Mix data sets with analytics tools for business results

As vice president of analytics and insight at Crossmark, Alex Siskos wanted to move his team’s role from transactional to strategic, from running reports to making decisions. But a hodgepodge of tools and manual processes for analyzing hundreds of terabytes of data annually stood in the way.

Crossmark provides consumer goods companies with sales and marketing services that it develops by analyzing shopper data from subscription services such as Nielsen, transaction data from point-of-sale systems and market trend data from industry associations. The company also pulls data from shipping forms, customer databases and emailed attachments.

[Related: 8 analytics trends to watch in 2015 ]

Siskos says that the time-consuming, resource-intensive process didn’t always yield what he calls “hidden relevant truths.”

“Depending on how good the person was at the tools, that’s how good your insight was,” he explains. “It was inconsistent.”

The process also created technical problems, he says. The sheer volume of data being brought together frequently generated error messages within applications or caused computers to freeze.

Andy Mulholland, an analyst at Constellation Research, says this is an issue for many companies as they try to use multiple data sets to harvest insights. “Blending sets of data together might well result in something meaningless unless done very carefully by an expert,” he says.

[Related: 5 Things CIOs should know about prescriptive analytics ]

Crossmark recently deployed a suite of products from Alteryx to combine and analyze data and tools from Tableau for visualization. The new systems let Crossmark “unlock the creativity of key insightful folks to improve the quality of the insights we’re delivering,” says Rob Saker, the marketing firm’s chief data officer.

For example, to ensure that a client’s product is in the right stores at the right time, Crossmark would blend disparate data—maybe POS figures, demographic statistics and geographical information—into a single form that analysts can use to see insights about consumer trends. That, in turn, can influence pricing and placement decisions.

Crossmark was able to do that in the past but could analyze only about 80 percent of the data it can now process, Siskos says. His team also built tools within Alteryx to let clients upload their own data sets into Crossmark’s systems efficiently and securely.

The company recently examined sales information for a product sold by a 5,000-store chain to determine how to increase revenues in certain regions. The Crossmark team zeroed in on 27 particular stores based on indications that a specific marketing strategy would work in their locales. “Before, I’d be in cross tabs of an Excel sheet trying to find things,” Siskos says. “Now I’m interacting with a map and it’s done in seconds.”

[Related: 4 ways to beat the Big Data talent shortage ]

Projects that once took days have been cut to a few hours—sometimes minutes. As Siskos explains, “It’s gone from someone trying to wrap their hands around how to do this to wrapping their minds around what to do strategically.”