What click farming is
Click fraud means clicks on your ads that you pay for and that will never become customers: bots, automated scripts, click farms and competitors burning your budget. In some verticals up to a third of the advertising budget goes to that traffic, and by definition none of it produces an enquiry. The mechanics are covered in what click fraud is.
Six signs to look for
| Sign | What it looks like in the data |
|---|---|
| Spend grows, enquiries do not | clicks and cost rise while conversions stay flat |
| A wave of very short sessions | paid sessions ending in seconds with no scroll |
| Repeating IPs and subnets | the same addresses clicking again and again across days |
| Clicks concentrated at odd hours | activity peaks when your audience is asleep |
| One placement behaves unlike the rest | a single source with a wildly different bounce profile |
| Enquiries with junk contact details | forms filled with nonsense, or the same details repeatedly |
How to estimate the share yourself
- •Take paid traffic only, over a period long enough to be representative,
- •Count sessions under a few seconds with no scroll and no interaction,
- •Break that share down by campaign and by placement — fraud is rarely evenly spread,
- •Multiply the junk share by your spend to get the figure in money.
The full calculation, with what the numbers usually come out at, is in how much ad budget bots consume. Telling an individual bot from a human is in how to tell a bot from a human in advertising.
Frequently asked questions
Is it competitors or bots?
Both. Part of it is automated bots and click farms clicking ads at scale; part is competitors, manually or by script, burning through your budget. For defence purposes the difference is small: both are filtered on behaviour, fingerprint and repeating IPs.
Do the ad platforms not filter bots themselves?
They filter part of the obviously invalid traffic and credit it back. But a significant share of bots and junk clicks gets through, particularly on display and in-app inventory. Your own antifraud catches what the platform’s filtering misses.
How quickly can we tell?
A single evening with your analytics is usually enough for a first estimate: look at sessions from paid traffic with near-zero duration, no scroll and no interaction, and compare that share across campaigns and placements.