Boston-based DataXu is doing something that many people consider impossible – near-real time analytics on Hadoop HBase, which is supposed to be a batch system. And this is no lab experiment. It is gathering all the data on what is happening with hundreds of companies on the Internet in real time to support analysis of marketing campaigns and positions for 400+ customers.

“Active Analytics is the idea that we are measuring everything in real time,” says DataXu President and CEO Michael Baker in an interview with Wikibon Co-Founder David Vellante originally webcast in June (full video below). “We measure everything that a Ford or AmEx does online and what people are posting, etc. It measures what’s working – what creative message with what groups of people in what context. That allows our 400 clients to redirect their marketing investments to double the efficiency of advertising online.”

DataXu customers use Active Analytics to accomplish three things:

1. Make yields better: They can get more bang for their advertising buck, get more customers, make more sales.
2. Engage with people: “In the digital lifestyle the consumer’s in control, cruising around on a mobile device, watching video on her iPad, or hanging out on Facebook.” Active Analytics allows advertisers to understand how to engage with people to drive sales.
3. Reduce cost: Active Analytics allows marketers to test campaigns in limited real life trials and adjust them for maximum effectiveness and avoid investing in campaigns that will fail or that have unintended consequences. This is much more accurate than the traditional approach of working with focus groups.

DataXu is a SaaS company that works by providing a front end user interface on customers’ laptops. This allows marketers to ask their own questions, design their own searches and analysis and get answers based on what is happening now rather than on a batch system providing static reports two days later. This speed is important in today’s high-speed digital environment where marketing executives need to stay out ahead of events, positive and negative.

The impact is to change marketing from a fuzzy practice based largely on guesswork to something approximating if not an exact science at least a fact-based art. It allows marketing campaigns to be cost-justified based on their returns, measured in terms of thousands of reactions captured off the Internet in DataXu’s Hbase database.

This, however, is only the beginning for Active Analytics. “Think about the Internet of things,” Baker says. “Today we think the amount of data coming from social media is huge. In a few years we’ll have a trillion connected sensors deployed in things, providing reports. They will be in the ballasts in lightbulbs, in valves in city water systems, electrical switches. And we will need ways to gather, process, and filter all that data in near real-time not just to produce a report but to make active decisions from replacing a bearing that is beginning to fail in a power generating station to changing that lightbulb.”


Watch live video from SiliconANGLE.com on Justin.tv