"Data-Driven Thinking" is written by members of the media community and contains fresh ideas on the digital revolution in media.
Today’s column is written by Nico Neumann, senior research analyst for programmatic strategy and analytics at the University of South Australia.
Procter & Gamble, one of the biggest global advertisers, has repeatedly announced that it would cut spending for companies that do not offer third-party measurement. Facebook, YouTube and Twitter, among others, started accreditation processes or conversations with the Media Rating Council (MRC) to allow independent metric reporting.
And France passed an anti-corruption law for advertising, stipulating that agencies can no longer buy and resell digital media to their clients unless they are separate entities. Moreover, holding group relationships must be disclosed and media owners must send invoices directly to the advertiser.
The French transparency legislation should serve as a blueprint for other countries as it is high time advertising used similar regulations and disclosure requirements as other industries that are characterized by agency-principal business.
Dual Platforms And Conflicts Of Interest
However, while the new French decree addresses the conflicts of interests of media agencies, the reporting role for other intermediaries appears less clear. Last year, TubeMogul openly attacked Google for its advertising business model. In its manifesto for independence, TubeMogul criticized Google for making money as a seller and buyer of inventory.
It is difficult to deny this conflict of interest: The buyer wants the lowest sales price and the seller wants the highest. Thus, one may argue that letting the same agent represent both sides of the transaction is like having the same attorney represent both the defendant and plaintiff in a legal case.
To be fair, let’s recall that Google is not alone here. Many ad tech companies, such as AOL, Rubicon, Videology and AppNexus, have a seat on both sides of the auction table. And this characteristic is not unique to advertising.
In real estate, we also often find “dual agency,” the term used when the same firm or agent represents both the buyer and the seller in the same transaction. Likewise, stock brokers can be both sellers and buyers in a market.
Lessons From Commodity Brokerage
In particular, if a key player owns several parts of the trading ecosystem, market manipulations would be tough to expose without scrutiny. We have seen some very unethical practices here in commodity trading.
For example, the investment bank Goldman Sachs acquired warehouses, which bought up tons of aluminum and apparently manipulated logistics operations, creating delays as well as delivery bottlenecks. The result was an increase in aluminum prices and stocks. Of course, at the same time these actions appear to have achieved huge profits for Goldman Sachs, which ramped up aluminum trading operations simultaneously.
A US Senate investigation further revealed that Morgan Stanley owned oil tankers, pipelines and storage facilities, while JP Morgan had purchased electricity plants, all of which could be used to create artificial shortages.
Programmatic Trading And Market Manipulations
The big question: Could ad tech be exposed to similar tactics or other market manipulations? Unfortunately, there seem to be enough degrees of freedom that allow key players of the media system to adjust rules or charge according to opportunity.
For instance, Google states that it may choose at any time to close an auction at a price below the reserve price. AdX also has the right to change the auction rules to its discretion, add fake bids and play with publisher floor prices. How will clients know whether or not any such intervention provides a disadvantage for them or may have artificially inflated the revenue of Google at the expense of others?
Another point of concern is that many exchanges and supply-side platforms (SSPs) take a dynamic hidden fee from the buyer too. Consider a case where the highest bid is $10 and the second bid, at which the auction clears, is $2. Because the exchange has insights into the gap between these two bids, they could use this information to maximize their income and charge a fee of nearly $7.90, about four times the price of what the publisher would get. In other words, with undisclosed and variable buyer fees, SSPs or exchanges can exploit the information asymmetry and milk the buyers.
The problem is that the described scenario skews the perceived value of placements and inventory. For the last example, an advertiser may think the $9.90 it paid represents the market value, although the quality may be much lower, as indicated by the floor price less than the $2 CPM demanded by the publisher.
Generally speaking, the opportunistic fee modifications of SSPs defeat the purpose of a second-price auction, which should provide some value (the money saved due to differences to the clearing price) to the bidders to induce truthful bidding.
Transparency: A Must In Presence Of Conflicts Of Interest
Of course, the SSP practices of charging buyers or playing with auction characteristics do not necessarily need to lead to controversial behavior or market manipulations. Dynamic fees do not automatically mean high fees. And representing buyers and sellers in an auction does not always result in betrayal of one side.
However, such scenarios represent conflicts of interest and therefore require trust and complete transparency. And this still seems to be lacking in our current ad tech system. Companies that own or influence a large part of the trading ecosystem – the warehouses, pipes and oil tankers of programmatic – especially have a high responsibility to signal to the market that everything is going by the book.
To foster market efficiency and allow all stakeholders to understand the supply chain and maximize their value, bid details should be disclosed to all parties, as well as any fees or experimental interventions or modified auction mechanics. Otherwise, any unethical behavior, malpractice or gaming of the system is nearly impossible to identify.