Our Ad Infrastructure Limits Digital Ad Trading

Investors in programmatic media companies envision a future where media buyers sit at the industry’s equivalent of Bloomberg Terminals, placing orders for 20 million young moms or 30 million middle class dads at a time.  Part of this vision also includes bringing some concepts over from finance that we’ve learned to loathe in recent years, like high-speed traders “earning” money based on their ability to execute orders quickly, in the milliseconds before another order goes through.  Or perhaps allowing speculators with deep insight into the markets the ability to, say, buy up Olympics inventory and resell it to advertisers or other investors who do not possess similar precognitive abilities.  In short, they want to make the programmatic media buying market look a lot like the financial markets.

But there are important differences between our infrastructure and the structure of the financial markets.  In order to make the media markets operate like the financial markets, a number of things need to change, namely:

How Ads are “Ordered”

Most ads that are placed programmatically are sold in real time (or something close to it) to the highest bidder.  If you were to execute a stock trade – say for 500 shares of AAPL – in the same way, it would be as if the order were broken down into 500 orders for a single share.  But there’s also another important distinction: Ads expire.  That is, they represent the opportunity for one advertiser to show an ad to someone at a very specific time – now.  Something has to fill the space.  Shares of stock can sit until their owner will part with them at a specific price.  Ads can’t.  So if The Powers That Be wanted the ad marketplace to mirror the finance markets, it would likely have to move off impression-by-impression RTB and move toward packages of ad inventory that can be run in the future.

The Pipes

If ad delivery is holding up your browser’s rendering of content on your favorite content site, then you’ve seen first hand how overtaxed our ad infrastructure already is.  This will be especially true in a post-Net Neutrality world.  Gaining any sort of speed advantage for a media trader would involve one of two things:

  1. Paying tolls to telecoms and backbone providers to prioritize bid-related traffic, or

  2. Doing what high-speed trading companies do in the financial markets – invest in new infrastructure to gain a time advantage over others.

The latter could involve laying new fiber or investing in other information transmission methods with the aim of gaining a time advantage.


Right now, if you assess the size of the total programmatic display opportunity from a dollar perspective, and think about what a media trader could break off in terms of a profit, one thing becomes clear.  We’re not talking about the astronomical profit opportunity that would merit trying to move programmatic ad buying off RTB or making large-scale investments in laying fiber.  Of course, the programmatic marketplace is predicted to grow, but it’s not there today.  We already have too many companies on the LUMAscape competing for a small slice of total digital ad revenue.

I have no doubt that programmatic plays a role in digital media buying, but if you’re entertaining the notion of high-speed media traders making a profit in the arbitrage space, that’s not something that’s going to happen in the next five years without some radical investment and change.

Tom Hespos is a contributor at The Makegood and Founder and Chief Media Officer at Underscore Marketing, a boutique firm that creates and manages digital marketing programs. Look for Tom’s column the 1st and 3rd Friday of every month.