“Half the money I spend on advertising is wasted; the trouble is I don’t know which half” – John Wanamaker

With the recent news of Facebook overestimating video views by up to 80%, we at addmustard have been wondering for many years how much activity is actually over-reported and overpaid by advertisers. In fact, it’s been confirmed that globally over $7.2 BILLION has been spent on fraudulent advertising activity this year alone – that’s 9% of all video impressions which are non-existent.

With this in mind, let’s take a look at how ad fraud happens, its implications for advertising, and how we can combat it.

 

How does it happen?

The main characteristics of ad fraud are as follows:

Fake sites

Creating a network of fake sites with little or no content which is sold into ad exchanges. Often they will slip through the net of smaller exchanges and accrue money until they are discovered. Once they are found out, it’s likely they will set up under another name and repeat the process.

No visibility

Publishers will often use tactics such as pixel stuffing where ads are crammed into a 1-pixel by 1-pixel unit at the bottom of the page, or ad stacking where multiple ads are layered on top of each other in a single unit. Both leading to high impression volume but zero visibility.

Domain masking

Hiding the URL of the website where the ads will actually appear allows publishers to lie about the content, and charge high CPM rates for ‘targeted content’ that’s not even there.

Non-human traffic

Also known as NHT, is where ad fraud bots are designed to drive fake ad impressions and simulate human traffic to site.

 

Why does it happen?

Most advertisers do not realise that they are paying for “bad” traffic owing to the low click costs and the large traffic volumes that fraudulent activity provides. And even though the traffic is not real, companies will often unknowingly choose fraudulent inventory over verified inventory to hit business targets.

There are also many stakeholders currently benefiting from fraudulent traffic, so it is doubtful whether fraud will ever be eradicated. Companies like Click Guardian who offer anti-fraud solutions and solely function as a result of fraudulent activity, do not have the incentive to eradicate this endemic industry problem.

 

How do we stop it?

The recent Facebook news calls for all platforms to allow third party tracking verification. At addmustard, we too noticed some time ago that interface data such as traffic and impressions can vary significantly with the third party tracking tools that we implement. But it is often difficult to reconcile the differences in impression reporting. Two years ago, we disputed the reports of the ad serving platform, challenged their data and saved the client hundreds of thousands of pounds.

This means that monitoring metrics like bounce rate, time on site and conversion rate is imperative to pinpointing the culprits. It also means that advertisers will have to consider investing in high CPMs on well-known exchanges in order to achieve the quality of traffic that they need. Although, it has been reported that ad fraud can happen, even on reputable exchanges, by up to 10%. At addmustard, we believe it can be significantly higher than this.

Finally, the automatic process of approving new sites needs to be reviewed. A more human approach will help stop fraudulent publishers being accepted onto ad exchanges, as well as allowing for publishers to be more accurately categorised based on their content, which will result in advertisers being able to choose the most relevant sites for their audience.

 

Have you been affected by ad fraud? Let us know your story by tweeting us @add_mustard.