The Need for Speed in Real-Time Bidding
March 5th, 2013 Tom Lorek
Recently, I was given the opportunity to give a Tech Talk at the Digiday Exchange Summit. The focus of my talk was how increasing the speed of data querying doesn’t just increase the efficiency of the business user but also fundamentally changes how that user does their job.
Speed allows employees to approach each business problem through a dynamic, data driven method that includes multiple angles and with multiple hypotheses about the solution. Real-time bidding (RTB) is one area where speed is needed for publishers and exchanges to maximize ad pricing and revenue. However, most companies utilizing RTB do not realize the full benefits of what it can actually offer because most analytics systems lack speed.
In the ad network era, reporting techniques involved limited .csv files e-mailed from the ad network to the media buyer on a delayed basis that could extend to only every month. Producing this information required a labor intensive process led by high-salary, resident data scientists that usually provided only top-line data in an aggregated manner. No detail was available unless the media buyer was a big enough client and demanded it. Because of the lack of detail and the delay in production, any opportunity to make strategic business decisions was lost.
The other problem with these reports is that they are created ad hoc to respond to problems as they crop up. That means any new problem may not be covered. Unless you’re a data scientist with access to the internal database, you can never begin with an unfiltered, complete view of the data. You always start at some previous problem’s ending point.
Amazingly, even in the RTB era, most reporting is still done through .csv files prepared over very long processing times and communicated through inefficient means. This reliance on antiquated analytics prevents trading partners from responding to the fluctuating nature of market activity. At Metamarkets, we are addressing this problem with reporting that ingests data in real-time, queries the data at ultra-fast speeds and which enables operations teams to drill down in real-time to see patterns and anomalies in their data and act to take advantage of them.
Check out these images to see how visual real-time reporting speeds up RTB.
If you’d like to see the presentation, please check out our youtube page.