Is Programmatic A Recipe For Fraud?

PubMatic president Kirk McDonald gave MediaPost's RTBlog a rosy prediction of the future of programmatic advertising, saying, in part: "Brand marketers will follow in the footsteps of Disney, Procter & Gamble, etc., by building internal programmatic buying competencies and pulling programmatic spending in-house."

Brands populate the membership of the Association of National Advertisers, which has warned them and anyone else who would listen that the online ad business is rife with fraud, wasting billions -- an estimated $7 billion, at least, this year -- on online ads that people do not see.

The problem seems to be getting worse each year, as the ANA said that a 2015 study found between 3% to 37% of their ad impressions were created by bots, up from a 2014 study where the bot traffic ranged from 2% to 22%.



A White Ops Inc./ANA study showed that display and video ads bought using automated systems had a significantly higher level of fraud compared to ads that were purchased via human sales forces. Video ads, one of the fast-growing parts of the online ad business, performed far worse than display ads bought “programmatically.”

“Theoretically, we [marketers] should be pulling back from the digital supply chain because of all the issues such as fraud, the lack of transparency, ad-blocking and viewability,” Bob Liodice, chief executive of ANA, told the Wall Street Journal earlier this year.

Nobody denies the cost efficiencies earned by programmatic, but surely brands are getting a little tired of the accompanying fraud and waste. Even Facebook and Google, which gobble up the majority of digital ad dollars (64%, according to Pivotal Research analyst Brian Wieser -- last year Google did $30 billion, Facebook $8 billion and they are doing even better this year) are not immune from fraud. Google has its share of ad fraud problems, and Facebook has admitted to a few "mistakes" such as how it counts video views.

And all of this before we even get to the ad-blocker discussion.

I suppose you could make the case that there is a concomitant amount of "fraud" in other major media, such as when a TV ad is served into a room where the audience has left or muted the sound, or is counted in C3 or C7 but is subject to ad-skipping.

Interestingly, almost every study shows that the more humans are involved in the buying process, the less fraud.

While Kirk McDonald has a vested interest in a bright future for programmatic (and it's hard to believe that as an industry we will ever go back to more labor-intensive ways of buying and selling ads), one can't help think that sooner or later, brands themselves will say enough is enough, and enforce whatever threats are implied behind their "demands" for greater accountability.

4 comments about "Is Programmatic A Recipe For Fraud?".
Check to receive email when comments are posted.
  1. Doug Garnett from Protonik, LLC, December 8, 2016 at 5:12 p.m.

    This is not surprise. In fact, there are a couple of good laws to remember about automation.

    Tom Goodwin noted Weiner's Law recently which says (based on abiation experience) that automation removes small errors while occasionally creating extraordinary ones... like digital buying fraud.

    And my good friend @shahinkhan noted that his team finds that with Artificial Intelligence, if AI fails, it's impossible for humans to diagnose

    And this can entirely undo any theoretical advantages of programmatic. (Note that those advantages for non DR advertising remain entirely theoretical because no one can quantify what was lost in order to get suppose efficiency.)

    As a guy who grew up in tech, I find it astonished by the ad biz gets sucked in so quickly by shiny digital baubles like programmatic. Everyone needs a healthy skepticism of tech solutions in order to find the ones that matter.

  2. Ed Papazian from Media Dynamics Inc, December 8, 2016 at 5:59 p.m.

    Exactly, Doug. Why can't people recognize that advertising reactions and consumer actions are done by people, not robots and the same thing applies to doing business---like making an ad buy and placing the ad in a user -friendly as well as advertiser-friendly context. You can't automate that.  Automation can help with the grunge work, like transmitting information, tabulating tons of numbers, etc. but that's about all. The rest is up to humans.

  3. Paula Lynn from Who Else Unlimited, December 9, 2016 at 11:26 a.m.

    Sounds like winning at a casino by putting in 3 quarters and getting back 2 telling you that you won and then telling the public about how much was won by the patrons. Dupes are a dime a dozen. Here, bots spit the ads out to bots seeing them and everybody gets charged (what, the price you pay for something includes all costs ?). Who profits ?

  4. Craig Mcdaniel from Sweepstakes Today LLC, December 9, 2016 at 5:14 p.m.

    I wrote on Tobi's story along the same lines about another school for thought on this subject.

    Before the internet, what did we have?  We had human involvement in the whole process of any kind of advertising. Humans have been for the most part, have been taken out of the process. Sure human's made mistakes and cost more per ad and a number of things that made human's value much lower.

    However, human can put out even in digital advertisement a far superior product that is customized to the advertiser's need.  I recently added a custom video with a sweepstakes inside a Intel ad for a WPP company.  This ad cannot cheat the advertiser or agency and because the bill is direct and not automated.

    The point is the digital advertising market has gotten to the point that only a human inputs the ad then someone waits for a check to cash.  There is no real involvement, other than a gitch, in the process. 

    Humans got to be a bigger part of the WHOLE process again and not just on the front and back end to fix the problems.

Next story loading loading..