How to track leads from Google Ads
See which Google Ads campaigns and searches bring enquiries and paying customers.
The fear is simple: a competitor or a bot clicks your ads and burns the budget. Google filters clicks it judges invalid before you pay for them, and much of what looks like fraud is a setting you can change. How to tell the two apart, and what to do about each.
Updated 29 October 2026, about 8 minutes to read
The short answer
Google filters clicks it judges invalid, such as bots, accidental clicks and clicks meant to run up your costs, and does not charge you for them. Invalid clicks it finds after billing come back as credits. Add the Invalid clicks column to see what it caught. Before blaming fraud, check your location option, search partners and search terms. For a small account, a paid click-fraud tool is rarely worth its fee.
Before you pay for anything or accuse anyone, check your own account:
If the clicks still look wrong after all six, gather the evidence below and ask Google to investigate. You can do that for traffic in the past 60 days.
Google defines invalid traffic as clicks and interactions that do not represent genuine interest in your business. Its examples:
So the competitor who clicks your ads is on Google's own list. Google's ad traffic quality team uses automated filters that scan traffic continuously, plus manual reviews and investigations, to catch it.
Invalid clicks found before your billing cycle ends are removed from your costs, so you are never charged for them and there is nothing to refund. They still show in the Invalid clicks column, which is why that column exists. Google's own reading is that a rise in invalid clicks shows its filters protecting your budget.
Invalid clicks found after you were billed come back as a credit, as a billing adjustment for their value. You find these in Billing, under Summary, in the Adjustments of each month, labelled Invalid activity, and they appear on your invoices. The Invalid Activity Credit Report in Report Editor shows the adjusted clicks, cost and conversion rate.
One quirk: credits for past invalid activity appear on the invoice but do not change the old campaign data in your account. If a report from last month looks worse than the bill, that can be why.
Many patterns that look like fraud have a simpler cause. Check these before anything else:
Each of these costs money and none of them is fraud. All of them are fixed in your own settings, for free.
After the settings are right, these patterns deserve a closer look:
Your website's server logs or analytics can show more than Google Ads does: the IP addresses, the devices and the Google click ID on each visit, which Google adds to every ad click when auto-tagging is on. Save them for the dates in question. They are what Google asks for if you request an investigation.
Google takes investigation requests through its Click Quality Form, for traffic from the past 60 days. Its specialists review the traffic and email you the findings. Before you send it, collect:
Check the Invalid clicks column for the same dates first. Google points out that third-party tools may flag traffic it has already filtered, and you were not charged for that. A request about clicks Google already removed will find nothing to credit.
Write the request plainly: what you saw, when, and why it does not look like real customers. Do not name or accuse a competitor in public while you wait. If you are right, the evidence is what counts. If you are wrong, you have made an enemy for nothing.
If you know the address a stream of bad clicks comes from, such as an office network, you can block it. Go to Campaigns, then Settings, select the campaign, open Additional settings and expand IP exclusions. Google allows up to 500 IP addresses per campaign, plus an account-level list that it merges with the campaign one. You can block a range by replacing the last part of the address with an asterisk.
IP exclusions are not available for every campaign type. Google says they are not available for video, hotel, App, Performance Max and Smart Display campaigns.
Know the limit before you rely on it. A phone on mobile data changes address easily, and many people can share one address, such as everyone in an office or on a cafe's Wi-Fi. Blocking addresses works against a fixed office. It does little against someone on a phone, and a wide block can shut out real customers.
Click-fraud tools watch the visits from your ads, flag suspicious patterns and add IP exclusions to your campaigns for you. They charge a monthly fee, and what decides whether it pays is simple: the value of the invalid clicks it stops that Google would not have filtered or credited anyway.
That is the catch. The tool cannot see Google's own filtering, so part of what it reports as fraud may be clicks you were never charged for. Its main weapon, the IP exclusion, is capped at 500 addresses per campaign and weak against phones.
A tool can make sense for a large account in a fiercely contested local trade, such as emergency locksmiths or personal injury law firms, where you have evidence of repeated invalid clicks that Google did not credit. For a small business spending a modest amount on search ads, the settings in this guide, a weekly look at the search terms report and an investigation request when the evidence is clear will usually do more for less.
Fraud worries grow when nobody looks at the account for a month and then the bill arrives. A short check every week, on the same day, catches problems while they are small:
Write the numbers in a sheet, one row per week. After a couple of months you will know what a normal week looks like for your account, and a strange one will stand out on its own. If nobody in the business has time for this, it is a sign the account needs a person to own it, whether that is you, a colleague or a paid manager. The weekly figures are also easy to send to one inbox automatically, as part of a weekly summary from your own automation.
A locksmith covering one city sees clicks double in a week with no extra calls, and suspects a rival. He works through the checks:
He did not buy a tool, did not ring the competitor and did not pause his ads. Most of the waste was in his settings, and the rest is now with Google. A trade business depends on its phone, so he also counts calls from the ads: see call tracking for a small business.
The best defence is knowing which clicks became customers. Save the click ID and the campaign with every enquiry, as how to track leads from Google Ads explains, and the fake clicks stop mattering as much because you can see the real ones. GGSites builds landing pages for ads that keep the click ID with each enquiry, and we do not run ad accounts, so we sell nothing that depends on your fear of fraud. If you are weighing Google's pay-per-lead option instead, see Local Services Ads or Google Ads, and for the cost of a page, our prices page.
They can click, but Google treats clicks meant to run up your costs as invalid traffic and works to filter them. Filtered clicks are not charged, and invalid clicks found after billing are credited. Add the Invalid clicks column to see what Google caught, and gather server logs if you see a pattern it seems to have missed.
Google does not charge for invalid clicks it filters before your billing cycle ends, so no refund is needed. When it finds invalid clicks after billing, it credits their value as an Invalid activity adjustment on a later invoice. You can see these credits in Billing, under Summary, in each month's adjustments.
Use Google's Click Quality Form for traffic from the past 60 days. Include your customer ID, the exact dates, the campaigns, ad groups and keywords involved, and server logs with IP addresses, devices and click IDs. Check the Invalid clicks column first, since clicks Google already filtered were never charged.
Only when you know a fixed address sends bad clicks, such as an office network. Google allows up to 500 excluded addresses per campaign, plus an account-level list, but not on every campaign type. Phones on mobile data change addresses easily and many people share one, so wide blocks can shut out real customers.
Rarely for a small account. They cannot see what Google already filtered, and Google notes that third-party tools may flag traffic it removed. Their main fix, IP exclusions, has a 500-address limit and does little against phones. Check your location option, search partners and search terms first, and compare any tool's fee with what it could recover.
Often because of the default location option, which reaches people in your area, regularly in it, or showing interest in it. Someone elsewhere searching about your town counts. Switch to Presence in the campaign's location options, exclude places you do not serve and check matched locations in the Locations report a week later.