GroupM's Rob Norman: Is the Headline Cost of Ad Fraud and Non-Viewable Ads the Real Cost?
No one disputes that digital ad fraud exists. The ecosystem is under constant and evolving threat from non-human traffic in the form of bots. Nor does anyone dispute the existence of non-viewable advertising impressions. Many ads appear in areas of sites that cannot be seen -- in part or whole -- by their intended viewer. The Association of National Advertisers has determined that this an issue of extreme importance to its members.
Forrester, in a report published on March 30, defines the issue in aggregate as invalid impressions and places a "nominal dollar wastage" in 2016 of $7.4 billion. This relates "only" to desktop and laptop internet usage and only to the United States.
Because this number has caused considerable alarm for advertisers, we believe it necessary to provide some validation, for those advertisers that invest in tools to protect themselves and especially for those that do not.
The key word in the descriptor is "nominal." "Nominal" refers to the aggregated value of all invalid impressions, assuming that advertisers paid the same market average price for them as they paid for valid impressions.
In its report, Forrester recognized that "real economic wastage" -- the actual money -- may be less than the nominal dollar wastage.
To add some context we have attempted to triangulate this data with other sources. $7.4 billion compares to $39 billion of U.S. desktop and laptop ad expenditure that the Interactive Advertising Bureau and PricewaterhouseCoopers reported for 2015, the most recent full year report. It can be inferred from the same PwC data that search constitutes 50% of that $39 billion. The Forrester report does not address search.
If we add the $7.4 billion nominal dollar wasteage to the $20 billion non-search PwC figure we get to a total of $27.4 billion. We accept that we are adding a 2016 estimate to a 2015 actual for the purposes of the exercise. But this suggests that 27% of the traffic is invalid, a significantly lower figure than Forrester calculates.
Equally hard and more important to reconcile, something Forrester did not attempt, is the relationship between nominal dollar wastage and real economic wastage. The real waste is of course the figure that counts.
Advertisers that pay for outcomes and engagement that only a human could deliver suffer zero loss from either fraud or viewability problems. Because bots can't buy anything, but can click, marketers that pay for clicks may indeed suffer loss.
Advertisers that deploy tools that allow them to pay only on a human viewability standard of their choice suffer zero economic loss: There is no premium incurred for paying for something when the alternative is paying for nothing. It should be noted that viewability is by far the biggest component of invalid impressions and the easiest to detect.
Advertisers that deploy fraud detection software and don't pay for non-human traffic suffer losses only to the extent that fraud detection is improving but never perfect. It's worth noting that the MethBot incident identified by White Ops as a $3 million-to-$5 million-per-day fraud was re-evaluated by the Trustworthy Accountability Group at $250,000 to $500,000 per day.
Publishers that serve non-viewable ads or serve ads to bots suffer zero direct loss as they did not create an opportunity for consumers to see the ad, nor did a bot deprive them of a human visitor.
Let's add another layer of context. Forrester estimates invalid impressions for mobile advertising at 1%. When the PwC/IAB report card for 2016 is in, it's likely that the total market value will be close to $70 billion. Half the total will be mobile and half of the total will be search. If all advertisers protect themselves, it is highly likely that the real economic wastage will be well below 2% for ALL digital advertising, giving a total real cost to advertisers of around $1 billion. It may even be much less.
Whatever the actual number is, it's nowhere near $7 billion -- and that alters the narrative.
$1 billion is a lot of money so there's no room for complacency in respect of real economic wastage, but there is room for context. Recently the market in digital advertising inventory has been described in the most unflattering terms. Vigilance must continue but it's time to move the focus of the debate away from nominal economic value and to the important issues of inventory quality, user attention, effectiveness, and the creation of business value. It's just as important to ensure that all protection tools are activated and that similar vigilance is conducted in markets outside the U.S. where viewability measurement may be less widely deployed and counter fraud measures less effective.