Click fraud is a subset of invalid traffic, not another name for it. Invalid traffic is the whole category of clicks that should not be billed; click fraud is the part of it where somebody intended to cause the charge. The distinction matters because it determines what you can do about it — and most advertisers use the more dramatic term for a problem that is actually the duller one.
Invalid traffic: the umbrella
The advertising industry splits invalid traffic into two tiers, and the vocabulary is worth knowing because it is what Google's systems, ad verification vendors and publishers all use internally.
General invalid traffic (GIVT)
Traffic identifiable as non-human by routine means: known search-engine crawlers, monitoring bots, declared data-centre sources, obvious duplicate activity. It is filtered by published lists and simple rules, and there is nothing deceptive about most of it — a search crawler indexing your page is doing its job, not attacking you.
This is the bulk of invalid traffic by volume, and it is largely a solved problem you never see.
Sophisticated invalid traffic (SIVT)
Traffic actively built to look human: hijacked devices, browser automation dressed up as a real browser, proxy networks, click farms staffed by people paid per click. Detecting it requires behavioural analysis rather than a list, and no detection system catches all of it.
This is where the money is lost, and where "click fraud" as most people mean it lives.
Click fraud: the intentional part
Click fraud is invalid traffic generated deliberately to cause a charge or to profit from one. Two motives cover almost all of it:
- To drain a competitor's budget, so their ads stop showing and yours face less competition.
- To earn revenue share — a site or app in an ad network generating clicks on ads it hosts, because it gets paid for them.
Intent is the whole distinction, and it is also the reason click fraud is hard to prove. A click looks identical whether the finger belonged to a bored competitor, a paid click-farm worker, or a genuine customer who changed their mind.
The third category nobody names, which is bigger than both
There is a large class of clicks that are entirely valid, entirely human, entirely unintentional, and entirely worthless to you: accidental taps.
Someone playing a mobile game brushes a banner. Someone on a cramped screen misses the button they were aiming for. No fraud, no bot, no policy violation — and no possibility of a refund, because nothing invalid happened. It is simply an inventory quality problem, and for most advertisers running Display it costs more than fraud does. See mobile app placements wasting your Google Ads budget, which is the specific and highly fixable version of this.
Why the label changes what you should do. Invalid traffic is largely Google's problem and is credited automatically. Click fraud is worth reporting, with evidence. Accidental clicks are entirely yours to fix, cannot be claimed for, and are fixed with a settings change rather than a support ticket. Calling all three "click fraud" leads people to file claims for the one category that was never going to be refunded.
Working out which one you have
Three questions, in order:
- Where did it serve? If the spend is concentrated in Display, apps or Search Partners, suspect inventory quality before fraud. Check placements first.
- Does it behave like a machine? Sub-second sessions, zero scrolling, submissions faster than typing, activity through the night in a daytime business. That points to automation. The full list is in how to tell if your Google Ads clicks are real.
- Is it targeted? Activity concentrated on your single most expensive keyword, during business hours, from a handful of addresses, looks like intent rather than a bot net. That is the competitor case — see is a competitor clicking my Google Ads.
The response that covers all three
Each category has its own remedy, and keeping up with all of them is a permanent job. There is one change that reduces the damage from every category simultaneously: report real business outcomes back to Google Ads rather than form submissions.
A bot, a click farm and a mis-tapping ten-year-old have exactly one thing in common — none of them ever becomes a paying customer. Once your bidding is learning from customers rather than form fills, all three stop influencing where your budget goes, without you having to classify any of them correctly first. That is what offline conversion imports do, and it is the argument behind how Claudphic Ads works: the click ID travels with the lead, and marking it Won tells Google which clicks were worth buying.