Revisiting Data Literacy
The needle just isn’t shifting!
Background
Exactly 2 years ago, I wrote an article on The Problem with Data Literacy (article #2). The topic got onto my radar again, triggered by a recent conversation with a friend. She mentioned that one of her clients was still struggling with improving the data literacy of their Sales & Relationship Management unit, despite the organisation having invested in creating interactive dashboards for their use. Her story resonates with many. I’m well aware of many organisations, including clients of mine, who continue to struggle with data literacy even today.
The term “data literacy” was coined by writer Paul Gilster in 1997, and since then, far too many data literacy training programmes have been launched in a great many organisations, and yet, the needle has not moved in proportion.
And so I dedicate my 105th article to revisiting and re-questioning how we are undertaking the task of increasing data literacy within corporate organisations.
(I write a weekly series of articles where I call out bad thinking and bad practices in data analytics / data science which you can find here.)
Data is NOT a Language
I read somewhere that data is a language, and that we need to teach it as such. In fact, the common-held definition of data literacy is the ability to read, write and communicate with data. As with the instruction of languages, it is scaffolded from basic to advanced, covering vocabulary, grammar, style, subjective and nuanced interpretation, etc. And we make a clear distinction between teaching a language versus teaching literature. This perspective has led to data literacy being deeply rooted in data analytics because that’s perceived as the foundational basics. From the curriculum to the instructional methods. We’ve been teaching Data Analytics 101 and calling it data literacy. But data literacy isn’t data analytics.
I think data literacy is simply the wrong term and the wrong focus. All of us are data-driven; no one is making decisions randomly. So we are obviously consuming data in our decisioning process. The challenge is whether those data are complete and representative. The purpose of data literacy training is to enable the corporate citizen to become more sensitive to (additional) information signals, and go beyond the familiar. When a corporate citizen views a dashboard, are they able to see useful information signals; signals that help them anticipate a coming future state, or signals that are counter-intuitive and makes them rethink their worldview? So the training needs to always be domain- and persona-contextualised, and outcome-focused. The focus shouldn’t be on statistics and data types.
The Blame Game
Another observation is that organisations often cite the lack of data literacy as the reason for the low adoption of their data-oriented solutions like dashboards. But that might be a fallacious assumption. Here is a list of possible reasons for the low adoption:
- The data-oriented solution is not appropriate. The metrics don’t provide any additional information signals to what is observably obvious.
- The corporate citizen is unable to interpret the metrics contained within data-oriented solution. This could be due to information overload or poor visual design.
- The corporate citizen is unable to translate the detected information signals into actions because they don’t have the resources or the decision-taking is not within their control.
- The data-oriented solution is not integrated into the corporate citizen’s workflow. For example, the corporate citizen has to “break away” from the flow of their day-to-day operating system to access a separate stand-alone application.
In each of these scenarios, the corporate citizen will not spend too much of their time on the data-oriented solution simply because there is little outcome utility or the inconvenience friction is too high.
Conclusion
Too often in the corporate world, we conflate solution adoption / utilisation with data literacy. We jump to conclusions when adoption is wanting. We don’t do sufficient diagnostics but blame the corporate citizen for their lack of self-motivation and learning discipline. Despite the decades, data literacy is probably misunderstood, and mis-measured. And solving it only gets more urgent each day as the knowledge and AI economy hurtles forward.
