December 5, 2012 – Big data means much more than technology or hoopla, says noted data quality expert Thomas C. Redman. The best, successful big data projects should be about getting quality enterprise information to the appropriate end users and letting managers make business-minded moves based on the results.

Redman, president of Navesink Consulting Group and author, led a discussion Wednesday entitled “Organizational Imperatives in the Era of Big Data.” It was hosted by the Harvard Business Review and sponsored by Hitachi Data Systems. As a basis for the insights presented, Redman dialed into his past in establishing applied research programs on data quality at AT&T as well as his ongoing work in establishing enterprise data management programs and practices as an outside consultant.

Perhaps obvious but no doubt vital, big data plans should start with an assessment of business need. How and where they could lead to better business decisions or refined information use across the workforce hold much more value than just piping in more data sources or dumping out more, possibly bad, data to customers, Redman says. And this lines up with another focus area when sizing up big data implementations: Think end-to-end “from the beginning.”

“There are many ways to make money from data. Big data is only one,” says Redman.

Taking on big data and advanced analytic capabilities also provides the opportunity to route (or re-route) data to the correct department and teams. For example, Redman says that analytic capabilities should be close to the line of business user for day-in and day-out improvement. Other analytic capabilities should rest increasingly further from the LOB for more fundamental discussions.

“At too many organizations, their top data person is some architect buried in the bowels of IT, and things just aren’t working. IT has proven to be a bad home for data, and step one is to get it out.”

And when that data is pointed in the right direction, there must be a concerted effort to get savvy, supportive people working on it. Data scientists can pull a lot from data, but they may not be so easy to find or it could take training of in-house talent. As the analytics roles are culled, enterprises should take a longer look at the bosses running data teams and experiments, what Redman called “switched-on managers.” Executive support and awareness of data digging brings in a solid business perspective on big data actions. And Redman says those actions should also come with some understanding from the C-suite that big data involves time and risk.

Enterprises should sew in data quality efforts to get the most out of big data. This is diligent work, but no “miracle” is required, and it may result in new benefits in clearing up hidden costs in accommodating for bad data, Redman says.

Finally, Redman says that businesses of all sizes must be ready for change not only prompted by a big data implementation, but from the findings and experimentations with the subsequent results. Here, agility is crucial, and Redman expects it may result in an “extreme rearrangement” of the top global companies over the next generation, particularly based on the moves by smaller, more nimble companies and the ability to farm out data mining and analysis.

“Sooner or later, data is going to be driven into every nook and cranny of your organization,” Redman says. “You cannot escape the need to build organizational capabilities and organizational structure [for quality data] to be effective.”