The Data Flow Is Like A Broken Water Main

“the executioner” opinion expressed below is written by Sam Temes, Associate, Platform Management, Ad Exchanges, at Razorfish.

The ExecutionerThe CEO of a major ad exchange recently likened the flow of data from real-time ad exchanges to a fire hose – I like to think of it as more of a broken water main. There is so much information coming out and so quickly that it’s impossible for most to do anything reasonable with it. Just participating in real-time bidding requires the processing of thousands of transactions per second, reliable data centers that can pass information instantaneously, and continuous up-time. As an agency buying on behalf of my advertisers, it is technology I know little about, but I know that my clients can benefit from the kinds of insights that result from it.

As more exchanges introduce real-time bidding in 2010 and ad networks, ad exchanges and data providers enter the platform space, there will be a rush to see who can deliver the best performance and the smoothest UI. But for a platform to truly meet an advertiser’s needs (and help grow the industry), it needs to have expert plumbing that both supports scale and speed, and also allows buyers to generate insights from the data that are both predictive and significant (not to mention interesting).

Having the best algorithm won’t be enough for the savviest agencies. We, on behalf of our advertisers, are already expecting to learn what makes exchange buys perform and what we can learn from performance to help advertisers’ broader marketing. The key to a best-in-class platform will be having an expert decisioning tool and the technology and analysis to make sense of the decisions for humans. As more players enter the demand side platform space, the competition for the best decisioning algorithm is sure to heat up. But in order for exchange based media buying to keep growing, we need the tools to prove to advertisers that we can make use of the information inside the box.

Follow Sam Temes (@SamTemes), Razorfish (@Razorfish) and (@adexchanger) on Twitter.

Enjoying this content?

Sign up to be an AdExchanger Member today and get unlimited access to articles like this, plus proprietary data and research, conference discounts, on-demand access to event content, and more!

Join Today!


  1. Good insights! I think one of the most interesting you’re drawing is between the decisioning algorithms (which will quickly converge to become a commodity among the smarter players) and the broader “meta-data” that can drive better performance from future campaigns.

    I’m a big believer that there’s an enormous amount of value in that meta-data, and will be very interested to see how agencies, clients, and their platform partners determine who owns that meta-data and what they can do with it.

  2. Amen. Just being able to deal with a bid request volume of tens of thousands of requests per second is really hard. But, you are absolutely correct that the data that comes from that pipe is a treasure trove for marketers looking for insights.


  3. Even if you do in fact know little about technology, you do have a deep understanding of what insights will be relevant and interesting to marketers. This knowledge is key to developing buying interfaces that meet the needs of the agency and the ultimate end client as well.

    As you allude to in your post, there is often a difference between the data that is predictive of performance and data that is insightful for clients. I find that clients are much less interested in specific exchanges and automated optimization methods for driving higher clickthrough rates. They are more interested in who is buying their products and why.