By Eric Picard (Originally published on AdExchanger.com August 5th, 2013)
I’ve been struggling lately with some oddities in how the “programmatic” media space functions. Ad-tech infrastructure — both the “standard” digital infrastructure made up of publisher’s ad servers (DoubleClick for Publishers, 24/7 Open AdStream, Adtech, OpenX, etc.) and the real-time bidding platforms that run alongside those legacy platforms — has drawn huge investment. But misunderstandings and false assumptions abound about how these technologies operate, their limitations and what types of businesses should use them.
My friend, Jed Nahum, wrote a great article last week about programmatic buying and selling, including the sometimes confusing ways in which we use the word “programmatic” because of the complexity we’ve created. What I loved about Jed’s article is that he laid out a taxonomy of four different ways the “programmatic advertising world” operates (or will soon operate):
- RTB/programmatic spot: The RTB world we all know and love.
- Deal ID/private marketplaces: Using the “RTB pipes” to execute buys that are similar to direct buys, but that aren’t guaranteed.
- Programmatic direct: Using non-RTB “pipes” to buy directly from publishers, including guaranteed buys previously supported only via a direct human sales relationship.
- Programmatic forward: The (still-to-come) extension of the Deal ID/private marketplace world to guaranteed/reserved buys over the RTB “pipes.”
I really like Jed’s taxonomy because it calls out the very real differences in how various constituents in our space operate. If I do a quick mapping of vendors to their various spaces here (and I’m bound to forget a few), you can see that the players are siloed in their approaches. Full disclosure: My company, Rare Crowds, sits across all of these boxes today – although we don’t have a bidder or an ad server. We sit above that layer and push into each of those various platforms, so in a sense we don’t compete with any of these companies, but would partner with any of them.
The problem I have with this is that the vehicles to programmatically buy media are locked to the plumbing (technology layer) that supports them. And yet they’re all programmatic. That would be fine if we didn’t have this “programmatic forward” component at the bottom — the one that all the RTB folks are working toward.
I still stand my definition of programmatic as any method of buying or selling media that enables a buyer to complete a media purchase without human intervention from a seller. I’ll push forward by saying that — contrary to popular belief –the technology layer, the plumbing, is irrelevant to the channel.
When I talk to people from the programmatic-direct world, they argue they’re the logical path for managing guaranteed direct-media buys because they’re directly plugged into publishers’ platforms. But when I talk to the RTB folks, they make a very good argument about how they can expose publisher inventory “directly” between a buying platform and the publisher’s ad server with a check-box configuration setting — and that it’s better for buyers because they can apply all their first-party data to the buys and get the best of all worlds.
The reality is this: Both sides are “sort of” right. But, in the end, it just doesn’t matter. All the vendors will ultimately plug into each other and liquidity will flow. The programmatic-direct vendors, though, need to make sure that they don’t miss the value proposition of all the various partnerships they should be creating.
One of the most significant developments in our space in the last few years was the DoubleClick Ad Exchange’s rollout of dynamic allocation, the next-generation technology that replaces the previous AdMeld product. Essentially it’s a switch that sits between the exchange and DoubleClick for Publishers and makes real-time decisions about how to allocate inventory to the exchange and for guaranteed buys. Maxifier offers a very similar product that will work with other exchanges (which is also super important, but I give the nod to Google because of its scale).
The reason this matters is that publishers need a dynamic-allocation technology that regulates the decision about which line items get the impression. The yield increases are significant, especially those coming through the exchange.
Dynamic-allocation technologies level the playing field between the RTB players and the programmatic-direct players. Buyers will ultimately need to be able to support both channels. This is critical for decision-makers at agency trading desks or large technical advertisers with their own platforms, such as eBay or Amazon.
They need a way to rationalize when they should be doing dynamic buys, controlled at the impression level (RTB), and when they should be doing direct buys in advance. Ideally, they need a system that sits above all of these various channels and allocates budget to them in advance, but that also monitors and optimizes how those budgets are allocated throughout the life of a campaign.
Regardless of what any one vendor will tell you, all the functionality of the current set of “legacy” ad-server technologies will be replicated in the RTB stack over the next few years. And the current lines that sit between those stacks will get blurrier. Anyone preaching any kind of secularism here is a bit suspect.