news.cuna.org/articles/115790-turn-big-data-into-small-data
2019-03-22_TurnBigData_115790_1
Randy Harrington leads attendees through the “Launch Experience.”

Turn big data into small data

The more data you have, the less certainly there is.

March 15, 2019

Harnessing big data requires credit unions to assess their “data ocean,” communicate about what they want to learn from the data, and test their assumptions, says Randy Harrington, CEO of Strategic Arts & Sciences.

He addressed the 2019 CUNA Marketing & Business Development Council Conference Thursday in Las Vegas.

“When we talk about marketing today, we’re talking about data,” Harrington says. “Using data requires having the right culture. It’s hard to do because it’s complicated.”

He offers four critical requirements for analyzing data:

1. Define the domain of your data. What data set will you use? “Identify the superset of data that drives your credit union,” Harrington says. “You’ll never be done with this—it will be ongoing.”

2. Democratize, or share, your data. “People must have access to your data. If not, it’s a problem.”

3. Make decisions using data. One of Harrington’s colleagues has a mantra about data: “[Forget] your gut.” Experiment with your hunches, but rely on data when making decisions.

4. Put your data back in the domain and repeat the previous steps.

“The idea is to model and use data, and deploy resources against that data,” he says. “Data must be combinable. I should be able to ask questions about the relationships of different data sets interacting to solve real business problems.”

The goal for big data is to make it smaller data. “You know you’re winning when turn your data ocean into a data lake, and then a pond, and then a puddle,” Harrington says. “Chunk it down into operable, bite-sized pieces. The more data you have, the less certainty.”

During the presentation, Harrington took attendees through the “Launch Experience,” an exercise where attendees had to solve a problem—free their disabled starship—by applying data modeling and decisioning principles to the data at hand.

By assessing, communicating, and testing the data, all of the attendees survived the crisis unscathed.

“You don’t have to get fancy with a data model,” Harrington says. “When you have a model, you can start to work.”

► Visit CUNA News for more conference coverage, and get live updates on Twitter via @CUNA_News@cumagazine@CUNACouncils, and by using the #MBDCouncil hashtag. Learn more about the CUNA Marketing & Business Development Council, a member-led professional society for credit union executives, at cunacouncils.org.