The core premise of data democratization is that every one of us can use, interpret, and work with data to make organizational decisions. Yet a recent survey found only 20 percent of business leaders consider themselves data literate and less than 25 percent feel proficient at working with data. Those are surprising numbers because these are people on the front lines of business using data to make decisions and drive future strategy.
As we shift from a world where IT departments solely managed data to our current reality where people who aren’t used to working with data must incorporate those skills, we face a significant, urgent, and daunting challenge around equipping workers with the right tools. Consequently, a debate has begun around when and where we need to introduce data literacy and data science programs throughout the educational process (such as Robot-Proof: Higher Education in the Age of Artificial Intelligence by Joseph E. Aoun, which I recently read on this topic).
(Almost) everything you need to know about data literacy you learned in school
It’s a common fear that being data-literate requires extensive data training. While some jobs indeed demand advanced skills in statistics or algorithm design, most of the fundamental skills needed for working with data are already taught as part of a typical K–12 education—we just don’t necessarily recognize them as such. But in fact, the same skills required to be data literate are developed in classes from music to chemistry to math to history. These include numeracy, ordering, and sequencing facts by dates, reading symbols or patterns, and looking at relationships for possible cause and effect. We learn to write essays to argue a position based on these facts, or learn to follow sequences to play a song. These skills also form the foundation of being able to interpret data.
Unfortunately, many people feel they aren’t sufficiently skilled to ask questions or have nuanced conversations about data in a workplace setting, particularly when confronted by the jargon of business or IT. That’s fundamentally wrong. What trips people up most when working with data tends not to be a lack of skills, but a lack of domain knowledge (or how to apply their skills to that domain), a lack of confidence, or time pressure (where deadlines are prioritized over examining, evaluating, and questioning the data or the conclusions). When we’re in school, we learn the terminology for different domains of knowledge as we encounter them and how to apply our skills to each new domain. Our requirements within an organization for data literacy are no different.
Building a data-literate organization
If you’re trying to build a data-literate, data-confident organization (whether as chief data officer or anyone championing data-driven decision-making and accessibility in your company), the first step is ensuring that everyone has access to the language of the business. A corporate-wide business glossary is one approach that puts the terminology at employees’ fingertips, helping everyone feel confident when discussing issues around data because they share a common reference. Another step is empowering people to learn and understand the problems relevant to their line of business and what data is important, which should be part of a comprehensive, continuous training program that lasts throughout employees’ tenure at the company.
But this should go beyond functional training to helping employees understand context. You need to break down the types of data your company works with and provide insight into where and how the data is acquired and where and how it is used. In trainings, ask guiding questions such as: What policies are relevant? Why is it important to collect certain pieces of information? This helps employees to ask the right questions on their own and grasp what data is relevant—how to sift through it, apply it to a problem, test it out, and identify whether it is of sufficient quality.
Two-way communication is key. We know that active engagement, discussions, and the ability to ask questions (particularly the “wrong” questions) reinforces and expands our knowledge, which is why training initiatives should be interactive. Lunch-and-learn programs, video recordings, and regular demonstrations (such as those common to agile development methodologies) expand the domain knowledge and contribute to a data-literate organization. Topics could range from the basics of finance to the use of the CRM system to reading BI reports to examples of data science explorations. I’ve seen firsthand how planning meetings incorporating design-style approaches can present common data, yet also allow everyone time to question that data (and underlying assumptions) to shape better business decisions that all can agree to.
Building a data-literate organization starts with the realization that your employees already learned the basic skills. Like transposing a song to a new key, what they need to do next is shift foundational skills and practices to new business domains. To do so, they need the right vocabulary and an understanding of what data comprises a given subject area. Through varied, interactive training methods and establishing open, positive communication channels—where questions are encouraged and dialogue takes place without fear of ridicule—you can instill the confidence needed to discover and appraise the data they’re using. And that appraisal is critical if we are to discern what is fit for our specific business purpose, with the right quality and free from bias.