Ad Fraud: we're looking in the wrong place
- Advertisers’ money lost to ad fraud every year is totalling $7.2billion. That’s the latest number we’re hearing.
- We are looking in the wrong place – at the website level. We need to be looking at the subdomain and user level where the majority of fraud occurs.
- Moving away from the basic mentality of just monitoring fraud to a granular auditing process in order to weed out all the inefficiencies.
- Shifting from post-bid to pre-bid blocking, whereby an automated system can pre-emptively stop the fraud BEFORE the crime is ever committed.
To put this into context, the UK Home Office feels the need to step up its efforts to crack down on “cyber-fraud”, such as credit card scams. But compared with ad traffic fraud, this problem is peanuts – just £30m.A recent UK survey, carried out by ad verification company Adloox, showed that less than 17pc of brands are aware of ad fraud, and less than 10pc knew if they had a fraud prevention service in place on their ad campaigns.We are dealing with organised, sophisticated criminals and highly intelligent, machine-learning robots mirroring human behaviour and engagement online.
What makes this $7.2bn figure even more concerning is that it is only based on surface-level fraud – looking at website names and scoring them on how viewable they are, and how “real” they look.
This approach may help to block brands’ ads appearing on inappropriate content such as pornography or terrorism websites, yet how many of the billion-plus live websites globally today are called www.fraud.com?
Fraudsters are hiding much deeper down the rabbit hole, and as a collective industry we are not doing enough to stop it. We are looking in completely the wrong place.
The proof is that the current accepted benchmark for non-human traffic (NHT) on an ad campaign is anywhere between 10 and 15pc. Adloox has reported ad traffic suppliers with more than 80pc hidden ad fraud across areas of their network.
There are fake sites everywhere, with no humans ever landing on the website. It is of great concern that in the past 12 months we have seen the majority of the top 100 Nasdaq advertisers appearing on these fraud websites. So can we combat this? Or do we need to accept that 10 to 15pc wastage is always going to be the norm? And is 0pc fraud therefore an unrealistic pipe dream?
Vendors offering only viewability do not help matters. One such company recently reported the “good news” that UK viewability is up to 54pc.
First, let’s be frank. Saying that nearly 50pc of a brand’s advertising is never seen is not a reason to pop the champagne. Worse still is that this number does not take into account the fraud, and fake (bot) views of the ads. So the number is inaccurate.
“Robots don’t buy Kit Kats” was one famous media agency quip from last year. But when agencies say they are hesitant to cut out bot fraud completely from their clients’ campaigns (“if we do, we’ll look like our performance conversions have gone down” being the justification), then there is a real issue of quality and transparency.
The solution lies in the detection and eradication of fraud at a much deeper, user level. By using a more robust, user-centric solution such as Adloox, which is able to spot patterns and symptoms of fraud in the traffic – 90pc of a publisher’s audience spending just two to four seconds on a web page, say – and being able to pre-emptively block this inefficiency before the bid (or ad impression) is even served, this can result in both savings and rapid uplift in campaign performance for both agencies and their advertisers.
NHT traffic goes down to under 0.2pct – no longer 10 to 15pc. Not only that, but viewability can actually reach more than 80pc, as the bought/sold traffic is now premium, human-only traffic. If more and more agencies and brands start auditing traffic quality rather than the legacy score of a website, perhaps then we will have cause to celebrate.