Big Data isn’t just about quantity, it’s about the way the data is often presented—unstructured, gathered from numerous sources, and inconsistently formatted

Big Data might be the Big Darling of the digital world, but the decision to invest in a solution that can help manage and analyze all that unstructured information is one that needs to first be carefully qualified and considered.

Despite what seems like petabytes of articles on Big Data solutions, Big Data benefits, and Big Data excitement, one of the key problems enterprises face is still definitional, according to Robert Boehnlein, president of Teradata’s marketing applications subsidiary Aprimo, at the Teradata Partners conference.

Namely: the term “Big Data” is reductive and leads to an array of interpretations.

“Some people say it’s massive amounts of information associated with relational databases,” Boehnlein says. But Big Data isn’t just about quantity, it’s about the way the data is often presented—unstructured, gathered from numerous sources, and inconsistently formatted. “It’s more than just lots of data, it’s the structure of that data,” Boehnlein says.

The early days of Big Data focused on simply acquiring and storing that data. Once companies understand what Big Data is, they still need to identify the right applications to structure, analyze, and take action on overwhelming and often chaotic data sets.

Magazine publisher Meredith Corporation is currently evaluating the benefits of diving into Big Data, says Craig Gard, the company’s director of marketing applications. Unlike companies hesitant to rip-and-replace existing technologies for new ones, Gard has already experienced three full-scale technology overhauls while at Meredith. Instead, the publisher’s considerations revolve around which solutions to use and whether a Big Data plan is even necessary at all.

“I think the technology is a little new and it’s still growing up and I’m not sure if we know which one to go to yet,” Gard says. He points to two possible solutions: Teradata’s Aster—known for its analytics capabilities—and Apache’s Hadoop—known for its open-source framework that enables the processing of large data sets. Both solutions, Gard says, have unique features—and consequently both will require different skill-sets from internal Meredith staff.

And like other corporations, financial resources are a significant consideration. “I think you have to be cautious on Big Data, meaning that there’s a cost to it,” Gard adds. “If we go down the Big Data path, we might have to invest another million dollars. Is that worth it?” Additionally, Gard wonders if it’s even necessary. Meredith might be better served simply grabbing scraps of digital information—“golden nuggets,” as described by Gard—instead of relying on a constant stream of incoming data.

Boehnlein recommends that companies interested in a Big Data implementation begin with smaller, more narrow goals where it’s easier and faster to see success, and then gradually expand the strategy. “My advice is start small. Go for a quick win,” he says. “It’s easy to put big aspirational goals on the table relative to Big Data. These technologies are emerging and you want to show value to the organization in a reasonable amount of time. Define something very narrow and very specific. Let’s try email marketing and develop some steps with the email marketing campaign. And from there expand into, say, mobile messaging and mobile marketing. Add value in layers as opposed to doing everything in one big fell swoop.