Charting the DSP's Progress

By: Joe Zawadzki  Published: December 20, 2011

Link to original article

It was a cold time in early spring 2007, but a hot time for the ad tech market.

Yahoo! was buying Right Media, Google was acquiring DoubleClick and Microsoft was scooping up AdECN and aQuantive. Major media companies were making big bets – $850 million, $3.1 billion, $6.1 billion – to acquire the technology assets to make display media saleable via technology. Over $10 billion dollars changed hands, with a commitment to push more and more inventory into these automated systems. Billions of impressions, tens of thousands of data points, pricing precision to the one-thousandth of 1%. A revolution was about to start.

But from the buyer’s perspective, not so much.

Arrayed against $10 billion dollars worth of investment by companies worth north of a half trillion combined, you had two to three media planners and buyers armed with email, phone, fax and Excel representing Fortune 500 marketers.

The team that founded MediaMath in the summer of 2007 included members who had started [x+1], incubated Right Media, informed DoubleClick’s display strategy and built its platform, swapped investments with an ad tech cloud computing company called AppNexus, took capital from a former Nokia executive who was also an investor in an academic technology in search of an application (DataXu), and advised a social media ad serving company for Slide and RockYou called Invite Media that there might be a better opportunity out there.

ClickZ broke the news of this DSP concept in March of 2008 with the headline “MediaMath Adds Layer of Efficiency to Already Efficient Ad Exchanges,” which MediaMath founders read as “MediaMath Adds Second Layer of Sprinkles to a 14-Layer Birthday Cake.”

As opposed to: “MediaMath revolutionizes the fundamental principles of marketing by mapping billions of consumer touchpoints daily to individual brand objective functions, in order to precisely evaluate and automatically execute on each transaction to produce 10x+ improvements to true business results with full transparency and deep insights.”

MediaMath founders felt misunderstood.

Stephanie Clifford of The New York Times nevertheless saw the parallels to financial services and the transformation wrought there once quantitative and automated buying took hold on the buy-side early, with her piece in July of 2008 entitled, “Leftover Ad Space? Exchanges Handle the Remnants.”

Here the takeaway was that while the company was onto a business model, its efforts were still regarded as secondary – if you could sell with relationships, do that. If not, sell with math. Also, that its CEO could stand to hit the gym every so often. Either that, or New York Times photographs add 20 pounds directly to the midsection.

Fast forward to today: Joanna O’Connell, a senior analyst from Forrester, has produced a deeply researched piece on demand side platforms. It’s a category that merits a Wave report. The Consumer Reports of B2B software has basically said for the first time “Yes, the DSP is, like the television, a product category. Many people will want one; here are the best ones.”

It’s a reasonably awesome outcome, especially if you know the history of the lead analyst.

She was a massive skeptic at the outset, having managed Capital One’s budget of hundreds of millions of dollars at Razorfish. It took over a year to secure a pilot, but once results were in she aggressively launched one of the first agency trading desks. (ATOM Systems within Razorfish and Varick Media Management within MDC vie for primacy in this regard.) She constantly pitted company against company to evaluate results and always maintained the option to build in house.

As evidence, this author has the dubious honor of having been fired by Joanna in the ATOM Systems days in favor of Invite Media for having been late with a pure self-service interface, though an acquisition by Publicis and some mutual stubbornness over pricing didn’t help.

Three years later the market has evolved. Now when brands and agencies pick a DSP, they focus on advertiser results, not merely the simplicity of the DSP's user interface, as in the early days. The DSP has grown from bid management on display exchanges to being the platform by which all digital media is managed via algorithm, with full transparency, and with automated execution across channel.

The Forrester DSP Wave isn’t about a set of tools – it is about a transformation in how marketing happens.

DSPs are the dawn of a new era in which CMOs sit down with CEOs and CFOs to manage sales and profitability in real-time. They will be able to walk out of meetings where the CEO says: “We really need to boost sales of this product, profitability of that product, and net present value of that customer segment,” and the CMO says, “done.” They, or their team, log in and change settings, and magic happens across channels, across segments, across the funnel.

We are not quite there yet, but we will be. The DSP is the operating system that makes it possible.