Every Google Ads account leaks money. Not through bad bidding or weak creative, but through search terms you never intended to pay for: job seekers, DIY researchers, competitor names, “free” hunters and product variants you don’t even sell. Negative keywords in Google Ads are the cheapest lever you have to stop that leak, and the only reliable way to use them is a repeatable routine rather than a one-off cleanup.
This guide gives you that routine: how to mine the search terms report for wasted spend, how to decide whether an exclusion belongs at account, campaign or ad group level, how to pick the right match type, and how to avoid the classic mistake of blocking queries that were actually converting.
Negative keywords in Google Ads: the short answer
A negative keyword tells Google Ads: do not show my ad for searches containing this word or phrase. It never triggers an ad, it only blocks one. Adding free as a negative keyword means your ad stops competing for “free crm software” while it keeps competing for “crm software pricing”.
Three things make negative keywords behave differently from the keywords you bid on, and almost every wasted-spend problem comes from ignoring them:
- No close variants. Negative keywords do not automatically cover plurals, misspellings, abbreviations or accents. Excluding
shoedoes not excludeshoes, and excludingAdidasdoes not excludeadiddas. - No synonyms or intent matching. Blocking
cheapdoes nothing for “low cost” or “budget”. - Negative broad match requires every word. A negative broad keyword only blocks a query that contains all of its words, in any order.
Negative keyword match types, with examples
Choosing the match type is where most accounts either over-block or under-block. Here is exactly how each one behaves. Complete Guide to Negative Keywords in Google Ads tackles the same question from another angle.
| Match type | How to write it | Blocks | Does not block |
|---|---|---|---|
| Negative broad | running shoes free | Any query containing all three words in any order: “free running shoes”, “running shoes for free” | “free shoes”, “running shoes cheap” |
| Negative phrase | “free running shoes” | Queries containing that exact word sequence: “buy free running shoes online” | “running shoes free”, “free shoes running” |
| Negative exact | [free running shoes] | Only the identical query, nothing more, nothing less | “free running shoes uk”, “free running shoes for men” |
The practical rule of thumb
- Single word, negative broad for themes that are always wrong for your business:
jobs,salary,torrent,crack,diy,meaning. - Negative phrase for modifiers that are usually wrong but occasionally attached to commercial intent:
"how to","for sale","open source","used". - Negative exact for a single expensive query you want gone without touching its longer, converting cousins.
If you hesitate between broad and phrase, start with phrase. It is the reversible choice. Broad single-word negatives are powerful and they are also how accounts accidentally lose 30% of their conversions overnight.
Step 1: Mine the search terms report for wasted spend
The search terms report is the raw material. In the Google Ads interface, open your Search campaign and go to Audiences, keywords and content > Search keywords > Search terms. You can also run it across the account from the campaigns overview. It is done convincingly by a studio that gets this right.
The pull that finds money in 10 minutes
- Set the date range to the last 30 days (use 60 or 90 days for low-volume accounts so patterns are visible).
- Add the columns you actually need: Search term, Match type, Campaign, Ad group, Cost, Clicks, Impressions, Conversions, Cost per conversion, Conversion rate.
- Apply a filter: Conversions = 0 and Clicks > 2 (or Cost above a threshold that matters in your account, for example 2x your target CPA).
- Sort by Cost, descending. You are now looking at your most expensive zero-return queries, top first.
- Work down until the cost per row becomes trivial. Stop there. Chasing 0.40 rows is not a good use of your hour.
Then do an n-gram pass
Row-by-row review catches the obvious. The pattern review catches the systemic. Export the report to a spreadsheet, split every search term into single words and two-word combinations, then pivot on cost, clicks and conversions per token.
What you get is a list like this, which tells you far more than 300 individual rows:
| Token | Cost | Clicks | Conversions | Verdict |
|---|---|---|---|---|
| free | 612 | 318 | 1 | Exclude broad, account level |
| template | 288 | 141 | 0 | Exclude broad, campaign level |
| vs | 203 | 77 | 6 | Keep, move to its own ad group |
| salary | 96 | 52 | 0 | Exclude broad, account level |
Note the third row. “vs” looks like research intent, but it converts. That single line is the reason you never exclude a word because it “feels” wrong.
The eight wasted-spend buckets to look for
- Free and pirated intent: free, gratis, trial, crack, torrent, download, apk.
- Jobs and careers: jobs, job, salary, career, hiring, internship, cv, resume, vacancy.
- DIY and information: how to, what is, meaning, definition, tutorial, examples, pdf, wikipedia, reddit, forum.
- Wrong commercial stage: reviews, complaints, cancel, refund, login, support, phone number (unless you want support traffic).
- Wrong price tier: cheap, cheapest, discount, second hand, used, wholesale, bulk (or the reverse if you are the budget option: luxury, premium, custom).
- Wrong geography: city, country or language terms you cannot serve.
- Wrong product or service: variants, sizes, colors, brands and models you do not stock. This is usually the biggest bucket in ecommerce.
- Competitors: only if competitor traffic performs badly for you. Test before you block it.
One important caveat: the search terms report does not show every query. Terms below a privacy threshold are hidden, so your reported spend will never fully reconcile with your keyword spend. That is normal. It also means theme-level exclusions and tight keyword structure do work that row-by-row review cannot.
Step 2: Choose the right level (account, campaign, ad group)
Where you add an exclusion matters as much as what you exclude. Use this decision table.
| Level | Use it for | Examples | Risk |
|---|---|---|---|
| Account level | Terms that are wrong for the entire business, forever | jobs, salary, porn, torrent, competitor brand you never target | High if you are careless. Nothing in the account can override it. |
| Shared list applied to campaigns | Reusable themes shared by a group of campaigns | Informational modifiers, free intent, unserved locations, out-of-stock brands | Medium. Easy to audit and remove, applies where you choose. |
| Campaign level | Separating campaigns from each other, protecting budgets | Brand terms excluded from generic campaigns, product line A excluded from campaign B | Low |
| Ad group level | Routing queries to the right ad and landing page | Excluding “blue widgets” from the red widgets ad group | Low, but tedious to maintain at scale |
The rule that prevents 90% of accidents
An exclusion goes to the account level only if you can finish this sentence honestly: “No campaign in this account, now or in the next twelve months, should ever serve on a query containing this word.” If you hesitate, use a shared list applied to specific campaigns instead.
Step 3: Add negative keywords, step by step
At campaign or ad group level
- Open the campaign (or ad group) in Google Ads.
- Go to Audiences, keywords and content > Search keywords, then open the Negative search keywords tab.
- Click the plus button, choose whether you are adding at campaign or ad group level, then paste one term per line.
- Write the match type into the term itself: plain text for broad, “quotes” for phrase, [brackets] for exact.
- Save, then check the list once to confirm the match types display as you intended.
Building a shared negative keyword list
- Go to Tools > Shared library > Negative keyword lists (in some accounts this appears under exclusion lists).
- Click the plus button, name the list using a convention you will still understand in a year, and paste your terms.
- Save, then use Apply to campaigns and select the campaigns that should inherit it.
- Repeat for each theme. One list per theme beats one giant list, because you can apply themes selectively.
At account level
Account-level negative keywords live in your account settings and apply to Search, Shopping and Performance Max inventory across the account. Treat this as a short, curated list of absolute exclusions, not a dumping ground.
Performance Max, Shopping and Demand Gen
These campaign types do not use a keyword list the way Search does, but you can still steer them:
- Performance Max: campaign-level negative keywords are available, plus brand exclusions and account-level negatives. Use the Insights tab and search category reporting to find themes worth excluding.
- Shopping: campaign and ad group negatives work as in Search. Feed quality (titles, product types) does the rest of the filtering.
- Search partners and Display: negatives apply to search queries, not to placements. Use placement exclusions for content.
Step 4: Organize your lists so they stay usable
Unmanaged negatives become a black box nobody dares touch. A simple naming and structure convention fixes it.
A structure that scales
| List name | Contains | Applied to |
|---|---|---|
| NEG – Global – Junk | Jobs, adult, piracy, academic | All search and shopping campaigns |
| NEG – Global – Free and DIY | free, gratis, diy, homemade, open source | All paid-product campaigns |
| NEG – Global – Informational | “how to”, “what is”, definition, pdf, forum | Bottom-funnel campaigns only |
| NEG – Brand | Your brand and its misspellings | Generic and competitor campaigns |
| NEG – Competitors | Competitor names and misspellings | Generic campaigns (not the conquesting campaign) |
| NEG – Geo not served | Cities, regions, countries out of scope | All local campaigns |
| NEG – Products not sold | Brands, models, sizes you do not carry | Shopping and generic search |
Keep a change log
Maintain a simple sheet with four columns: date, term added, match type, level and reason. When performance drops six weeks later, this log is how you find the cause in two minutes instead of two hours. It is also what lets a new team member touch the account without breaking it.
Watch the limits
Google Ads caps how many negatives you can store: a few thousand per shared list, a limited number of lists per account, a large but finite number per campaign, and a much smaller cap for account-level negatives. Exact numbers change, so check the current figures in Google Ads Help before planning a mass upload. The practical takeaway: curate, do not hoard. Ten thousand negatives is usually a symptom of loose keyword structure, not good hygiene.
The repeatable monthly routine
Here is the cadence that keeps waste down without turning into a full-time job.
| Frequency | Task | Time |
|---|---|---|
| Weekly (high-spend accounts only) | Top 20 zero-conversion search terms by cost, last 7 days. Block the obvious. | 15 min |
| Monthly | Full search terms pull, n-gram analysis, add exclusions, promote winning terms to keywords, log changes. | 60 to 90 min |
| Monthly | Check the negative keyword conflicts report and fix anything blocking an active keyword. | 10 min |
| Quarterly | Audit every list: remove obsolete exclusions, re-test competitor and “cheap” blocks, review list assignments. | 45 min |
| On every launch | Apply the standard global lists before the campaign goes live. | 5 min |
The monthly session, in order
- Pull the search terms report, last 30 days, all Search and Shopping campaigns.
- Filter to zero conversions and meaningful cost, sort by cost descending.
- Tag each row: Exclude / Keep / Promote to keyword / Needs more data.
- Run the n-gram pivot to catch themes hiding in the long tail.
- Sanity check every candidate exclusion: search the full report for that token and confirm total conversions are zero or negligible.
- Assign level and match type using the tables above.
- Apply via shared lists wherever possible, direct campaign negatives only for structural separation.
- Log the changes with dates and reasons.
- Review two weeks later: impressions, clicks, conversions and CPA per campaign, compared to the previous period.
How to avoid blocking queries that convert
This is where negative keywords go wrong, and the damage is invisible: you do not get an alert saying “you just blocked 18% of your revenue”. A longer analysis is available for anyone who wants it.
Six guardrails
- Always test the token before excluding it. Filter the full search terms report for that word across all time, not just the 30-day window. If it has any conversions, downgrade from broad to phrase or exact, or skip it.
- Never exclude a word that appears in one of your active keywords. Check the negative keyword conflicts report after every batch.
- Beware of short and ambiguous words:
pro,plan,service,app,online,near me. Each of these has ended someone’s good month. - Do not exclude a competitor name at account level unless you are certain you will never run a conquesting campaign or a comparison page.
- Add exclusions in batches you can undo. Twenty terms per session with a clear log beats 500 pasted from a generic “negative keyword list generator”.
- Question generic downloadable lists. A word like
revieworusedis toxic for one advertiser and profitable for another. Filter every list against your own data before applying it.
How to spot that you over-blocked
Compare the two weeks after your changes to the two weeks before, at campaign level:
- Impressions down sharply and conversions down proportionally: you cut muscle, not fat. Roll back the most aggressive negatives.
- Impressions down, conversions flat or up, CPA down: exactly what you wanted.
- Impressions down, budget suddenly underspent: your negatives are too tight or your keyword coverage was thinner than you thought.
Bonus: connect exclusions to what happens after the click
Negative keywords only fix the top of the funnel. If you cannot see which campaign, ad group or search theme produced a signup, you will keep excluding the wrong things. Consistent, correctly tagged destination URLs (campaign, source, medium, content) let you cross-check Google Ads data against your analytics and CRM, so you can distinguish “expensive query” from “expensive query that becomes a customer 60 days later”. If your UTM tagging is inconsistent across campaigns, fix that before you start pruning aggressively, ideally with a URL builder that enforces a single naming convention across your team.
Negative keyword examples by business type
| Business | Typical exclusions |
|---|---|
| SaaS | free, open source, crack, tutorial, “how to build”, jobs, salary, api docs, login |
| Ecommerce | used, second hand, wholesale, diy, pattern, “how to make”, brands and sizes not stocked |
| Local services | jobs, courses, licence, “do it yourself”, out-of-area cities, “free quote” if unprofitable |
| B2B lead gen | internship, cv, template, pdf, definition, “case study” (if not gated), student |
| Education | free course, torrent, answers, cheat, salary, jobs |
FAQ
How do you write negative keywords in Google Ads?
Type the term the way you want it matched: plain text for negative broad (free trial), quotes for negative phrase ("free trial"), brackets for negative exact ([free trial]). One term per line. Remember that negative keywords ignore close variants, so add plurals and common misspellings explicitly.
Why do I only see negative keywords in Google Ads and no regular keywords?
You are almost certainly on the Negative search keywords tab rather than the Search keywords tab, or you are looking at a campaign type that does not use bidding keywords (Performance Max, Demand Gen, Display, Video). Those campaign types accept exclusions but do not have a positive keyword list. Switch tabs, or open a Search campaign. A fuller account is out there.
How do I find negative keyword lists in Google Ads?
Open Tools > Shared library > Negative keyword lists. From there you can create a list, edit its terms, and see or change the campaigns it is applied to. To see the negatives active on a specific campaign, open the campaign and check the Negative search keywords tab, which shows both direct negatives and applied lists.
What are examples of negative keywords?
The universal ones are job-related (jobs, salary, hiring), free-intent (free, torrent, crack), informational ("what is", definition, pdf), and DIY (diy, homemade, "how to make"). Beyond those, the best examples come from your own search terms report, because they depend on what you actually sell.
How many negative keywords should an account have?
There is no target number. A tightly structured account with clear keyword themes may need a few hundred. A broad-match-heavy account may need thousands. If your list keeps growing every month with no drop in wasted spend, the problem is structure or keyword match types, not the number of negatives.
Do negative keywords affect Quality Score or impression share?
They do not directly change Quality Score, but by removing low-relevance impressions and clicks they usually improve CTR and conversion rate on the traffic that remains. They will lower your impression share for the excluded terms, which is the entire point. Watch lost impression share due to budget: it often improves, because your budget stops being consumed by irrelevant clicks.
Can negative keywords block a search term I want?
Yes, and this is the main risk. A single broad negative like pro can silently block dozens of profitable queries. Always test the token against your full historical search terms data, prefer phrase over broad when in doubt, and check the negative keyword conflicts report after each batch.
Key takeaways
- Mine the search terms report every month, filtered to zero conversions and sorted by cost.
- Add an n-gram pass to find themes the row-by-row view hides.
- Use account level only for absolutes, shared lists for reusable themes, campaign and ad group for routing.
- Default to negative phrase, use broad for clear junk, exact for surgical fixes.
- Log every change, and review impressions versus conversions two weeks later.
Done consistently, this routine takes about an hour a month and it is usually the single highest-return hour in a paid search account.