The human touch will lead us out of the ad fraud storm
By David Hartley, Product Innovations Director at Data2Decisions, 29 January 2016
Last week, the Association of National Advertisers (ANA) and White Ops, a leading digital ad fraud detection and prevention service, put out its annual study on digital bot fraud, which estimated a $7.2 billion loss for advertisers. Whilst the immediate impact of online fraud is wasted advertising spend, there will inevitably be a longer-term impact on advertising effectiveness.
There has already been some headway made to tackle the problem - the ISBA, IAB and IPA set up the Joint Industry Committee for Web Standards to establish good practice. Whilst the Trustworthy Accountability Group has also said it will start certifying advertisers, publishers and industry middlemen to make them more responsible for bot traffic.
Whilst these are steps in the right direction, as an industry we need to be more effective. Let’s be frank, this problem isn’t going away and we need to step up.
Fraud artificially increases the number of impressions in order to increase the advertising cost. Therefore, if performance data includes consumer activity that never actually happened, it can skew our understanding of effectiveness which goes on to inform future investment decisions.
We now operate in a world where we can directly link the spend on digital to revenue uplifts. Although bots can mimic some online behaviours like ad views and clicks, there are still many human behaviours they cannot, like interacting with ads and making an actual purchase. Despite the evolution of AI and the speed at which bots have evolved, spending our hard-earned is, for now, a behaviour that remains uniquely human. By keeping this front of mind, the number of impressions becomes irrelevant as a metric.
The battle between true performance data and the fraudsters will be ongoing, so it is important to focus on analytics that will help manage the inaccuracies in the data. Unfortunately, many advertisers are still grappling with the relatively basic issues of how to collect, store and organise the avalanche of data in a coherent way. This in turn it makes it impossible to see the holistic ROI picture.
A lack of faith in basic performance data coupled with an inability to understand the true revenue impact of their digital spend is a dangerous combination for advertisers and a bad omen for a growing channel that presents so many opportunities for brands. Online media owners must act together to prevent a collapse in trust before it is too late.
So what needs to happen? We are already seeing specialist companies such as Forensiq building clever algorithms to detect bots. The technology exists so it is now incumbent on advertisers, agency groups and legitimate online publishers to work together to weed out the corners of the internet that are infected and starve them of the cash they need to perpetuate themselves.
In my opinion, however, measuring revenue growth is as bot-proof as it gets right now and one of the best ways to ensure an accurate view of your advertising effectiveness.
Media owners are increasingly asking for evidence of the effectiveness of their online advertising. This has been coupled with an increased investment by advertisers and agency groups in IT infrastructure and data management platforms to organise and store the relevant data. This increased access and velocity of data can be used to create new analytical products that can prove the true impact of digital spend on revenue. Combining advanced attribution modelling techniques with 'digital ecosystem' modelling allows us to develop an approach to get to a highly granular read on revenue growth performance that bypasses the bots entirely.
The industry needs to stick together for the long haul and face this fight head on. By doing so we will overcome ad fraud and reap the benefits of online advertising. Ad fraud, we are ready for you.
This piece was originally published by [_PerformanceIn_](http://performancein.com/news/2016/01/29/human-touch-will-lead-us-out-ad-fraud-storm/)