Fraud in ads.

What fraud traffic is

Fraud traffic is visits that never actually happened: they are created not by buyers but by bots and unscrupulous placements in order to get paid for impressions and clicks. For the advertiser it means wasted budget and distorted statistics.

The consequences are more dangerous than the spend itself. Reports show activity that does not exist, the cost per click looks acceptable, yet no leads appear. Decisions are then made on corrupted figures: channels that do not really work get scaled up.

Where it comes from

The sources vary. Some traffic comes from junk placements in partner networks, where an ad impression is paid for regardless of who saw it. Some is the result of click fraud: automated clicks on ads, sometimes from competitors.

A separate category is incentivised traffic: real people who were paid to click through. Formally this is not a bot and the behaviour looks human, but such a visitor will never become a customer.

All the sources have one thing in common: they can fake a visit but cannot fake interest. That is exactly how they get caught.

How to spot it in reports

The first sign is an abnormal bounce rate on a specific slice. If bounces on one ad type or one placement are sharply higher than on the rest for the same offer, that is a reason to investigate, not to blame an unlucky creative.

Next, look at the details of the visit: time on page of a few seconds, no scrolling, identical page depth, spikes in traffic during the night, traffic from regions where you do not operate. Individually each sign can be explained; together they form a picture.

It is useful to compare behaviour across sources. Normal traffic is always varied: some people read, some leave immediately. Suspicious traffic is suspiciously uniform.

What to fix in the settings

Most fraud is cut off by campaign settings, not by specialist services.

• Switch off placements with abnormal bounce rates — in the networks this is done manually and from a list that should be reviewed regularly.
• Narrow geography and ad schedule to where you genuinely have customers.
• Build negative keyword lists so you do not pay for impressions on queries unrelated to a sale.
• Separate search and networks into different campaigns: blended statistics hide the problem.

The quality of the ad itself matters too: the more precisely it describes the offer, the fewer accidental clicks, and the more clearly anomalies stand out against them.

Protection tools

Ad platforms filter some invalid clicks on their side and do not charge for them. This is baseline protection, and it does not catch everything.

On top of that come monitoring systems that track anomalies in traffic and block suspicious sources before they affect campaign results. Connecting them makes sense when the ad budget is large enough for the losses to exceed the cost of the service.

And the mandatory minimum for everyone is goals set up in analytics. As long as you judge a campaign by clicks, it is impossible to tell live traffic from fake; as soon as you count leads, everything falls into place by itself.

What not to expect from fighting fraud

You cannot remove it completely — some share of invalid visits is always there. The task is not to reach zero but to stop it influencing decisions.

And do not blame fraud for every poor result. More often the cause is simpler: settings that are too broad, a weak offer or a landing page that does not continue the ad. Check for fraud after you have ruled out those three things.

Frequently asked questions

Will money be refunded for click fraud?

Platforms themselves do not charge for some invalid clicks, but retrospective compensation is rare. It is better to count on prevention.

Can a competitor click on my ads?

It happens, but the scale is usually exaggerated. The sign is a series of clicks from similar addresses without a single action on the site; this is visible in reports.

How to tell fraud from simply the wrong audience?

By the uniformity of behaviour. The wrong audience behaves in different ways but does not buy; fraud behaves identically and does not interact with the page at all.

Should ad networks be switched off entirely?

Not necessarily. Networks bring cheap reach but require manual cleaning of placements — if there is nobody to do it, switching them off really does turn out to be more profitable.

What to do next

Open the placements report for the last month and sort it by bounce rate. The top rows are your first list for switching off, and it almost always saves money without losing leads.

We clean up campaigns and monitor traffic quality — that is advertising setup. Next, you may find useful our pieces on what to do if advertising brings no clients and why you have a website but no leads.