By D.J. Murphy, Senior Editor, Digital Content for Data Universe

The global economy is inexorably moving toward a data-centric future. The collection and analysis of data along with leveraging the insights uncovered by that analysis are driving productivity gains for organizations committed to digital transformation, but at the same time are upending traditional notions of labor and the application of human resources.

“Firms at the technological frontier have broken away from the rest, acquiring dominance in increasingly concentrated markets and capturing the lion’s share of the returns from the new technologies,” Zia Qureshi, director of strategy and operations for The World Bank, wrote in an article for The Brookings Institution.

Qureshi noted productivity growth has stagnated in organizations that have not embraced a data-first mindset. Labor demand has shifted, he said, toward higher-level skills, hurting wages and jobs at the lower end of the skill spectrum.

Realigning and Building Data Talent

The delay in development of those higher-level skills has created the much-discussed talent gap for those searching to fill tech roles. And, while tech companies have vacuumed up a huge share of data science professionals and data analysts, the recent upheaval at those firms, which shed hundreds of thousands of jobs, will benefit more traditional organizations as digital transformation trickles down to them.

“There's a ton of really great tech talent that's been tied up in activities that may not be creating that much value,” Erik Brynjolfsson, the physicist who directs Stanford University’s Digital Economy Lab, told an interviewer in early December. “The real need is to have tech folks go and transform the rest of America, manufacturing, retailing, finance, health care.”

Existing “tech folks” can’t achieve all they need by themselves, though. Businesses will be looking for a new generation of talent that doesn’t exist yet. So far this year, the U.S. economy has produced more than 200,000 new tech jobs, many in “non-tech” sectors such as manufacturing, according to research by CompTIA. To fill those jobs, they will need to be heavily involved in accelerating the data literacy of their existing staff, creating a class of educated employees who can fill those jobs.

Developing Data Literacy

Data Literacy is the ability to read, analyze, understand, and effectively communicate the story of data to inform and enable data-led decisions. Currently, tech leaders feel their employees are not in a position to help them transition to a data-led strategy. According to the Data Maturity Index from data management consultancy Carruthers and Jackson, 64 percent of data leaders polled in a recent study said they believe most or almost all of their employees are not data literate.

“To overcome this challenge, organizations need to create the right environment to train employees to become data literate, not only introducing staff to new terms and concepts but also reinforcing why data knowledge is critical to helping them improve their own department’s operations,” the report’s authors wrote. “By creating an environment where people are given the capacity to learn and fostering a culture that promotes the role of data, organizations can build confidence, and reduce feelings of fear or misgivings about data.”

Creating an environment in which they can effectively address the discomfort some people have with data and communicate how it will enhance employees’ day-to-day work life and value they can provide current and prospective employers is prerequisite to actually implementing a program.

But once the process has begun, organizations interested in boosting the data literacy of their employees must have a clear understanding of the current situation, establish an effective training program, and have the ability to measure the effectiveness of the initiative.

1.       Assess

Building data literacy starts with knowing existing skills and agreeing on the level of proficiency for different job types in your organization. According to IT consultancy Gartner, a thorough baseline assessment will answer these questions:

• How many people in your business can interpret straightforward statistical operations such as correlations or judge averages?
• How many managers are able to construct a business case based on concrete, accurate and relevant numbers?
• How many managers can explain the output of their systems or processes?
• How many data scientists can explain the output of their machine learning algorithms?
• How many of your customers can truly appreciate and internalize the essence of the data you share with them?

2.       Design

When building the actual training program, it’s important for organizations to remember that data literacy is aimed at business lines beyond data science and analytics teams. Using common language understood by all participants at all organizational levels is a must, data science expert and author Piyanka Jain said in an MIT Management Sloan School interview.

“Establish a common way of talking about data throughout the organization,” Jain said. “Analysts collaborating with marketers might work on a project to optimize ROAS (return on ad spend), but if the marketers don’t understand the term, they might not end up using the best insights. Establishing that common vernacular is very important to establishing a culture of data,” Jain said. “It's basically a civilization and its language. If you don't have language, civilizations don't work.”

And, according to Ashley Howard Neville, senior evangelist at Tableau, “literacy” does not mean training someone how to use a specific tool. Less formal discussions and case studies will benefit less technical users.

“The reality is we actually throw people into training way too early,” Neville said in a report on data literacy from Forrester commissioned by Tableau. “Whenever you’re learning something new, you need to build a mind map of concepts and conceptual understanding versus jumping right into how to use the tool. And establishing internal communities actually allows us to start building our data vocabulary in a way that is less intimidating. I can sit and listen while I eat my lunch to someone else talking about how data was impactful in the organization, and I can use that to build my data acumen.” 

3.       Measure

Finally, a successful data literacy program should ensure that specific objectives are met at its conclusion. An organization’s Data Management team can use a combination of quantitative and qualitative measures based on the results of participant surveys. Post-program metrics should answer questions like “How many workers are asking the right questions?”, “How many workers are making informed decisions based on data?” and “How many workers are able to communicate data-driven findings to others?”

Measurement should describe not only technical qualities of a training program but non-technical aspects like employee interest in and attitude toward data, and overall willingness to become data advocates within and outside their peer groups. 

Moving Forward

According to Mark Bates, senior consultant at Carruthers and Jackson, the consultancy’s report shows that many companies are at least engaged in formalizing data roles, which is a step toward creating a culture that rewards building data literacy.

“Positive and engaged people with good data literacy skills drive curiosity, value and business performance creating strong advocates for data who will improve and mature the existing data culture,” he said. “This can then encourage organizations to further define data roles and responsibilities, therefore promoting both a data organizational structure within which data can thrive and a positive approach to valuing it as an asset.”