Researching real time bidding is a heady experience. The vision of ad placement running like a stock market – values flashing across the room, changing by the second as harried brokers shout and signal – what an image! Of course, that’s not how things will really look. Instead, the millions (and later billions) of daily transactions will happen in the cool sanctity of the cloud, with very little human intervention at all. At least, that’s the concept.
In the advertising world around the corner, Don Draper will be replaced by a pimply quant in blue jeans and sketchers. He will sell algorithms instead of ideas. It will be a world where databases compete for advertising. The media outlet with the best knowledge of its audience will hope to marry its avails to an advertiser whose knowledge is just as good. Such a union promises remarkable ROI as sales skyrocket. The key, of course, is the data.
Do media outlets know enough about their audiences to allow RTB to thrive? Right now, even though volume is burgeoning, the demand for audience data is still primitive. RTB still moves mostly remnant inventory – even though TubeMogul recently coined the term “premium remnant” to describe price jumps for in-stream video ads during the recent NCAA basketball tournament. Although the title is confusing, the idea has impact. As RTB becomes more efficient, remnant inventories should shrink. As this occurs, the remaining inventory will become more valuable. Eventually – these days, that means within a few months to a year – the algorithms that make the buys will demand more data to fuel their efforts.
Facebook and Google are both trying hard to meet the data challenge. Each has fielded data meant to feed these complex formulae and make them more successful. Tomorrow’s ad industry may be a very strange place indeed, with A/B tests run within minutes, and audience data quality making the difference between success and failure for a Website, a TV station, or a newspaper.
Will creative still matter? The current answer is yes. It is a waste of time, proponents say, to find the right audience only to deliver them an ineffective message. Will creative have to keep pace with the data? Will ad builders have to work 24/7 altering messages to satisfy a changing data mosaic? Or will messages be simplified so they can serve more than one audience? The answer probably lies somewhere in between. No doubt programmers are at work right now, trying to automate creative to meet the time demand of RTB.
A long time ago, when the cloud was called time-sharing, I worked with a tech who taught me a great lesson. “See that key,” he said, pointing to the ENTER key, “that’s the U $ Spend key. Once you hit it, if you made a mistake, there will be no time to correct your error. The computers are too fast. Before you can draw another breath, you could owe my company a million dollars.” The tech taught me to always test what I wanted to accomplish on a small subset, and only to push that key when I was certain everything would work properly.
The work to automate advertising proceeds at a feverish pace. Most observers think (and I agree) that RTB and what it will become is here to stay. But I have to wonder as the pace accelerates, is that old tech (or his grandson) still around? Or is some poor soul about to push the U $ Spend key?


