Evidence for advertisers

Getting the clicks you pay for takes your own measurement

Written and reviewed by Abisola Tanzako, 18 August 2026. Figures from Imperva, Juniper Research, the ANA, Google Ads Help and a ClickPatrol case study of 21 UK agencies.

Getting the clicks you pay for is not automatic. Dashboards count clicks, leads and conversions, but part of that traffic is bots, click farms and click fraud that land as invalid traffic. Google, Meta and TikTok supply the inventory, measure the result and send the invoice. Your number is the share you measure independently.

The case, on video

A ClickPatrol partner walks through what invalid clicks are, why they pass as paid traffic, and why a platform invoice is not an independent count.

What the reports measured

These are dated, named sources. They are averages from other people's data. They explain why a clean-looking account can still leak. They do not replace a reading of your own clicks.

Figure	Finding	Source
51%	of internet traffic was automated in 2024	Imperva, Bad Bot Report 2025
22%	of online ad spend lost to ad fraud in 2023	Juniper Research, Sep 2023
36¢	of each programmatic dollar reached a consumer	ANA, programmatic transparency study
6%+	media savings in one ClickPatrol agency group	21 UK agencies case study

Three facts about paid clicks

These three points sit underneath every dashboard that looks healthy. None of them needs a ClickPatrol login to be true. They come from the reports above and from the platforms' own help pages.

A large share of ad traffic is not human

In 2024, 51% of all internet traffic was automated. Bots are now the majority online. Juniper Research estimated that 22% of online ad spend was lost to ad fraud in 2023.

The party measuring your traffic also profits from it

Google, Meta and TikTok supply the ad space, measure the results, send the invoice and decide for themselves what counts as invalid. Independent of any single platform, the ANA found that only about 36 cents of every programmatic dollar actually reaches a consumer.

What you do not verify yourself, you cannot trust

You can only steer a budget once you measure your own traffic independently. Platform filters shrink the problem. They do not make it disappear, and they do not publish your number.

Signs in your own account

Each of these is something you can check today, without a vendor report. Two or more at once is reason to measure the traffic mix.

Clicks rise, conversions do not
Volume jumps. Sales, calls or form fills stay flat. The extra clicks never had intent.
The daily budget runs out earlier
The same cap used to last the day. Now it burns through by mid-afternoon.
Sessions last 0 to 5 seconds
People who meant to buy do not bounce that fast, again and again.
Clicks arrive from outside your geo
You target one set of regions. Reports show another.
The same IPs keep coming back
Repeat patterns on the same addresses or devices are a tell, not a coincidence.
Conversion rate slips while spend holds
Your conversion rate falls and the budget does not. The mix changed, not the offer.

The industry numbers

The figures in the table are the headline. The detail below is why those headlines keep showing up in paid search, not only in cheap display inventory.

Bots are now the majority of the web

In 2024, for the first time in a decade, more than half of all internet traffic was automated (51%), and "bad bots" made up 37%. That is the traffic that drains ad budgets.

Ad fraud is a measured, large-scale problem

Juniper Research estimates that 22% of online ad spend was lost to fraud in 2023 (about $84 billion, rising toward $172 billion by 2028), with search set to become the largest fraud channel over the next five years. Those are typically the clicks you pay the most for.

Most programmatic money never reaches a person

The ANA found that only about 36 cents of every programmatic dollar reaches a consumer. Made-for-advertising sites alone absorb 21% of impressions and 15% of spend, and the ANA's 2025 benchmark still puts annual waste at $26.8 billion.

Even protected channels leak

Industry anti-fraud standards prevented an estimated $10.8 billion in losses in US display and video advertising in 2023. Roughly $979 million still slipped through. Protection shrinks the problem. It does not make it disappear.

Why click fraud keeps happening

Click fraud is not a glitch in one campaign. It persists because someone is paid for it, bidding systems treat it as success, and the firm that invoices you also writes the invalid-traffic rules.

Someone profits from every invalid click

Click farms, automated bots and publishers that earn on ad clicks all have a direct financial motive to generate them.

Your own bidding can amplify it

If bots look like clicks and conversions, Smart Bidding scales the bad sources instead of cutting them. You end up optimising on contaminated data.

The scorekeeper is also a player

Google filters invalid clicks, decides what qualifies and issues credit rather than a refund. An entire apparatus (MRC and IAB detection standards, Google's own Ad Traffic Quality team) exists to police this, which tells you how structural the problem is. Even platform metrics are not beyond question: in the US, advertisers are pursuing Meta in court over reach figures they say were inflated. The point is not that one platform is bad. It is that the party keeping score also benefits from it.

Measure the share on your campaigns

Every figure above is an average from someone else's data. Yours could sit well below or well above it. Click fraud protection is how ClickPatrol does that measurement: it scores clicks on 800+ signals, then shows which share looks human and which share does not. In a ClickPatrol case study of 21 UK agencies, the group recorded more than 6% media savings after fake traffic came out of the mix. That is one cluster, not a promise for every small business account.

The free 7-day trial shows the share on your Google, Meta and TikTok campaigns. Setup for Google Ads takes under a minute. A card is required for a one-time €0.01 verification charge. You are not billed for the trial itself.

Start My Free 7-Day Trial Book a Demo

Sources

Imperva (Thales), Bad Bot Report 2025: imperva.com
Juniper Research, Quantifying the cost of ad fraud (Sep 2023), via PR Newswire: prnewswire.com
Juniper Research, via MediaPost (Sep 2023): mediapost.com
Google Ads Help, Invalid clicks: Definition: support.google.com
Google Ads Help, About invalid traffic: support.google.com
ANA, Programmatic Media Supply Chain Transparency Study (2023): pr-2023-06-programmaticstudy and media-programmatic-transparency
ANA, Q2 2025 Programmatic Transparency Benchmark (Aug 2025): pr-2025-08-programmatictrans
TAG, 4A's, ANA and IAB, 2024 US Ad Fraud Savings Report (Oct 2024), via PR Newswire: prnewswire.com
CNBC, U.S. Supreme Court rebuffs Meta bid to avoid advertisers' lawsuit (Jan 2025): cnbc.com; case DZ Reserve v. Meta Platforms, Cohen Milstein: cohenmilstein.com
ClickPatrol, 21 UK agencies case study: more than 6% media savings

All sources are independent research firms, industry associations, the platforms' own documentation, court records, or ClickPatrol first-party case evidence. No other fraud-detection vendors are cited.

Questions advertisers ask

Google rarely pays a cash refund. It issues credits or adjustments in Google Ads billing when it agrees a click was invalid. ClickPatrol reports invalid traffic as a Google Certified Click Tracker. Manual review credits take roughly 4 to 8 weeks, and contested clicks have a low approval rate. The distinction is on ClickPatrol's Google Ads refunds page.

Google defines invalid clicks as interactions it judges accidental or malicious, including bots and repeated clicking. Google decides what qualifies and how much credit to issue. ClickPatrol scores the same click on 800+ device, network and behaviour signals, so you can see a share that Google's own filter left in the report.

The same company sells the click, measures the click and decides whether the click was valid. Filters exist, and they catch a share. They are not an independent audit. Advertisers who want a second reading use a Google Certified Click Tracker such as ClickPatrol, then compare that share with the platform report.

Under a minute for Google Ads. Create the account, connect the ad account, and ClickPatrol starts classifying clicks. There is no site code required for that Google Ads path. Meta and other channels use a site tag or GTM so ClickPatrol can build a suspicious-traffic audience you then exclude.

Yes. The free 7-day trial shows how much invalid traffic is hitting your ads and what that costs. A card is required for a one-time €0.01 verification charge. You are not billed for the trial itself. Start my free 7-day trial from the button on this page, or book a demo if you want a walkthrough first.