Advertisers expect that all online content is viewed by human audiences – genuine consumers who have the potential to buy their products or services. That isn’t always the case. The digital advertising industry is plagued by an increasing amount of fraudulently generated non-human traffic (“traffic fraud”) which is then paid for by advertisers.
A report from the Association of National Advertisers and WhiteOps in December last year indicated that bot-generated traffic accounts for as much as 11% of display ads purchased, and 23% of video ads.
The headline figure from the report is that traffic fraud will cost the digital advertising industry $6 billion this year.
How does it work?
The price of an advert is usually dependent on the number of impressions. If an intermediary is able to sell inventory and subsequently generate false non-human traffic to increase overall impression count, the advert will be sold at a falsely inflated price. This is primarily an issue for the advertiser, as ultimate buyer of the ad, since the advertiser will be paying amounts which do not correlate to the actual number of human impressions.
From a technical perspective, there are various ways the numbers of ad impressions can be distorted but to focus on just two:
Page Fraud – this is where there are multiple adverts on a page, some of which are ‘viewable’, others are hidden behind other adverts or compressed to a minute size. The adverts are technically displayed and an ad impression registered but they are not actually viewed by the end user; and
Bot Fraud – this is a more pervasive type of ad fraud which can be perpetrated by a publisher or an ad network and needs real-time protection to prevent it from happening. Bots find their way on to unsuspecting user’s machines and mimic human behaviour by interacting with adverts behind-the-scenes and creating fake ad impressions.
How is the industry responding?
The IAB is clearly aware of this issue and the fact it could undermine the credibility of online advertising and curtail the rise of online ad spend. In response, it set up a working group to look into ad fraud and come up with recommendations and best practice, primarily for agencies to follow. In September 2014, the working group established the following set of principles which the IAB has encouraged all players in the ad delivery ecosystem to sign up to.
Following on from this, in December of last year, the UK’s Joint Industry Committee for Web Standards (JICWEBS) – an independent body that defines best practice and standards for online ad trading – held the first cross-industry technical group meeting designed to tackle online ad fraud. The body includes publishers such as the Financial Times and Trinity Mirror and advertisers including Reckitt Benckiser.
JICWEBS is aiming to publish anti-fraud good practice principles for the UK market in July 2015 and, beyond that, to announce the first companies to be accredited for meeting industry-agreed standards to reduce the risk of fraudulent ads being served.
JICWEBS is set to make an interim announcement in February to update the industry on the progress in defining the different types of online ad fraud and the appropriate guidelines. Adtekr will keep you updated.
What are companies doing to combat ad fraud?
Whilst the true scale of the problem is difficult to quantify across all the digital space, there are companies which are developing services to combat ad fraud:
Integral Ad Science, a leading player in this area, has recently announced the opening of an Anti-Fraud Lab to conduct research and development that will help advance anti-fraud solutions in the advertising industry. Integral’s services are currently used by key ad tech companies as a means of removing fraud from the ad delivery process and it is the increasing take-up of services like these which indicate that ad fraud is a real concern and needs to be swiftly addressed by all in the ad delivery chain.
comScore has recently developed “Trust Profiles” which enable ad buyers to assess the quality of ad space and to judge which sites have the most non-human or “bot” traffic running through their properties, as well as which ones offer the most “viewable” ad impressions. This “Trust Profile” can then be plugged into programmatic ad buying software to ensure those systems purchase only “high-quality” inventory. Since last month, a number of large ad tech companies have been trialling the trust profiles, these include MediaMath, Turn, The Trade Desk, Rubicon Project and Eyereturn.
Forensiq – another company which works for ad tech companies to detect ad fraud – recently carried out a study to paint a picture of the scale of the fraud and how it operates in practice.
Forensiq focused on a particular type of ad traffic fraud whereby malware is installed on a user’s computer which then loads websites of its choosing in the background. This also loads the accompanying ads and the malware generates click-throughs and impressions, whilst the ads themselves are never actually visible to the user. The malware is designed to replicate user activity and is therefore difficult to detect when publishers report impressions and clicks to agencies and advertisers.
Forensiq’s video shows how a single ‘infected’ machine can generate 10,000 fraudulent ad impressions over the course of 24 hours:
Is it working?
The increasing problem of traffic fraud has finally reached critical mass and we are now seeing a concerted effort from the major players in the adtech market together with industry bodies such as the IAB to find a solution. It is still early days and it remains to be seen as to who will win the development arms race, fraud detection technology or bot creators. These first steps should however be welcomed by advertisers, since a united front across the industry has the highest likelihood of success in the war against the bots.