Ad Fraud includes the deception of ad traffic, i.e. generating fake traffic, fake downloads or fake installs.
As opposed to the believe of many people that ad fraud just includes non-human (i.e., bot) traffic, the definition actually has to go a bit further than this. While non-human traffic is for sure an essential part of the definition of ad fraud, we have to consider that human traffic is as well part of the definition.
So when talking about ad fraud, we have to clearly differentiate between human and non-human traffic:
In this case the traffic on the affected websites is generated by click bots which automatically call up certain websites. This way the bots create artificial user profiles that are served with matching ads. So basically these ads are played out on a website without any human getting in touch with it. This way click bots make advertisers pay for fake customer contacts. Besides a website owner can increase the prices for advertisements on his website by generating fake traffic with click bots.
Within this type of ad fraud the fake traffic is generated by paid users or fraudulent publishers. Since this kind of ad fraud is caused by real people it is even harder to detect and to protect from than traffic generated by click bots. There are many different ways of human fake traffic, e.g. click farms. These usually pay low wage workers to click on ads or fill out forms, resulting in valueless impressions, clicks and conversions.
Ad Fraud in Mobile Advertising:
Apart from that ad fraud has become a serious threat to the mobile advertising business. There are not only click bots, but also bots that generate fake app downloads. These bots imitate real users with a certain smartphone and automatically install certain apps. This way they make advertisers pay for fake installs. The advertiser will, in most cases, think that a real user installed his app, although he actually did not acquire a new user.