Learning from role models who are already successful at their job - as part of our interview series, editor Aline Prigge meets Vida Brychcy, Director Product & Engineering Test and Consumer Consulting at Chip.
Apart from low viewability, ad fraud is one of the problems the digital advertising industry is closely concerned with at the moment. Ad fraud mainly includes non-human traffic, also referred to as invalid traffic (IVT). Sites are visited by bots that generate unseen and therefore false impressions. Some bots can be recognized easily, others are difficult to detect, as they create user profiles and imitate human behavior. Admittedly, with 2.2 per cent, the proportion of potentially deceptive online ad impressions is very slim in the German market, according to Bundesverband Digitale Wirtschaft (BDVW) (BDVW quality report 2017). Nevertheless, advertisers expect more transparency. BurdaForward Advertising will therefore measure its IVT inventory regularly from now on and publish the figures transparently online.
Carsten Sander, director of product & development at BurdaForward Advertising, provides some background information concerning the new service in our interview:
Why is invalid traffic an important issue for BurdaForward Advertising and your customers?
We want the quality of our inventory of programmatic and direct advertising to be as good as possible. The measurement parameter has to be viewability generated by real users of our services, which does not include ad impressions generated by machines. This is what our clients expect when booking services with our goodvertising marketing network. Transparency is an important part of our goodvertising initiative. We have dealt with the IVT issue because it is very important to our clients at the moment. Fortunately, our IVT figures are extremely low and we would like to communicate that, of course.
What practical problems does IVT involve for marketers?
Invalid traffic cannot be recognized in advance in every case. Often an invalid impression only comes to the attention of our ad server provider as part of their downstream analysis. Therefore there may be differences in the counting systems, and sometimes the delivery figures even have to be corrected up to 12 hours later. Understandably, such variances cause distrust, and the ex-post analyses result in extra effort on each side, for buyers and sellers. The percentage of inventory generated by machines should not be relevant for marketing, but it partly cannot be anticipated as such. There are some known producers of IVT, but new ones appear every day. Detecting these new ones is different depending on the procedure and the algorithm of the ad server provider. This then again leads to different figures in the chain.
What does your approach to solving this problem look like?
In terms of measuring, we collaborate with the three top providers MOAT, Meetrics und Integral Ad Science. We use two approaches: The first involves MOAT measuring the IVT percentage as continuously as possible in almost every one of the advertising campaigns we deliver to our clients. MOAT is our current service provider for measuring the viewability of our clients’ advertisements. In addition, we have our other measurement partners Meetrics and Integral Ad Science send us their figures for our network once a quarter. All three measurement results are then published. Furthermore, our Google adserver has a built-in IVT analysis tool, which is generally able to correct the delivery figures within the first 12 hours of the following day. We will disclose the IVT rate for our network on a website specially created for theses analyses from now on. Basically, we use the options offered to us by our partners to measure the quality of traffic in our network.