Data scientists might be in demand, but data literacy starts with leaders. Leaders need to trust and understand data well enough to make good decisions, and they must also drive literacy efforts throughout the organization and create a culture of trust in data.
Data literacy — the ability to work with and understand data to drive business impact — “is as essential a skill as negotiation, communication, management, people management, all of that,” said Piyanka Jain, president and CEO of Aryng, a data science consulting company. Executives need to lead with numbers to gain competitive advantage, she said. “Your competitors are.”
Historically, business school programs have offered courses on data modeling, querying, and architecture, but rarely look at broader data strategy, said Barbara Wixom, a principal research scientist at the MIT Center for Information Systems Research. “Leaders have to understand and appreciate that void, because it puts even more pressure on them to be literate and drive literacy in their organizations,” she said.
Here’s what leaders need to know about data literacy — from how to establish trust to being appropriately skeptical — and steps to getting there.
Trusting data and promoting its use
Modern leadership skills include trusting in data to guide decisions and knowing when to question results.
Trusting in data can be difficult for leaders, Jain said, because they are good at decision-making but often rely on their intuition. Asking someone to give away some power and accept guidance from analytics is hard. But pairing judgment and intuition with insights from data makes leaders stronger.
Leaders aren’t necessarily involved in creating or analyzing data, but they often make decisions based on analytics. “The goal for a leader, from a data literacy perspective, should be, ‘How can I be a fast but effective consumer of analysis that is produced by my organization?’” said MIT Sloan professor of the practice formerly a senior vice president at Salesforce.
Leaders also are responsible for establishing data literacy in their organizations.
Important steps include defining what data literacy means for your company and for different roles, establishing a baseline of data literacy skills; building a culture of curiosity around using data, and defining success. (For more details, see “How to build data literacy in your company.”)
It’s also important to establish purposeful data vocabulary beyond buzzwords, Wixom said.
Data literacy requires “top leaders committing to not just really learning the language but then using the language in understandable, consumable ways so that everyone across the organization will get on board,” she said. Wixom said her research has found that it’s particularly important to distinguish between data and a data asset, which has been purposely prepared for future value creation.
Understanding your firm’s data processes
Data literacy requires hands-on, ongoing effort. It’s not a matter of completing one training session, Jain said — it’s a fundamental shift in thinking.
Changing how leaders make decisions takes time. But it’s faster and easier with a systematic approach. “If you think about data literacy as training and a checkbox, you will waste your time and your money, and you will waste the time of your organization,” Jain said. Creating new business value through data and defining what data literacy means to leaders and their organizations should be part of the process. “It’s a commitment to long-term learning and changes,” she said.
While reading case studies is a good way to learn about what goes into data initiatives, it’s hard to understand data, and data terms, unless you’ve worked with it, Wixom said. She recommended that leaders get involved with data projects. “Data is such an active kind of concept,” she said. Talking about data quality can be vague but trying to work with or remediate a bad data set helps users appreciate why it matters and what’s needed to overcome challenges.
Wixom suggested that leaders focus on answering three main questions:
How does a data initiative lead to actual financial returns? Data monetization — how data leads to economic returns — is a vital component of a data strategy. “Being familiar with examples of initiatives and how they create value in different ways — I think that’s really important for literacy for a leader,” Wixom said. “If you’re trying to establish a strategy that involves some collection of initiatives, you need to really understand those initiatives and how they contribute to your firm’s financial performance.”
What are the actual practices behind creating data assets? Leaders need to understand the capabilities that support data strategy — things like master data management, metadata, or data catalogs — at a foundational level. “This is stuff that’s very tangible,” Wixom said. “When you’re talking about making investments in these types of practices, leaders need to really pay attention and hear what's involved.”
How are people learning about data and from data? Leaders “have to really appreciate how they're growing the organization in its data literacy,” Wixom said. Consider establishing centers of excellence or embedding data experts in the company
Being savvy consumers of company data
At the same time, leaders need to be able to evaluate data and be skeptical when appropriate.
Leaders should keep the following in mind:
- Before being shown data, think about what you expect to see. That way, “the contrast between what you expected and what is actually showing up will just jump out at you,” Ramakrishnan said. This will often quickly highlight the most relevant parts of a report. “This is a habit worth cultivating, and you’ll get better at it with practice,” he said.
- Remember that data is uncertain. There is almost always a degree of uncertainty in data, Ramakrishnan said. He cited statistician John Tukey, who said “The combination of some data and an aching desire for an answer does not ensure that a reasonable answer can be extracted from a given body of data." Leaders will need to live with this uncertainty, he said, or ask their team to get more data to provide more information and reduce uncertainty.
- Use the “common sense” test. If you’re making an important decision based on data, put it to the test. “If something’s true … chances are, different data paths lead the same truth,” Ramakrishnan said. “If you’re about to make an important decision based on one analysis, try your best to get another team or another data set to be analyzed to see if it points in the same direction.”
- Don’t confuse causation and correlation. “Whenever you're looking at an analysis that suggests that some factor is driving some outcome, always try to figure out how much of it is correlational and how much of it is causal,” Ramakrishnan said. “They should really brainstorm with the management team on what else could explain this alleged cause-and-effect relationship.”