ANA Details The Supply Chain Fees That Agencies Try To Hide

The Association of National Advertisers’ (ANA) inquisition against nontransparent programmatic media buying rages on.

The ANA on Thursday released a study illuminating the often murky fees taken throughout the messy programmatic supply chain, from advertiser, to agency, to trading desk, demand-side platform (DSP), exchange, supply-side platform (SSP) and publisher. Read the study.

The survey, conducted from May 2015 to April 2017 in conjunction with Ebiquity, the Association of Canadian Advertisers and ad tech company AdFin, examined metadata on 16.4 billion media impressions from seven major advertisers. That spend flowed through five DSPs, one trading desk and six agencies.

Participating marketers spanned 30 major brands in auto, CPG, banking, fashion and travel. All impressions surveyed were part of fully disclosed programmatic buys.

On average, 42% of each programmatic dollar was spent on nonworking media – meaning it went toward tech or agency fees. Only the remaining 58% reached the publisher. In some cases, the survey found a 30% to 70% split.

“When the advertiser spends a dollar and only a fraction ends up with the media company, one needs to question whether the intermediaries are adding value or charging too much,” said Bill Duggan, group EVP at the ANA.

Demand-side technologies and services extracted a big chunk of change. When isolating the demand side, 72% of every dollar went to the exchange. Breaking down the 28% that constitutes nonworking spend, 12% went to execution costs, 9% went to third-party targeting data, 6% went to agency fees and 1% to “other.”

On average, demand-side fees added 45% to the cost of display inventory and 35% to the cost of video inventory.

While members of the supply chain have a right to make money, buyers must educate themselves on where that money goes in order to negotiate fees, said Andrew Altersohn, CEO of AdFin, which analyzes programmatic transactions.

“You need to know the cost of each element of the supply chain to have any chance of negotiating, finding better suppliers, cutting out a piece or whatever you might do,” he said.

Advertisers that understand the supply chain can optimize spend on things like third-party data and figure out how to make their dollars work harder, he added. For example, advertisers can ask smarter questions when they can see a certain DSP is working better with a certain SSP, or that buys are more effective with third-party data bought from a specific source.

“When you look at costs holistically across the supply chain, you can arguably optimize the supply chain,” Altersohn said. “When the advertiser has this information, they are in more of a position to ask smart questions to their partners.”

Notably, fees taken by agency trading desks were underrepresented in the ANA’s report. More than 95% of the bids the ANA analyzed were not bought through agency trading desks due to nondisclosure agreements in agency contracts. Only seven out of 58 marketers who expressed interest in participating were able to obtain the necessary data needed from their agencies.

Agencies also discouraged several advertisers from participating in the study and bottlenecked requests for data, even when clients had signed disclosed contracts.

“We had a lot of very rapid interest when we announced this,” Altersohn said. “Slowly but surely, they just disappeared. If you want to know where your money is spent, what it buys and the value you’re getting, it should not require what we went through to figure that out.”

Inability to measure buys from agency trading desks may have reduced the average demand-side fees found by the report, the ANA said.

But it’s not just the agencies who stood in the way. DSPs have inconsistent and limited processes for mapping insertion order to ad server data. The report called for more standards, protocols and cost benchmarks to ensure advertisers have access to the information they need to analyze programmatic media spend.

“The advertiser has to at least understand they have a lot more leverage than they sometimes act like they do,” Altersohn said. “It’s almost as though they’ve thrown their hands up to say, ‘It can’t be done I won’t even try.’

“What we’re saying here is it absolutely can be done.”


Popular On AdExchanger Right Now:

2 Comments

  1. But did those campaigns deliver on their KPIs? Surely that has to be taken into account also as relevant context to those data and tech costs?

    Reply
  2. I disagree with everything in this “report” as it not only shows a complete lack of misunderstanding of the programmatic supply chain but is very self-serving.
    Let's use this as an example....you're buying a new car. It has several "pass through" components all with their own costs and suppliers. Tires, headlamps, maybe a custom engine, radio, paint, etc that all contribute to the final price. Do you ask the auto dealer what all of those individual costs are to understand the supply chain? No. You're looking at the price sheet on the window for the complete car.
    Programmatic is the same. It's an assembly line of sorts with the publisher cost just one component of it. To isolate how much is ending up at the publisher is looking at a singular data point that is not indicative of the whole. It also shows the ANA’s old school viewpoint on media when you’d go media direct for TV, radio, newspaper etc. Those rules though don’t apply in digital.
    As the previous poster mentioned, advertisers should be looking at KPIs to judge the value…not the media cost. What if the $6CPM media delivered an average CPC of $.50 and 25% ROAS where as the $3CPM media delivered an average CPC of $1.00 and only a 10% ROAS? What’s more “efficient” in this example? Or knowing how many advertisers work, would that $6CPM even be considered in the first place (my personal experience tells me “no.”).
    Media cost by its own measure is not “value” nor “success”, it’s just a cost. We live in an era of performance and programmatic is about performance based marketing. The end result is where the value lies, not random supply chain cost benchmarks.

    Reply

Add a comment

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>