Programmatic Rocket Scientist To Skeptical Ad Exec: "Robots Don't Buy Things"

At least not yet, Rocket Fuel CMO Eric Porres said in response to Jen Brady Thursday morning at OMMA RTB. Brady, a self-described skeptic of programmatic media-buying, is founder and CEO of Chicago-based agency Fred & Associates, and she said that as far she’s concerned, programmatic should be spelled “problematic.”

In fact, If I’m not mistaken, she used that term in a Freudian slip when she said “fraud” in programmatic media buying “totally exists,” and then went on to describe her “problematic partners” in the sector.

Or maybe she meant it literally.

In either case, she picked it up from PHD Chief Operating & Digital Officer Craig Atkinson, who kicked off of the aptly named “Can A Robot Make You Cry” panel discussion, sharing a personal anecdote about a malaprop his son made when he asked his father, “Why do you keep talking about problematic media” around the house?



Much of the panel focused on the same discussion preoccupying similar panel discussions across the industry and articles and columns in the industry trade press, including MediaPost’s.

That said, Porres believes the problematic parts of programmatic may be over-exaggerated, because the technology skeptics say is contributing to “fraud” and “non-human traffic” can also be used to shed light on the most human behaviors. You just need to know what to look at.

“Robots don’t buy things,” Porres pointed out, as an example of a metric people could utilize to understand what portion of their programmatic media buys are actually being served to humans.

Later in the morning, during a separate presentation about how a giant consumer marketer is tackling and taming the problem, Kellogg Co. Senior Manager of Digital Media Jim Kiszka demonstrated how Kellogg’s internal trading systems now scientifically account for it. Basically, he showed how they are able to factor what portion of their programmatic audience deliverability is actually viewable to people, and they benchmark their buys -- the prices they pay and how they measure the return for it -- based on that.

Currently, he said that approach has increased Kellogg’s reach of viewable impressions vs. the rest of the industry. He’s no Polyana either, noting that Kellogg’s trading team doesn’t expect to ever get to 100% viewable delivery, but it can improve and Kellogg can account for the portion that isn’t viewable. In other words, they’re using scientific methods to account for and plan their media buys.

The approach Kellogg is using, based on Kiszka’s presentation, is lightyears ahead of others in the industry and I recommend other clients and their partners take a close look at it. It is the right approach to what is still a problematic form of media-buying. Just like the other problematic areas of media-buying that the industry has confronted, adapted and improved on over time.

In terms of potentially problematic partners, I know Rocket Fuel is not one of them. And it’s not just beause of how scientific they approach the marketplace. It’s how they make their market inside of it.

“I’m being measured on my ability to deliver a human audience,” Porres explained, adding, “We don’t get paid if we don’t… We can’t afford to deliver ads to non-humans.”

That’s reason enough for me. But I don't buy media, personally or programmatically. But I do know the difference between fraud and Freud.

2 comments about "Programmatic Rocket Scientist To Skeptical Ad Exec: "Robots Don't Buy Things"".
Check to receive email when comments are posted.
  1. Tom Cunniff from Tom Cunniff, October 3, 2014 at 12:14 p.m.

    If all that algorithms measured was clicks, I'd be a skeptic too. But the range of data points a company like Rocket Fuel evaluates are far broader -- including actual purchase behavior. One advantage of their machine learning approach is that fraud gets tougher and tougher to pull off: the algorithm relentlessly learns from both real human signals and fake bot signals. This learning happens 24/7/365. Fraudsters originally turned to online ads because they were a much softer target than financial services companies. But this is changing fast, especially with the information sharing done by IAB's Traffic of Good Intent Task Force.

  2. Henry Blaufox from Dragon360, October 3, 2014 at 1:43 p.m.

    Fraud reduction is a work in progress. The industry won't solve this in just a week or two. But by analyzing the ever growing data and more precisely detecting patterns that are suspicious, the detection and then prevention level will improve, to the point where fraud isn't lucrative for the swindlers, or a material financial hit to the advertisers, agencies, publishers and tech firms. Then it becomes a matter of catching and preventing new tricks for gaming the system as technology evolves.

Next story loading loading..