The ability to collect and process data has increased exponentially in the past 20 years and the cloud is at the core of this dramatic change. Businesses can now store and easily access large amounts of information, and while we have the technological tools to mine this data, what’s often lagging is our understanding of it. To get the most from it, everyone in the organization needs to understand how to access it, keep it secure and think critically about its potential use cases and applications.
What is Data Literacy?
While it’s true that data scientists spend their days crunching and analyzing data, that doesn’t let the rest of the organization off the hook. In fact, data literacy is becoming an important universal skillset. Even the ability to take data from a spreadsheet and create a table is a necessary requirement in many jobs. But the ability to shape that data, present it accurately, and critically assess the value of the information, is less common.
To increase data literacy in your organization, examine each person’s role and to what degree they might require “data education.” Consider these three core components of data literacy: understanding, access and analysis.Understanding. In today’s digitally driven environment, everyone should attain this level of data awareness. It’s about grasping the overall concept of data: where it originates, how to keep it secure and why it’s valuable. It’s also important for people to be familiar with the tools for using and presenting data, like spreadsheets, tables, and visualizations. Without a foundational understanding, employees can’t adequately do their part when it comes to compliance and data integrity.Access. Having data is great, but users need to be able to easily access that data and put it to valuable and acceptable use. They need to understand which data is important to their work, how it is organized and structured and how to use tools to access it in safe and secure ways.Analysis. Ideally, all employees should be able to think critically about data and derive insights from it. In some roles, use of statistical and analytical methodologies may be required. However, the processes involved in analysis should be transparent and explainable to people who aren’t data scientists. Everyone should be able to determine if the analysis is sound. Where is the data from? Are conclusions based on an inadequate or flawed dataset, creating unintentional bias? How, where, what and whose data is collected, sampled, and interpreted has a profound impact on the end results.
The Role of Data Governance
Let’s say your employees are data literate– able to understand, access, and analyze data. Does this mean anyone has permission to access any data? In a word, no. With the immense amount of information available, culled from so many sources, privacy and security must be scrupulously maintained. It is, in fact, an ethical issue. Hence, structuring and controlling the data at every level becomes paramount.
Data governance is how organizations manage, use, and protect data. It encompasses quality of data, maintenance, access, and security. At every stage of its lifecycle, data must be governed. This lifecycle begins with the acquisition of data, and continues through storage, synthesis, usage, publication (through analytics and products), archival, and purge. For example, acquisition is not just about the original provenance of the data but the frequency and reliability of updates. At some point, data must be deleted, too. The organization may lose the rights to it or may not need it anymore. This type of data hygiene is all a part of proper data governance.
With the scope of data governance wide-reaching, to help guide organizations, industry associations like the EDM Council have been created to elevate the practice of data management as a business and operational priority. When it comes to data governance, there are some critical foundational elements:
Establish an internal structure to oversee data. Robust data governance establishes roles and responsibilities at every level of the organization, from enterprise-wide to business units. Typically, data governance policies, standards, processes, and success measures are centralized. Business units are responsible for implementing data governance.
Determine where and in what state the data should reside. Many businesses store data in a “data lake,” composed of different zones depending on the type of data stored. To keep the data secure, organizations create rules about how it’s handled, whether it should be encrypted, if it can be transferred, etc. There may be different rules for a particular state or country. This might affect the ability to transfer data or create a need to protect the data in specific ways. For example, the European Union
has its own set of data privacy rules as do states like California.
Define and enforce who can access the data. This means designating which data can be viewed and by whom. A common practice is to provide access based on zone, type of data and user profile, or permissions based on use case. These permissions are embedded in the code base. It’s also critical to regularly audit history and use. If someone wants access to sensitive data, for example, it should require multiple layers of approval to review the purpose and ensure the data will stay secure.
Evaluate the purpose. Part of the permission process is to ask how the data will be used. For example, is it for research or analysis, or new product development? Some organizations have instituted data ethics boards
to evaluate requests for data access. The board reviews the idea and its potential uses. It can also provide meaningful direction and feedback to ensure data is being used fairly and in compliance with both legal requirements and the company’s own standards.
Make every employee a data steward. Establishing policies and governance structures are critical. But to make the process resonate among employees, all employees must feel responsible and held accountable for how they create, modify, and use data.
The Value of Data Literacy
It’s of no value having data if your people don’t understand it or can’t access it for meaningful applications. For too long, too much organizational data has been stuck in data silos. By being able to share data across the enterprise, new insights can emerge. Marketing needs to know what is happening in the supply chain before it can strategically launch new offerings. HR needs to see all the hiring activity in the company — even contractors hired directly by a business unit — if it’s to properly budget and forecast. Creating a culture of data literacy has the power to inform better business decisions and drive greater outcomes.
