Invalid Clicks in Google Ads: What They Are, Where to See Them and How to Get Your Money Back

You pay for a click, not for a person. Between the two stands Google’s filter, which decides which click was genuine and which was invalid. Most of this work is invisible: filtered clicks simply never reach your bill. But Google shows the result in dedicated columns, credits you for what slipped through and accepts requests for investigation.

Below is a practical breakdown: what Google treats as invalid clicks in Google Ads, where you can see them in your account, how automatic credits work, how to file a request and what to do if your own data shows far more suspicious traffic than was credited.

What Google counts as invalid clicks

Google defines invalid clicks as clicks that do not reflect genuine user interest. They include:

• intentional manual clicks — meant to drive up an advertiser’s costs or the earnings of a site owner showing the ads. What this looks like when competitors do it is covered in competitor click fraud;
• automated clicks — robots, scripts, auto-clickers, malware on infected devices;
• extra clicks with no value — for example, the second click of a double-click, which Google does not treat as separate interest;
• inflation on partner placements — sites and apps in the Display Network that provoke accidental or fake clicks;
• click farms — real devices and people clicking for money; more in clicks from click farms.

Important: “invalid” is not the same as “did not buy”. A real person who looked at the page and left is a valid click, even if they brought you nothing. Google judges the intent and source of a click, not its outcome.

The broader classification this terminology belongs to — general and sophisticated invalid traffic — is covered in invalid traffic, and the main schemes used to defraud advertisers in the overview types of ad fraud.

How Google handles invalid clicks

Protection works in three layers.

Real-time filter

Every click is checked the moment it happens. If the system deems it invalid, it is not charged and does not appear in the Clicks column or in CTR and CPC calculations. That is why most advertisers never see the bulk of the filtered junk.

After-the-fact review

Some fraud is visible only at scale: recurring patterns, linked placements, series of clicks across several accounts. If Google finds such clicks later, after you were charged, it automatically credits your account.

Investigation on request

If you believe Google missed something, you can request an investigation. Specialists review the activity and decide whether a credit is due.

Where to see invalid clicks in your account

Columns in the campaigns report

1. Open Campaigns.
2. Click the columns icon → Modify columns.
3. In the Performance group, tick Invalid clicks and Invalid click rate.
4. Save — as a separate column set if you like.

Invalid clicks is the number of clicks Google filtered out and did not bill. Invalid click rate is their percentage of valid plus invalid clicks. The columns are available at the account and campaign level, but not for keywords or ads, so you cannot use them to pinpoint which keyword is being clicked.

How to read these numbers

• Compare campaigns with each other. If one has a rate several times higher than its neighbors, see how it differs: Display Network, search partners, geo, bids.
• Watch the trend by day. Break the report down by day or week. A spike in invalid clicks often coincides with a spike in suspicious traffic that the filter only partly caught.
• Do not be scared by the number itself. A high rate means the filter is working and you are not paying for that junk. The danger is what got past the filter — and that is not in this column.

Credits for invalid activity

Credits for clicks found after billing appear under Billing → Transactions (or Summary and Documents, depending on your interface) as Invalid activity lines with a date and amount. A credit usually relates to a past period and is not tied to a specific campaign in the interface.

Where invalid clicks tend to be higher

The risk depends heavily on where the ads are shown. A rough guide by placement type:

So it pays to analyze campaigns separately by network. If one campaign mixes Search and partners, segment the report by network — the difference in invalid rate and conversions is often obvious right away.

Signs the filter may have missed

The columns only show what Google caught. What it missed shows up in your own stats:

• CTR is rising, conversions are not. Especially if the rise is in one campaign or keyword.
• Spend runs out early in the morning. The daily budget is gone in the first hours, and leads, if any, come later from entirely different visits.
• Repeats from the same addresses and subnets in server logs.
• Visitors without JavaScript — the server sees them, analytics does not.
• Leads with junk data: non-existent phone numbers, identical names, a form filled in within a second. What to do about them is covered in bots submitting forms.

If you see several of these signs, you have material for a request.

How to request an invalid click investigation

Before filing, Google advises checking whether the account itself explains the change: bid changes, new keywords, search partners switched on, seasonality. If there is no explanation:

1. Define the period and campaigns. Specific dates and campaigns where the picture is unusual. Google only reviews recent activity, so do not wait — check the current time limit in the help center.
2. Collect data on your side. Click times, visitor addresses and networks, gclid, landing pages, on-site behavior. Server logs and a visit log are worth more than analytics screenshots.
3. Match clicks to gclid. The click ID is the only reliable bridge between your logs and Google’s data. How to store it is covered in click ID and SubID.
4. Describe the pattern with facts. Not “we are being click-bombed” but “on the 12th–14th, campaign X got 400 clicks from 30 addresses in one subnet, 380 of them without running JavaScript and with under a second on the page”.
5. Fill in the form to request an invalid click investigation in the Google Ads help center, giving the account ID, campaigns and dates and attaching your data.

The decision is Google’s. If it confirms the issue, it credits your account; if not, it explains that the clicks look valid. Google does not disclose the details of its investigation so as not to teach fraudsters how to get past the filter.

Why there is more suspicious traffic than was credited

This is a common situation, and there are several reasons for it.

• Some of it has already been filtered. Your tool sees the visits but does not know Google did not charge for them. Compare against invalid clicks, not total clicks.
• Different definitions. A person on a VPN, a returning visitor, a click followed by a slow exit — a third-party filter may call it “suspicious”, Google calls it a valid click.
• Not everyone sees bots without a browser. If a visit did not run JavaScript, analytics does not see it, but the server does. More on why numbers from different systems diverge in bots in Google Analytics 4.
• Sophisticated fraud gets through. Bots on residential addresses and click farms really do slip through sometimes. This is where your data becomes the basis for a request.

What to do if not everything was credited

A request is not your only tool. Often it is cheaper to shut off the source than to fight for every credit.

Exclude the source

• IP exclusions. In campaign settings you can list up to 500 addresses and ranges. This works against specific persistent visitors and poorly against botnets with rotating addresses. A comparison of approaches is in JS check or IP blocking.
• Search partners. If the invalid rate is higher in campaigns with partners, turn them off and compare the results.
• Display Network placements. The “Where ads showed” report lists sites and apps. Exclude placements with clicks but no engagement or conversions, especially mobile apps with accidental taps.
• Geo and time. If the junk comes from specific regions or at night, narrow your targeting and ad schedule.

Do not train the algorithm on bots

The most expensive consequence of invalid clicks is not the spend itself but the learning. If bot leads get counted as conversions, Smart Bidding starts looking for similar traffic. Send only confirmed conversions to Google Ads — see sending conversions to ad platforms.

Filter on your side

Google does not refund everything and is not obliged to let only people onto your site. Your own visit checks do two jobs: junk never reaches your form or offer, and you get a log for your requests. ArtisanClo filters out bots, data center traffic, VPNs and proxies, and shows a reason code for every click in its log — handy to attach to a request to Google. Its built-in tracker counts spend, conversions, profit and ROI per campaign, so you can see how much money actually went on junk traffic. More on the features page.

Redo the math

If invalid traffic is consistently high and credits are low, factor it into your cost per click. How to estimate the losses in money is covered in how much budget is lost to bots. Sometimes it is more honest to admit that a campaign on a given source is unprofitable than to fight endlessly for credits.

What not to do

• Do not click your own ads to “check” them — those are invalid clicks too, and they skew your stats.
• Do not file a request without data. A screenshot of “lots of clicks, few leads” is not evidence.
• Do not send raw exports from a third-party tool without reconciling them against what Google filtered: half of the entries will turn out to be already accounted for.
• Do not exclude addresses in bulk at random — it is easy to cut off real buyers, such as corporate networks and mobile carriers, where many people share one address.

If your account is dealing not with fraud but with restrictions, it helps to know the Google Ads advertising policies.

Summary

Google Ads filters out most invalid clicks on its own: those caught in real time are not charged, those found later are credited back. You can see the result in the Invalid clicks and Invalid click rate columns and in the billing transactions. If you are sure something was missed, collect logs with gclid, describe the pattern with facts and request an investigation. And so as not to depend on Google’s decision alone, exclude junk sources, do not send bot conversions to the ad account and check visits on your side.