A new study conducted by Metamarkets, a provider of interactive analytics for programmatic marketing, reports that despite the growth in the programmatic marketplace, existing analytics tools are limited, unable to process some 600 billion ad impressions daily.
Metamarkets cites the fact that most marketers rely on static reports to track campaigns and the data becomes stale quickly. In addition, most databases can be probed for specific questions, but the process is slow and cumbersome. A data scientist or other specialist must set up the query, and the business user then waits hours or days for the results to come back. By the time an answer arrives, it may have already become irrelevant.
Among the study’s findings:
Metamarkets’ research touts interactive analytics, an approach to exploring and visualizing data that puts the full power of big data into the hands of the people who need it.
Interactive analytics, according to the research, focuses on access to real-time data overlaid with historical data to offer context, offering marketers a complement to programmatic media buying in the form of "fresh data, self-serve dashboards and the ability to slice and dice data and drill into any specific metric, as opposed to having a predetermined set of questions to pick from like many traditional analytics tools offer," Mike Driscoll, CEO, Metamarkets told Real-Time Daily via email.
The self-serve aspect enables marketers to use an intuitive, point-and-click interface. The white paper argues that interactive analytics helps turn each user into his or her own data scientist to explore all aspects of the data.
Metamarkets cited DoubleVerify as an example of the importance of the self-serve function. DoubleVerify authenticates the quality of digital media for brands. Since deploying Metamarkets, the firm has shifted from a centralized data analysis approach, to integrating responsibility for data analysis within each account team.
"... most databases can be probed for specific questions, but the process is slow and cumbersome. A data scientist or other specialist must set up the query, and the business user then waits hours or days for the results to come back."
Welcome to the wonderful world of traditional old school BI systems. Get in the queue with everyone else and wait your turn.
Henry: Why do clients accept such slow "batch" turnaround? Days is simply unacceptable. Jim
Henry - Yes, it's precisely these old school BI systems that don't enable interactive analytics that needs to be swapped out. Slow databases were a problem when data points numbered in the millions, but a few days of programmatic data routinely result inbillions of records. So the problem has been compounded. There are a number of next generation database platforms that do enable fast queries at scale, including one we invented and use at Metamarkets (Druid.io).