Start free

Google Ads Click Fraud

Google Ads click fraud: what Google filters, what you still pay for, and how to check

Where invalid clicks show up in your account, which junk clicks your own settings buy, how to ask Google for a credit, and when a blocking tool is worth paying for.

Cover: Google Ads click fraud, a large $0 for invalid traffic Google detects before billing

Short answer

  • Google doesn't charge for invalid clicks it catches before billing, and the Invalid clicks column counts them. Clicks it catches later come back as a credit on the invoice, so check Billing too.
  • Many clicks that feel like fraud are valid by Google's rules. Presence or interest location targeting and search partners are on by default, so split clicks by network and location type first.
  • A Click Quality Form request covers the last 60 days, and Google won't say which individual clicks it judged invalid.
  • Google's click filters can't stop fake form leads. reCAPTCHA, server-side validation and qualified-lead goals are your side of the job.
  • There is no industry-standard invalid click rate, so treat any vendor's fraud percentage as that vendor's own claim.

Split clicks by network and location in Claude or ChatGPT

Start free

3-day free trial · no card needed

Google Ads doesn't charge you for click fraud it catches before billing, and the Invalid clicks column shows how much it filtered. Clicks caught after billing come back as a credit. You still pay for clicks Google counts as valid, often bought by your own location and network settings, plus fake leads on your form.

Checked 13 November 2026 against Google Ads Help1.

We run a performance marketing agency and build AdCtrl, which connects Google Ads to Claude or ChatGPT. It doesn't detect fraud, block clicks or get refunds, so there's no blocker to sell on this page, only the checks we'd run first. For a full settings audit, use our Google Ads audit checklist.

  • $0 charged for invalid clicks caught before billing1
  • 60 days how far back Google investigates a request
  • No standard for a "normal" invalid click rate

Does Google charge you for click fraud?

Google doesn't charge for invalid clicks it detects before billing. In Google's words, "You aren’t charged for invalid traffic detected before billing." Traffic it finds after you were billed comes back "in the form of a billing adjustment", and Google counts competitors "who manually click to increase your spend" as invalid activity too1.

Google files click fraud under a wider term, invalid traffic: traffic that "doesn't represent genuine interest in your business", from bots and accidental clicks to people who click on purpose.

So start by asking which bucket a suspicious click sits in. On clicks caught before billing, a vendor pitch frames Google as crediting clicks back later, while Google's page says those clicks are never charged:

TrafficGuard's blog says

TrafficGuard "blocks the invalid ones before they are charged rather than crediting them back later".

Google Ads Help says

"You aren’t charged for invalid traffic detected before billing." Credits only cover what Google finds after the invoice.

Screenshot of Google Ads Help, About invalid traffic, with the note highlighted: You aren't charged for invalid traffic detected before billing.

Screenshot of Google's invalid traffic help page, read on 13 November 2026, with the note highlighted: invalid traffic detected before billing isn't charged.

A tool can catch clicks Google misses, and Google may already have filtered the clicks a tool reports. TrafficGuard's own page points readers to the same Invalid clicks column for what Google already filtered, and says to "Start with Google Ads controls"2. You can see Google's side in your own account.

How do you check invalid clicks in Google Ads?

You check invalid clicks in the places Google reports them, and each one counts something different. The Invalid clicks column counts clicks you were never billed for. Credits for clicks caught later sit in Billing, not in the campaign table:

WhereWhat it showsHow to open it
Invalid clicks columnClicks Google filtered and didn't chargeCampaigns → Columns → search "Invalid clicks" → Apply
Invalid click rateFiltered clicks as a share of all clicks, filtered plus billed (the API metric's definition)Column picker, if your account offers it
Invalid traffic cardEstimated savings and credited cost for the date range; rolling out, so not in every account yetCampaigns → Overview
Billing summaryCredits labelled "Invalid activity"Billing → Summary → Month card → Adjustments
Invalid Activity Credit ReportClicks, cost and conversion rate adjusted for credits, plus a "Credited clicks" columnReport Editor

Five places Google reports invalid clicks. The two columns count clicks filtered before billing, the Invalid traffic card shows both savings and credits, and Billing and the credit report show credits for clicks caught after billing.

Source: Google Ads Help, About invalid traffic, and the Google Ads API v26 metrics reference1, 3

A high invalid click count next to a small credit is normal, because the two numbers cover different clicks. In a made-up example account, 800 clicks are filtered in a month and never charged, and a later $12 credit covers other clicks caught after the invoice.

Google also freezes campaign metrics at the end of the month, so credits "won’t retroactively adjust your campaign data". To see campaign numbers with credits taken out, use the credit report.

Google Ads also removes filtered clicks from its own reports while your analytics may still record the visit. That's one reason GA4 and Google Ads disagree, covered in why Google Ads and GA4 conversions don't match.

Which bad clicks does Google count as valid?

Google counts a click as valid when a real person in your targeting clicked, even if that person was never going to buy. Those clicks are billed like any other, and they can look a lot like fraud in the reports. A bad click can sit in one of four buckets, and each shows up in a different part of the account:

Comparison of four kinds of bad click: filtered by Google and never billed (Invalid clicks column); caught later and credited (Billing adjustments, credit report); valid by Google's rules but bought by your settings (network and location reports); fake leads on your form (your CRM)

The four places a bad click can sit. Only the first two are invalid traffic in Google's sense; the third is bought by your settings and the fourth happens on your own form.

The third bucket is the one you control. Google's own FAQ on traffic spikes lists campaign changes, like new keywords or location settings, as a cause, and suggests switching from "Presence or interest" to "Presence". These are the settings that buy junk clicks:

SettingGoogle's defaultWhat to look for
Location optionPresence or interest, which Google recommends for Search; it includes people only interested in your areaClicks from people outside your area, with cost and no conversions
Search partnersOn for new Search campaigns; Google doesn't show which partner site ran the adPartner clicks with cost and no conversions in the network split
Search termsYour match types decide which searches qualifySearches that cost money and don't fit what you sell
Display and Performance Max placementsAds can run on sites and apps you didn't pickDisplay placements with spend and no results; for Performance Max, placements you wouldn't want your ad on

Settings that buy junk clicks Google counts as valid. Each is a choice you can change, so check them before paying for a blocker.

Source: Google Ads Help on location options, the Search Network, Performance Max and placement reports4, 5, 6, 7, 8

The check takes five steps. Steps 1 to 4 only read; step 5 changes a campaign, so it needs someone allowed to edit it.

  1. Split clicks by network.

    Manual: in the Campaigns table, segment by network, and compare Google Search with Search partners and Display for cost and conversions9.

    With AdCtrl: write "Show clicks, cost and conversions for each Search campaign by network for the last 30 days."

  2. Split clicks by location type.

    Manual: compare clicks from people in your targeted area with clicks from people interested in it, then open the campaign's Settings, then Locations, and check the location option.

  3. Read the search terms.

    Manual: run the weekly search-term routine with the Keyword column on.

  4. Check where Display and Performance Max ads ran.

    Manual: on Display, review the placements with spend and no results; on Performance Max, the placement report shows where ads ran, not results, so use it to spot sites you don't want. Exclude those placements and use content suitability settings to filter content.

    With AdCtrl: write "Show cost and conversions for each Display and Performance Max campaign for the last 30 days", then review and exclude placements in Google Ads.

  5. Change the setting that bought the clicks.

    Manual: in Google Ads, switch the location option to Presence, or switch search partners off under Networks in campaign settings.

The audit checklist's settings checks cover the same networks and location options from the audit side. Our rule there holds here: switch off only what the split shows isn't converting.

Is a competitor clicking my ads?

A competitor clicking your ads is invalid activity in Google's own definition, so Google treats it like any other invalid click: filtered if caught before billing, credited if caught later. Google won't tell you who clicked: details on "specific clicks" stay confidential, so the account can't show you a competitor by name.

You can see the pattern. A jump in the Invalid clicks column means Google caught more. A jump in billed clicks with no matching rise in conversions, from one place or one keyword, is worth a closer look in your own server logs.

Clicks from outside your target area have ordinary causes too. Your ad can show to someone elsewhere whose search names your location, or who searches on a country domain such as google.fr when you target that country. Google also says it can't guarantee location accuracy in every case, since location comes from signals like IP address and device settings.

How do you get a refund for invalid clicks?

You ask for a credit with the Click Quality Form, Google's invalid traffic investigation request, and it covers activity within the last 60 days1. Specialists review it, which "typically takes several business days", and you get the findings by email. If they find invalid traffic the filters missed, Google tells you about any credit.

Google says an investigation "won’t be able to confirm if specific clicks were marked as valid or invalid", so the answer is a credit or no credit, not a click-by-click verdict. Gather this first:

  • Your 10-digit Google Ads customer ID
  • The exact dates of the suspicious activity, inside the last 60 days
  • The campaign and ad group names where it happened
  • The keywords that drove the unusual traffic
  • IP addresses from your web server logs with repeated clicks or clicks without conversions
  • Browser and device details for those visits
  • The GCLIDs (Google click IDs) of the suspicious clicks
  • For bad leads, a spreadsheet with each lead's GCLID and what happened when you called or emailed
  • One or two lines on why it looks wrong, such as a click spike without more conversions
Checklist of what to gather before a Click Quality Form request: customer ID, dates within the last 60 days, campaigns and ad groups, keywords, IP addresses, devices, GCLIDs, lead outcomes and a short explanation of the pattern

What Google asks for before an invalid traffic investigation. GCLIDs are what let Google map a visit back to a specific ad click.

Without GCLIDs and server logs, a request has little to work with, and the 60-day window keeps moving. If your site doesn't keep either, start now:

Do IP exclusions and click-fraud tools work?

IP exclusions work for an address you're sure about, and Google lets you exclude up to 500 IP addresses per campaign, or exclude addresses across all campaigns at account level, Performance Max included11. The catch is shared addresses: blocking an office or mobile carrier address can block real customers on the same network.

Paid click-fraud tools are a separate category. Google says third-party tools "often report different data than Google Ads" and often flag traffic Google already filtered. The options describe themselves like this:

  • Google's filters: Filters before billing, credits what it finds later, Click Quality Form for the rest

    (Built in)

  • IP exclusions: Up to 500 per campaign, or account level across campaign types

    (Built in)

  • TrafficGuard: Scores "every click in real time" and says it connects to Google Ads, Performance Max included (its blog, 13 Nov 2026)

    (Vendor)

  • ClickGuard: Covers Google Ads, Microsoft Ads and Meta Ads and "responds to fraudulent activity in real-time" (its homepage, 13 Nov 2026)

    (Vendor)

  • Other blocking tools: Paid tools that monitor clicks; check each one's own page for coverage and pricing

    (Category)

Both descriptions come from the vendors' own pages, read on 13 November 2026, and neither page names the Google Ads setting the tool changes to stop a click2, 12. There's no honest spend threshold for when a tool pays off, so our view rests on the account's shape:

Skip a paid tool if

  • The account is mostly Search, on Presence, with search partners off
  • The Invalid clicks column shows Google filtering and conversions look normal

Consider one if

  • Display, Performance Max or partner traffic carries a large share of spend
  • Clicks are expensive and you need your own logs to back a credit request

Fix the form first if

  • The problem is junk leads, not junk clicks
  • Leads arrive with fake details or numbers from outside your market

Why do I get fake leads when invalid clicks are low?

Fake leads get through because Google's filters judge the ad click, not your form. Google is direct about it: "Google systems cannot track or control visitor actions after they click your ad"1. A bot or a person can fill in your form even after Google marked the click invalid and didn't bill it.

The fix sits on your site and in your goals. Google recommends reCAPTCHA and server-side validation on lead forms, a double opt-in by email where it fits, and campaigns that optimise for goals like "qualified lead" instead of every form submission. Sending real lead outcomes back from your CRM gives the bidding something better to aim at; Google Ads offline conversions covers the import side.

If the clicks are real and nothing converts, look at the landing page, the offer and tracking next; Google Ads clicks but no conversions will cover those.

Run the check with AdCtrl

AdCtrl runs the network and location reads in Claude or ChatGPT, so you see which setting bought the clicks before you change it in Google Ads13. The network read is in step 1 and the box above; for the location split, write:

In AdCtrl, write: only reads your accounts
For each Search campaign, split clicks, cost and conversions by location type, people in the area versus people interested in it, and by country, for the last 30 days.
AdCtrl report card with example data: clicks, cost and conversions for three fictional campaigns split by network, Google Search, Search partners and Display

A real AdCtrl report card for the network read in step 1, filled with example data: three made-up campaigns split by network. The campaigns, numbers and dates are all examples. The card shows Display under its API name, Content.

If the location split shows money going to people outside your market, switch the location option to Presence in the campaign's settings in Google Ads.

Read the Invalid clicks column and Billing credits in Google Ads: AdCtrl's report cards don't include the invalid-click metrics. Switching search partners off or the location option to Presence on a running campaign also happens in Google Ads, as do IP exclusions and the Click Quality Form.

The GAQL behind the check (for scripts and API users)

GAQL, the Google Ads Query Language, runs the same check from Google Ads scripts or the API. Skip this part if you work in the Google Ads interface. This query reads invalid clicks next to billed clicks for each campaign and network:

SELECT campaign.name, segments.ad_network_type, metrics.clicks, metrics.invalid_clicks, metrics.invalid_click_rate, metrics.cost_micros, metrics.conversions FROM campaign WHERE segments.date DURING LAST_30_DAYS

metrics.invalid_clicks is the "Number of clicks Google considers illegitimate and doesn't charge you for", and metrics.invalid_click_rate divides filtered clicks by filtered plus billed clicks3. segments.ad_network_type separates Google Search, Search partners and Display.

The location split comes from geographic_view, which "reports metrics at either actual physical location of the user or an area of interest", one row per country14:

SELECT campaign.name, geographic_view.location_type, geographic_view.country_criterion_id, metrics.clicks, metrics.cost_micros, metrics.conversions FROM geographic_view WHERE segments.date DURING LAST_30_DAYS

location_type returns LOCATION_OF_PRESENCE or AREA_OF_INTEREST, country_criterion_id is a country ID rather than a name, and cost_micros is in micros, where one million is one unit of your currency. For a longer window, swap DURING LAST_30_DAYS for a BETWEEN date range.

Frequently asked questions

What is a normal invalid click rate?

Google says "There is no industry standard for a normal invalid traffic rate". Vendors publish their own figures, such as TrafficGuard's claim that Search click fraud "typically" runs "between 14% and 22%"2, but that comes from the vendor's own analysis. Compare your rate with your own history instead.

Can I see which IP address clicked my ad?

Not from Google, which keeps details about specific clicks confidential. Your own web server logs hold the IP addresses, and the Click Quality Form asks for them.

Does invalid traffic use up my daily budget?

Invalid clicks caught before the end of the month are restored to the campaign budget, Google says, and its systems try to spend that money on valid traffic within your normal monthly limits. How that limit works is in Google Ads spending more than the budget.

Does Google remove conversions from invalid clicks?

Some. Google removes invalid conversions that came from invalid traffic, and that can happen after they appear in your account, so conversions can drop slightly over time.

Can AdCtrl detect click fraud?

No. It reads your campaigns by network and by location in Claude or ChatGPT, so you can see which setting bought the clicks. Spotting fraud and getting credits stay with Google's filters and the Click Quality Form.

Should I turn off search partners?

Only if the network split shows partner clicks costing money without converting. Search partners are on by default for Search campaigns, and Google doesn't show which partner site ran a Search ad, so the split is the only evidence you get5. If spend falls further than you wanted afterwards, see Google Ads not spending.

What to do next

Start with the third bucket, clicks your own settings bought, because it's the one you control: split clicks by network and by location type. Then add the Invalid clicks column and check Billing for "Invalid activity" credits. If something still looks wrong, gather GCLIDs and server logs and file a Click Quality Form request inside 60 days.

Learn more: the Google Ads audit checklist runs these settings checks across a whole account, and the negative keywords guide handles the search terms.

Try it: connect AdCtrl with the getting started steps and paste the network split prompt above. Try any plan free for 3 days, no card needed13.

Sources

  1. About invalid traffic, Google Ads Help. Accessed 13 November 2026.
  2. How to Protect Your Google Ads From Click Fraud, TrafficGuard, 21 September 2026. Accessed 13 November 2026.
  3. metrics (Google Ads API v26 field reference), Google for Developers, 11 November 2026. Accessed 13 November 2026.
  4. About advanced location options, Google Ads Help. Accessed 13 November 2026.
  5. About the Google Search Network, Google Ads Help. Accessed 13 November 2026.
  6. About Performance Max campaigns, Google Ads Help. Accessed 13 November 2026.
  7. About targeting for Display campaigns, Google Ads Help. Accessed 13 November 2026.
  8. Find or create placement reports for your Performance Max campaigns, Google Ads Help. Accessed 13 November 2026.
  9. Use segments in your tables, Google Ads Help. Accessed 13 November 2026.
  10. About auto-tagging, Google Ads Help. Accessed 13 November 2026.
  11. Exclude IP addresses, Google Ads Help. Accessed 13 November 2026.
  12. Power PPC Performance with Fraud-Free Traffic - ClickGuard, ClickGuard. Accessed 13 November 2026.
  13. AdCtrl docs: Google Ads and Meta Ads in Claude or ChatGPT, AdCtrl, 12 November 2026. Accessed 13 November 2026.
  14. geographic_view (Google Ads API v26 field reference), Google for Developers. Accessed 13 November 2026.

Keep going