Bad Neighborhoods in the Digital Marketing Metropolis
By Andrew Frank | July 08, 2013 | 2 Comments
Reports of rising crime in the ad ops sector, particularly the precincts associated with programmatic media and data exchanges, have been mounting steadily over the past year. Mike Shields, reporting for Adweek, deserves a great deal of credit for tackling the story head-on and naming names. In a typical report, Shields makes the case that “Suspicious Web Domains Cost Online Ad Business $400m per Year.” The problem has attracted the attention of John Battelle and the IAB, but as Battelle notes, “the bad actors are currently ahead of the good guys, and worse, many in our industry are turning a blind eye, hoping the problem goes away in time, without too much publicity.” That hope is becoming less likely.
I’ve been collecting first-hand research on this for a while now, with an eye toward a report and pitch that will be featured at Gartner’s Symposium this fall, but it’s worth relating this to our Digital Marketing Transit Map, to frame the question of how a marketer should deal with this. One could always take the approach of avoiding the programmatic media region altogether (if not digital advertising itself), but for those who still believe in the potential of real-time media markets to reach an elusive audience with well-timed, targeted messages based on demographics, product searches, and other real indications of purchase intent, it pays to know the landscape.
In particular, it pays to know where to go for help. The marketing police (I think you know them) are generally found in the marketing ops neighborhood, but their crime lab – which has the tools you need to fight fraud – is in the data ops region, where fraud detection is high on the list of big data use cases in non-marketing domains, and should be no lower in marketing. Using tools such as A/B testing and web analytics, data ops can help detect common fraud patterns such as click fraud in near-real-time, as opposed to the more typical situation in which it’s discovered months later while reviewing invoices. Impression fraud (often generated by botnets like Chameleon) is a bit more difficult but also detectable with the right tools in place.
In addition to media fraud, buyers must beware of fraudulent data from third-party providers. Here, again, data ops can help by applying traffic analysis to weed out bad targeting data. In any event, don’t venture into this area without a clear strategy for monitoring performance, and a plan to react quickly to degradation. Predictive analytics can help, as any deviation from predicted campaign performance should set off an alarm and mobilize your defenses.
There are shady players in any metropolis, but look for a concerted effort in the second half of this year to clean up this digital marketing neighborhood.