Loading tutorials…
Loading tutorials…
Negative keywords are the highest-leverage tactical lever in Google Ads — small effort, large recurring ROI. The catch: doing them once is worthless. Doing them monthly compounds for years. Here is the workflow specialists actually run.
Who this is forOwners running Google Ads campaigns who suspect they are paying for queries that will never convert. If your Search Terms report has any results that make you wince, this tutorial is for you.
What you'll need
Step 1
Campaigns → All campaigns → Insights & reports → Search Terms. Date range: last 90 days. Sort by cost descending. The top 50-100 terms drive most of your spend.
Click Campaigns → All campaigns → Insights & reports (left nav) → Search Terms.
Set the date range to "Last 90 days." Less than 30 days has too little data; more than 90 mixes in stale queries.
Add the columns you need: Search term, Match type, Campaign, Ad group, Clicks, Cost, Conversions, Conv. value.
Sort by Cost descending. The top 50 search terms drive most of your spend regardless of how many keywords you have.
Export to CSV (the download icon, top right) so you have a working copy. Mark up the CSV as you triage — Google's UI does not let you take notes per row.
Quick triage: for each of the top 50, ask "Does this represent a real buyer for me?" Mark each row Green (yes, intent-aligned), Yellow (maybe, depends on context), or Red (no, never going to convert).
Step 2
Group your Red queries into themes. Common themes: job-seeker queries, informational queries, competitor names, geographic mismatches, free/DIY queries.
Pattern 1 — Job-seeker queries: '[role] jobs,' '[product] career,' '[brand] salary,' '[skill] hiring.' These almost never convert for B2B/B2C products. Even if you do hire, you do not want to pay $5-15 per click for it.
Pattern 2 — Informational queries: 'how to,' 'what is,' 'tutorial,' 'guide,' 'free,' 'DIY,' 'definition,' 'meaning.' These are awareness-stage searches — usually no buying intent.
Pattern 3 — Competitor brand names: '[Competitor] login,' '[Competitor] support,' '[Competitor] vs [Other Competitor].' Decide deliberately whether you bid on competitor terms; if not, exclude them.
Pattern 4 — Geographic mismatch: searches with city/state/country names you do not serve. Geo targeting catches most of these but not all — some users in your geo search for products meant for other geos.
Pattern 5 — Synonym mismatch: queries that share keywords with your product but mean something different. Example: 'apple pie recipe' if you sell Apple-branded products.
Pattern 6 — Brand safety / off-color: queries that mention adult content, illegal substances, or competitor disparagement. Exclude these regardless of conversion potential.
Pattern 7 — Long-tail noise: 5+ word queries that almost never repeat. Worth excluding the recurring noise terms (e.g., "free download," "is it worth it") even if the full query is unique.
Step 3
Negative match types differ from positive: Broad negative blocks any term containing the word; Phrase blocks exact phrase order; Exact blocks the exact term only. Use Broad sparingly — it can over-block.
Negative Broad Match: [free] blocks any search term containing 'free' anywhere — 'free shipping,' 'free trial,' 'is it free,' 'free vs. paid.' Useful for clear single-word junk terms. Dangerous if the word has both negative and positive uses ('free shipping' is good intent for many products).
Negative Phrase Match: ['free trial'] blocks search terms containing the exact phrase in order — 'looking for free trial,' 'free trial signup.' Does NOT block 'trial of the free version' because word order differs. Use for multi-word junk phrases.
Negative Exact Match: [[free trial]] blocks ONLY the exact term 'free trial,' nothing longer or shorter. Use sparingly — almost too narrow to be useful.
Default recommendation: Phrase Match for most negatives. Broad Match for clear single-word junk (jobs, salary, careers, hiring, DIY, tutorial, definition, meaning). Exact Match almost never.
Decision rule: if the word has both junk and intent uses, Phrase Match the bad phrase. If the word is always junk regardless of context, Broad Match it.
Example: "free" alone Broad Matches both "free shipping" (good) and "free download" (bad). Use Phrase Match: ["free download"], ["free template"], ["free version"].
Step 4
Ad group level: blocks at the smallest scope. Campaign level: blocks across that campaign. Shared lists: blocks across multiple campaigns. Use shared lists for account-wide junk.
AD GROUP LEVEL: useful when a query is irrelevant for one ad group but valid for another. Example: "iPhone case" is junk in your Android Case ad group but valid in your iPhone Case ad group.
CAMPAIGN LEVEL: blocks the negative across all ad groups in that campaign. Most account-specific junk lives here.
SHARED NEGATIVE LISTS: Tools → Shared Library → Negative keyword lists. Build a master "Junk Words" list that you apply to every campaign. Common starter list: jobs, careers, salary, hiring, free, DIY, tutorial, how to, definition, meaning, lyrics, image, photo, picture, video, song, movie, Reddit, Wikipedia, YouTube.
To apply a shared list: open the campaign → Keywords → Negative keywords → click 'Use a saved list' → pick the shared list.
Maintain 2-4 shared lists by theme: 'Account-wide junk' (universal noise), 'Brand exclusions' (your own brand if you bid separately), 'Competitor names' (if you do not bid on competitors), 'Audience-mismatch' (e.g., job seekers).
Why shared lists: when you add a new junk term once, every campaign benefits. Without shared lists, you add the same negative to 10 campaigns separately and miss some.
Step 5
After adding negatives, wait 24-48 hours then check the Search Terms report again. The blocked terms should be gone (or showing 0 impressions).
Open the campaign → Keywords → Negative keywords. Click the campaign or ad group level. Click "+ Add negative keywords."
Paste your negative keywords, one per line. Wrap Phrase Match in quotes ("free download") and Exact Match in brackets ([[free download]]).
Save. Negatives are usually active within 15-30 minutes.
Wait 24-48 hours. Then re-pull the Search Terms report for the same date range.
Verify: the queries you negative-keyword'd should now show 0 impressions in the most recent 24-48 hours. They may still appear in the report for the historical date range (because they served in the past), but new impressions should be zero.
If a query is STILL serving despite the negative, possibilities: (a) you put the negative at ad group level but the query came through a different ad group; (b) match type mismatch — Phrase Match negative did not catch the variant; (c) the negative is on a different campaign; (d) reporting lag, wait another 24 hours.
Step 6
Calendar reminder: first Monday of every month, 30-min Search Terms mining session. Refresh shared lists quarterly.
Set a recurring calendar reminder: first Monday of every month, "Google Ads — Negative keyword mining."
Each session: pull last 30 days of Search Terms, sort by cost descending, scan top 50, add 5-15 new negatives.
Cadence matters more than session depth. 30 min/month for 12 months beats 6 hours every other quarter.
Quarterly: review shared lists. Some negatives become obsolete (e.g., a competitor goes out of business and you can drop their name). Some need broadening as new junk patterns emerge.
Annually: review the entire negative keyword list against current business. Have you launched new products that overlap with what you previously excluded? Negative keywords get sticky and outdated faster than you think.
Step 7
Apply shared lists to Performance Max campaigns (rolled out late 2025). Use account-level brand exclusions for brand protection across all campaigns.
Performance Max negative keywords: PMax now accepts negative keyword lists (rolled out late 2025). Open the PMax campaign → Negative keywords → add your shared lists.
Account-level Brand exclusions: Tools → Brand exclusions → set up your own brand and competitor brands here. Applied to all PMax campaigns in the account automatically. Prevents PMax from cannibalizing brand Search.
Negative keywords also apply to YouTube and Display campaigns (campaign-level). For Display, also add 'Placement exclusions' (Tools → Placement exclusions) to block specific websites or apps where your ads serve.
When you find a particularly bad query in one campaign, add it to your shared list AND apply the shared list everywhere — including PMax. One discovery, fleet-wide protection.
Common mistakes
Adding negatives once and never again
What goes wrong: Search terms drift constantly. New products, new search behaviors, new junk patterns emerge weekly. An untouched negative list from 90 days ago is letting through hundreds of new junk variants. Effective CPC creeps back up.
How to avoid: Calendar reminder, first Monday of every month, 30 minutes. Same time every month so it becomes habit. Less than 30 min/month is the most cost-effective protection in the account.
Using negative Broad Match for terms with mixed usage
What goes wrong: Negative Broad [free] blocks 'free shipping' (good intent) along with 'free download' (junk). You lose conversions because the broad negative was too aggressive.
How to avoid: For ambiguous words, use Phrase Match negatives on the specific bad phrases. Reserve Broad Match negatives for words that are always junk regardless of context (jobs, salary, careers).
No shared negative lists
What goes wrong: You add 'jobs' as a negative on Campaign 1 but forget on Campaigns 2-10. The same junk pattern leaks through 9 other campaigns. Wastes $200-800/mo per missed campaign.
How to avoid: Build 2-4 shared lists (Account-wide junk, Brand exclusions, Competitor names, Audience mismatch). Apply to every campaign. Maintain centrally.
Adding negatives at wrong scope
What goes wrong: You add 'iPhone' as a campaign-level negative on a campaign that has both Android Cases and iPhone Cases ad groups. Now your iPhone Case ad group cannot serve.
How to avoid: Add negatives at the smallest scope that solves the problem. Ad-group level for ad-group-specific exclusions. Campaign level for campaign-specific. Shared list only for truly account-wide junk.
Ignoring Quality Score impact
What goes wrong: Letting irrelevant queries match drags down Quality Score on the matching keywords (because CTR is lower for irrelevant queries). Lower QS = higher effective CPC. Compounds with budget waste.
How to avoid: Aggressive negative keyword discipline lifts Quality Score over 30-60 days. Track top-spend keyword QS before and after mining sessions to see the lift.
Skipping PMax negative keywords
What goes wrong: PMax did not accept negative keywords until late 2025. Many older accounts have not updated their workflow to add PMax negatives. Junk queries keep flowing through PMax even when blocked in Search.
How to avoid: Apply your shared negative keyword lists to every PMax campaign. Tools → Shared Library → Negative keyword lists → attach to PMax campaigns.
Recap
Done — what's next
How to lower Google Ads CPC without losing conversions
Read the next tutorial
Hand it off
Negative keyword work is invisible until it stops happening. Then CPC drifts up, irrelevant clicks creep in, and the account quietly loses 15-30% efficiency over 6 months. EverestX Google Ads specialists who do this monthly typically pay for their entire fee through negative-keyword work alone — usually $40-80 per monthly session, recovering $200-1,500 in wasted spend depending on account size.
See specialist rates
Healthy accounts typically have 200-1,500 negative keywords across shared lists and campaign-level. There is no upper limit that matters — every negative that prevents one bad click pays for itself instantly. Worry about volume only if negatives are blocking legitimate traffic (rare, but check by reviewing the Search Terms report after each session).
Yes, but lightly. Brand campaigns have inherent intent. Add negatives for clearly off-pattern queries (e.g., 'YourBrand jobs,' 'YourBrand layoffs,' 'YourBrand vs. CompetitorYouDon\'tWant'). Avoid aggressive negatives on brand — you want broad capture of intent variations.
Mechanically the same — both campaign types accept negative keyword lists. Operationally different — PMax is more opaque, so you cannot see exactly which queries are being blocked. Apply negatives to PMax based on what you see in Search campaign Search Terms (the patterns repeat) and via PMax's broader Insights → Search categories.
Yes — if you put a negative that conflicts with a positive keyword you bid on, the negative wins. Example: bidding on 'running shoes' but adding [running] as a Broad negative blocks every variant. Always cross-check negatives against your active keyword list before saving.
Usually within 15-30 minutes for new auctions. The Search Terms report has a 24-48 hour reporting lag, so verifying via the report takes a day. The auction itself respects new negatives immediately.
Several third-party tools (Optmyzr, Adalysis) automate Search Terms analysis. They are helpful for scale but not necessary — most accounts under $20K/mo spend can be managed manually in 30 min/month. The judgment of which queries are junk vs. opportunity is hard to automate well.
Google Ads
CPC creeping up while conversion rate stays flat means you're paying more for the same outcome. These six tactics drop CPC 20-40% on most accounts. Some take 1 hour. Some take 1 month. All matter.
Google Ads
A Search campaign is the highest-intent ad format Google offers — people typing exactly what they want, right when they want it. Setting it up well is also the workflow Google works hardest to flatten into 'just trust the algorithm.' Here is the real build.
Google Ads
Performance Max is Google's most opaque ad product — and for accounts with the right ingredients, also one of the highest-ROAS. The catch: every default Google offers is tuned for Google's revenue, not yours. Here is the build that gives you back control.
Google Ads
Your campaign is enabled, budget is set, and Google says "Eligible." But spend is at 5% of daily budget and impressions are flat. Here's the diagnostic sequence specialists run.
Google Ads
DIY Google Ads is a great idea — until it isn't. This is the honest framework: when the cost of self-managing exceeds the cost of hiring, and how to tell which side you're on.