than Smart Bidding can learn from, without changing what actually serves.
What is the difference between the search terms report and the keywords report?
The keywords report shows the terms you bid on. The search terms report shows what people actually typed to trigger those keywords. With broad match and close variants, the two can differ dramatically - which is precisely why the search terms report is where wasted spend is found.
Can negative keywords hurt performance?
Yes, if you are too aggressive. Over-negating starves Smart Bidding of the exploratory traffic it uses to find new converting queries, and volume erodes over a few months. Block clearly irrelevant intent decisively, but give ambiguous terms more data before you cut them.
