“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 Ari Lewine, co-founder and chief strategy officer at TripleLift.
RTB has revolutionized digital marketing. That said, it still has a number of well-known hurdles to overcome, most notably quality, fraud and viewability – each of which has spawned its own subindustries. While these issues justify the attention, there are a number of others that fly below the radar but expose the same degree of systemic problems in the RTB ecosystem.
On the buy side, the irrational valuation of third-party data threatens to undermine much of the promise of RTB. Meanwhile, on the sell side, RTB has introduced circular – and frequently infinite – auctions, the worst form of the publisher daisy chains that it was meant to address in the first place.
Third-Party Data: Overvalued
One of the greatest strengths that RTB has introduced is the ability to overlay cookie data, ostensibly helping clients more efficiently and effectively target consumers, wherever they may be. The reality, however, is that while brands have had success retargeting by leveraging valuable first-party data, the same cannot be said about third-party data. The unfortunate truth is that much of the available third-party data available simply lacks quality and reliability.
I recently checked the registry of several data providers and discovered that I was in more than a dozen mutually exclusive segments. I was both a mom with two kids and a male with “some college,” my income was in virtually every HHI group, and while I belong in the 25-to-34-year age segment, I was accounted for in other age groups.
More telling, however, is the fact that the number of BMWs sold in the US hasn’t really changed in recent years, but the cookie pool for BMW intenders has grown by an order of magnitude. This begs the question: What is the incentive of a data provider – quality or quantity?
Similarly, data providers are in the business of licensing out access to their pool of users on as many different axes as possible. So the same pool of users – the same exact impressions on the exchange – is being purchased simultaneously by hundreds of different buyers that are targeting all the different buckets that I appeared in. This, in turn, causes already-dubious user valuations to result in further-increased inventory valuations, compounding an already-serious ROI question.
The problem with user data thus becomes simple: If a single budget of $100,000 is targeted using only an intelligent inventory projection on the exchange, can you spend an equivalent amount paying for both user data and highly competitive impressions? The data is mixed at best. And if third-party data isn’t truly ROI-efficient, that calls into question many of the fundamental justifications for RTB.
Publisher Daisy Chaining: Repurposing Impressions
RTB would have solved publishers’ daisy chain headache, which is the proliferation of too many independent open exchanges with discrete sets of demands and different auction rules and price floors. RTB would have provided a single source where all demand could value an impression simultaneously if two rules were upheld.
First, the same impression would be valued exactly the same by demand sources, regardless of the exchange on which that impression is being purchased. Second, all demand sources would be equally present on all exchanges. Given that both of these are false, it is also not true that RTB solved the issue of daisy chaining. It made it significantly worse.
Fragmentation, caused by numerous independent exchanges – often with unique demand and auction mechanics – provides incentives for ad tech players to recycle impressions across as many exchanges as possible, trying to find the highest valuation. One can no longer rely on an impression being “fresh.” This undermines the sanctity of every single impression available on the exchange. There are a number of companies that have created an arbitrage business model out of this weakness by buying an impression in one exchange and selling it for a higher price in another.
One possible solution would be to append a unique exchange-specific auction ID to the bid request in any subsequent auctions, as well as all previous IDs for that impression. This would allow the buyer to follow the path of the impression, identify market inefficiencies and help platforms and exchanges address companies that simply arbitrage the same impression from one exchange to another.
If we can figure out a way forward as an industry in solving these two challenges of third-party data and exchange gaming, which are difficult but conquerable challenges, the programmatic RTB market will flourish at a level that we all thought was possible.
Follow Ari Lewine (@AriLewine), TripleLift (@TripleLiftHQ) and AdExchanger (@adexchanger) on Twitter.