Tips to Avoid Drowning in Data
It is estimated that the world now produces 7.5 septillion bytes of data every day. To put that in perspective, a septillion has 24 zeros. Just a little while back, in 2012, the estimated worldwide data production figure was 7.5 quintillion bytes per day. A quintillion has 18 zeros.
Not all of the data produced today is customer data. In fact, most of it has no corporate value, going instead into predicting weather and traffic patterns, compiling government statistics, producing economic forecasts, and more. But with the Internet of Things (IoT), social media, and other data producers, there is still a vastly growing pool of customer information that is only going to keep expanding.
“Data about customers comes fast, and it comes from everywhere,” confirms Chris Bergh, CEO of DataKitchen, a data and analytics solutions provider.
But having tons of customer data means nothing for companies unless they can turn it into actionable intelligence that they can use to influence or alter the bottom line. That translation might not happen immediately, but it has to happen eventually if data projects are to have any hope of sustaining themselves.
Anjala Krishen, a professor of marketing and international business at the University of Nevada-Las Vegas’s Lee Business School, argues that data collection should never be the end goal. Instead, companies should be able to use the data toward some end goal.
At marketing platform provider RollWorks, for example, the primary goal of data collection is to help marketing teams better identify their known and unknown customers, according to Robin Bordoli, its president.
With all the data available, companies need to be aware of what they are collecting and why, according to Krishen.
“There is a fire hydrant of data,” she says. “There’s a lot of hype about the new ways to get data. But companies are storing data that they will never use. They are storing some data that is not actionable. You have to limit what you source.”
“Just because you have a lot of data doesn’t mean that you have a lot of value,” adds Chris Nicholson, CEO of Skymind, a business intelligence and enterprise software firm. “A lot of data isn’t useful for what you want to predict.”
Experts agree that companies are collecting too much extraneous data that will not lead to good, actionable intelligence, and doing so only adds to costs—for wrong conclusions, slower processing (the more data, the longer it takes to process), and for storage.
One example of collected data that typically doesn’t get used or doesn’t provide a lot of value, according to Krishen, is beacon data, which many retailers collect in their brick-and-mortar stores. While retailers try to connect store traffic data to sales, it’s a very inexact science: Some customers might walk the aisles with little or no intent to make a purchase that day, while others go in with a specific need and plan to buy only that one item.
Additionally, there will be times when customers queue at certain locations (like in front of an Apple store before the release of the newest iPhone) that can provide positive word of mouth that goes far beyond just collecting the number of people in line, according to Krishen. “Crowds attract crowds.”
Nicholson adds that there needs to be a specific purpose for data collection. “Define what you want to find out before you start collecting the data,” he says, pointing out that the goal can’t be as generic as merely wanting to sell more than the competition. “Instead, data should be collected for more defined reasons, like reducing customer churn or increasing the size of customer shopping baskets in e-commerce,” he says.