How to stop click fraud on Google Ads

To stop click fraud on Google Ads: use Google's Invalid clicks data, cut bad sources, protect Smart Bidding, claim credits, then automate when lists stop scaling.

1. Add Invalid clicks columns; read what Google filtered.
2. Segment by network, placement, location, device, hour.
3. Exclude bad placements; review partners and Display.
4. Exclude repeat IPs at account level (covers Performance Max).
5. Use only hard conversions as primary conversions.
6. Request an investigation within 60 days.
7. Automate detection when the 500-slot lists fill.
8. Measure traffic quality, then CPA or ROAS.

What Google already does about invalid clicks

Google filters invalid clicks automatically, and when it finds invalid traffic after the billing cycle it issues an invalid activity credit in the next cycle, when appropriate and possible. Advertisers can add the Invalid clicks column to the campaigns data table to see that filtered volume.

What Google does not do is explain its decision per click, IP address or GCLID. You get aggregate columns and, sometimes, credits, not a click-level audit trail. See does Google already protect campaigns, what is click fraud, what is a GCLID, and Google Help About invalid traffic.

Why click fraud is a Smart Bidding problem, not only a budget problem

Performance Max, Target CPA and Target ROAS learn from the clicks they see. When bots, scrapers and click farms look like real visitors, the algorithm treats those profiles as the audience and bids to find more of them; conversion rate falls and CPA rises.

Budget waste is visible; the quieter cost is a bidding model that keeps chasing the wrong people. Fix the data Smart Bidding learns from, not only the invoice. Walkthrough: how invalid clicks break Smart Bidding. Google overview: About Smart Bidding.

How to stop click fraud on Google Ads in 8 steps

Step 1: Add the invalid clicks columns and read what Google already filtered

Add "Invalid clicks" and "Invalid click rate" to the campaigns table, then compare by campaign over 30 to 90 days.

In Google Ads, open Campaigns, click Columns, Modify columns, and add Invalid clicks and Invalid click rate. Sort by Invalid click rate and scan 30, 60 and 90 day windows. Spikes on Search partners, Display or one campaign are your first leads. Google filters invalid clicks automatically and may credit invalid activity it finds after billing. Honest limit: the columns show what Google already caught. They do not prove what slipped through.

Step 2: Find where suspicious clicks come from

Segment by network, placement, location, device and hour, then compare with GA4 engagement and CRM lead quality.

Break the same date range by Network (Search, Search partners, Display), Placement, Location, Device and Hour of day. Look for clusters: high clicks, thin engagement, weak CRM lead quality. In GA4, check engagement time and any bot filter you already run; cross-check paid sessions against form fills or purchases. Guide: filter bot traffic in GA4. Honest limit: segmentation finds patterns; it does not name a competitor or botnet.

Step 3: Tighten targeting where invalid clicks cluster

Exclude placements that send clicks without engagement, review Search partners and Display, and check location settings.

Where Display or partner placements show clicks with no on-site engagement, exclude those placements. Turning a placement off stops ads serving there; it does not credit past clicks. Review whether Search partners or Display are worth the invalid click load for that campaign. Confirm location options target people in or regularly in your locations, not people interested in them, when that matches your offer. More on Display: Google Display Network click fraud. Honest limit: exclusions cut future spend on bad inventory. They do not reverse billed clicks you already paid for.

Step 4: Exclude repeat IP addresses at account level

Build an account-level IP exclusion list for addresses you can justify from logs as repeat, non-engaging clickers, and keep campaign lists for temporary or campaign-specific blocks.

Campaign-level IP exclusion lists hold up to 500 addresses and do not apply to Performance Max. Account-level IP exclusions apply across campaigns including Performance Max. An asterisk can replace the last block of digits when you need to cover a justified range. Only exclude IPs you can justify from your own logs or reports. Never exclude a shared network on one spike. Step-by-step UI: exclude IP addresses in Google Ads. Google Help: Exclude IP addresses. Honest limit: static IP lists do not stop rotating VPNs, proxies or mobile carriers.

Step 5: Protect Smart Bidding data

Set only hard conversions (completed purchase or submitted lead) as primary conversion actions; keep soft events secondary or out of bidding.

Only hard conversions (a completed purchase or a submitted lead) should count; soft events such as page views or form starts teach the model the wrong pattern. If you run a click tracker that classifies each ad visit (see Step 7), you can also fire a Google Ads conversion or GA4 key event only when the visit was classified as legitimate, using the tracker's classification cookie as a condition in Google Tag Manager. Pair this with account-level exclusions so junk profiles stop feeding the model. Deep dive: how invalid clicks break Smart Bidding. Related: why Google Ads conversions drop. Honest limit: cleaner conversion signals help learning; they do not replace source exclusions.

Step 6: Request credits for invalid clicks Google missed

If you have evidence Google missed invalid activity, request an investigation within 60 days and watch Billing for Invalid Activity or Invalid Clicks credits.

Google's window to request an invalid traffic investigation is 60 days. Credits appear in Google Ads under Billing, Transactions (or Documents) as "Invalid Activity" or "Invalid Clicks". They are credits or adjustments, not cash refunds. Manual review takes about 4 to 8 weeks; Google does not guarantee a credit for every flag. Google does not explain its decision per click, IP address or GCLID. Process and expectations: Google Ads refunds and invalid activity credits. Honest limit: a credit is not guaranteed; the clock starts when the clicks occur.

Step 7: Automate detection and blocking when manual lists stop scaling

Move to a click tracker when the 500-slot lists fill, sources rotate through VPNs and proxies, or junk leads keep arriving after manual exclusions.

Look for Google Certified Click Tracker status, signals beyond IP, layered risk (allowed, monitored, blocked), an account-level tracking template, and a monitoring phase before any exclusions go live. Compare options on best click fraud protection software. Product page: Google Ads click fraud protection.

ClickPatrol is a Google Certified Click Tracker that checks every click against 800+ data points and blocks only high-risk sources, so shared office and mobile networks are not banned on one IP. It installs as an account-level tracking template. On a paid plan it adds known bot servers to the exclusion list as soon as the account is connected. It reports invalid activity to Google so Smart Bidding stops learning from those profiles. Monitor Only Mode lets you watch the scoring before anything is excluded. Detection detail: how detection works. Template help: tracking template merger. Certified list: List of certified Google click trackers. Honest limit: automation scales exclusions; you still own conversion setup, whitelists and threshold rules.

Step 8: Measure whether it worked

Write the hypothesis first, freeze other major account changes, then judge traffic quality before bidding metrics.

First, did suspicious, invalid, repeated or low-quality traffic stop entering the campaigns? Second, did cleaner traffic improve CPA, conversion rate, lead quality or ROAS? Write a hypothesis, freeze unrelated experiments, and run long enough for Smart Bidding and lead cycles to settle. Honest limit: change bids, creatives and exclusions in the same week and you cannot isolate the fraud fix.

How to block competitor clicks specifically

IP exclusion stops repeat clicks from the same static address. That works when a rival sits on a fixed office IP and clicks your ads every morning. It fails when they use VPNs, proxies or mobile networks that rotate addresses.

For repeat clickers, a click fraud tool can add click threshold rules that block an IP after too many clicks in a time window, for example 3 clicks in 10 minutes or 5 clicks in 1 day. Start generous, because real people click twice. In that tool, whitelist IPs that must never be blocked (office, agency, testers, known customers) before you tighten thresholds; a blacklist keeps known bad IPs blocked. Signals beyond the IP matter once lists fill or sources rotate. Full playbook: competitors clicking your ads.

Click fraud on Microsoft Ads and Meta

A click tracker's tag can classify ad visitors beyond Google too. On Microsoft Ads, use audience exclusions plus IP export. On Meta, classify visitors through GTM and exclude the suspicious audience; the tool does not hard-block on Meta. Guides: Bing Ads click fraud and Meta Ads click fraud.

Checklist

• Add Invalid clicks and Invalid click rate columns; review 30 to 90 days by campaign.
• Segment by network, placement, location, device and hour; cross-check GA4 and CRM.
• Exclude low-engagement placements; review Search partners, Display and location settings.
• Move repeat bad IPs to an account-level exclusion list (Performance Max covered).
• Primary conversions = hard actions only; soft events off bidding.
• Request invalid traffic investigation within 60 days; check Billing for credits.
• When lists fill or proxies rotate, evaluate a Google Certified Click Tracker with monitoring first.
• Hypothesis, freeze other changes, then measure traffic quality and CPA or ROAS.

See it on your own campaigns first

See it on your own campaigns first. The free 7-day ClickPatrol trial is monitoring only: it shows which clicks look invalid without blocking anything. Blocking starts when you choose a paid plan. Fit for self-advertisers: small businesses. Prefer a human review first: free click fraud analysis.

Sources

• Google Ads Help, About invalid traffic
• Google Ads Help, About columns in your statistics table
• Google Ads Help, Exclude IP addresses
• Google Ads Help, About Smart Bidding
• Google Ads Help, List of certified Google click trackers