By Lisa Rabasca Roepe, contributor
These days, it’s not just data scientists and analytics teams who are being called upon to read, use, and manipulate data to make important business decisions. Employees in all roles are expected to interpret data, as well as know when to be skeptical of what’s in front of them. They’re being asked to use data to argue a point, tell a story, or predict future trends that could impact the bottom line.
However, according to a recent report from Qlik, a data analysis company, only 24 percent of business decision-makers, from junior managers to C-suite executives, feel confident reading, working with, and analyzing data. Even that number seems inflated to Jordan Morrow, global head of data literacy at Qlik, and author of the report. “We think we are a bit more sure of our ability than we actually are,” he said.
Although companies are trying now more than ever to capitalize on the data they collect, when the workforce isn’t data literate, they are not going to make the best business decisions, he added. For Morrow, while the world does need data scientists, not everyone needs to become one. Instead, “they just need to feel comfortable with data and find areas where they want to learn.”
Here are five data lessons to teach your employees.
1. Democratize Data for Wider Access
According to Morrow, one of the first steps organizations should take to get the most out of their data is to ensure their employees have the right access to it. As a member of The Data Literacy Project, a group of data analytics thought leaders that encourages companies to teach employees how to be data literate, he advocates for a more open data policy.
For example, at DashThis, a marketing analytics firm, data is available to all employees through regular lunch-and-learns, workshops, and presentations in order for every department to understand what is collected, how it’s stored, and how to better use it. “These efforts enable every department to read and use data on their own,” said Marie Lamonde, content marketing and communications specialist at DashThis. “The more employees know about the different data we can use and how it can help their decision-making process, the more new ideas and possibilities have emerged.”
2. Look for the Story
Jim Kyung-Soo Liew, Ph.D., an assistant professor of finance at the Johns Hopkins Carey Business School, believes data is an undervalued asset in many companies. “Companies will collect data and store it, but not really use it,” he said. Because data can be messy and there is no standard way to house it, he continued, it’s not always easy to find patterns.
To turn potential into value, Liew recommends starting with a business problem that the marketing, sales, or operations department faces—a practice he relies on in his day-long course at Hopkins, where business executives learn to extract meaning from big data. Once the problem is identified, he explained, employees can use data to look for a story that shows a relationship between two or more things. For instance, the data might show that sales go up after a football game, so the best place to advertise your product may be at a football stadium or during a football game.
However, Liew warns that you need to validate your story with more than just raw data: “The data needs to reveal a story that makes sense to the people in your industry and your users.”
3. Find Insights, Not Observations
Companies tend to make decisions that can impact the bottom line based on their observations, not insights from data, Morrow noted. For instance, a marketing team might observe that sales increased by 20 percent and assume it was because they ran a Twitter campaign or offered customers a coupon during that same period. That observation could potentially be misused, Morrow continued, if you immediately infer that, because sales increased during the campaign, you should run more campaigns.
“The reality is that it could be one factor out of 20 that influenced the consumer to buy,” Morrow explained. It could be related to a change in the economy, the weather, or the timing of the campaign. Without data, it’s difficult to gain insight as to why customers responded to the campaign. Everything else is just an observation. “The insight is the why behind things, the observation is what happened.” Encourage staff to not just rely on what they observe—that sales increased during the campaign—but to use the data to find out why sales increased.
4. Visualize Data in a User-Friendly Way
The most effective method for employees to understand data is by visualizing it, said Cecilia R. Aragon, director of the Human-Centered Data Science Lab at the University of Washington. “You have to blend the stats with an understanding of how humans perceive data.”
Understanding how to present data—putting context around both the numbers and visuals—is important. “People think it is easy—’just make a bar chart’—but it takes more than that,” Aragon said. Knowing how to draw a graph correctly and understanding when to use a pie chart versus a bar chart is just as important as understanding how to read the data.
5. Beware of Bad Data
Most employees want to believe data is accurate. In reality, data can be biased, misrepresented, or simply incorrect. Experts warn it’s important to teach employees to be skeptical of data that doesn’t make sense or appears to be flawed. “Don’t assume that trend lines will continue indefinitely,” said Drew Farnsworth, a partner at data-center design firm, Green Lane Design. If data appears to be inaccurate, question whether it is biased or calculated incorrectly.
To drive this point home, Farnsworth often gives new employees a sample set of data that has a bug in the program to teach them that data is imperfect “There’s a real tendency when you see a model that gives you a specific number to say, ‘This is right!’—when really it should just be a guideline,” he said.
The Data Road Ahead
Though these data lessons are invaluable, Morrow recommends keeping in mind that every employee doesn’t need to have the same skills when it comes to data literacy, or even the same level of comfort using data. Understand where each employee currently stands regarding his or her comfort level with reading and working with data, he said. Then, build on that individual’s strengths.