Why Are We Still Data Illiterate?
The strange case of non-digestion.
Background
Data literacy as a global movement has around for at least a decade; started in 2014. Compared to the other 2 types of literacy, namely Digital literacy and AI literacy, Data literacy lags significantly. By some estimates, only 25% of the workforce is data literate today. Compare this with the fact that 53% of the entire world’s population is digital literate. Even the new kid on the block, AI literacy, has 29% of the world’s population as active Gen AI users!
My 140th article is a discourse on why data literacy seems to have hit a limit, and how we might rethink the nature of this problem from a first principle perspective, and thereby work towards a more effective resolution.
(I write a weekly series of articles where I challenge or shed new light on the practice of data analytics / data science which you can find here.)
A Quick Comparison
The table below provides us with a simple comparison of the literacy rates of Digital, Data and AI, collectively known as the 3 core skills of the modern knowledge economy. As can be seen, data literacy is well behind.
Data literacy, often called the “second language of business,” is the most uncommon of the 3 core skills in the 2026 workforce. This is despite the fact that data-literate employees command a 26% salary premium. From various surveys, only an average of 16% of the global workforce reports feeling “fully confident” in their ability to read, work with, and present with data. What’s holding them back?
I believe there are 3 fundamental reasons for this ongoing struggle:
- Lack of a utility-based feedback loop.
- It’s improperly taught.
- Not having the courage to impose strong penalties.
Utility-Based Feedback Loop
Digital and AI literacies are anchored on tools, and they have a natural utility feedback loop — use it, get some value, use it some more. In contrast, data literacy is not tool-anchored; it is a cognitive ability. For data literacy to work, it needs to have a utility-based feedback loop. I would argue that this is the most important consideration, as it is the root of self-motivated learning. So, what is the utility of being data literate?
Just as digital and AI are means to ends (i.e. connectivity and productivity, respectively), data is just a means to making better decisions. Good decision-making isn’t always immediately obvious, creating latency in the feedback loop. Enter Gen AI and context engineering (i.e. prompting). We can now use this technique to evaluate the quality of our inputs, interpretations, analyses, and arguments to ascertain whether it would lead to better decisions. It’s not perfect, but it’s a starting place to create a quicker and more consistent feedback loop.
Teaching the Wrong Things
The word “data literacy” evokes maths and statistics. It evokes spreadsheets and charts. If data literacy is about making better decisions, then the focus should be about information and knowledge. And being data literate means having the ability to extract information and knowledge from both structured and unstructured data.
Defining data literacy as business skill doesn’t help. It is a universal skill, and it needs to be taught in school as such. ALL academic disciplines need data literacy because data isn’t just numbers; it’s the underlying symbolic foundation for information and knowledge. Back in Sep 2023, I wrote an article entitled “The Problem with Data Literacy”. I argued then that the issue with data literacy was not recognising that data sensemaking was the foundational cognitive competency required. Data sensemaking is simply the ability to apply a frame or perspective to interpret information from the data, including how various data elements inter-relate to each other. This frame would also allow one to develop various initial hypotheses.
There are many ways to incorporate data sensemaking (and hence, data literacy) into various academic disciplines. For example, in history, teachers can overlay thematic timelines to analyse how with major conflicts and revolutions shape and impact civilisation. Instead of just reading about the Industrial Revolution, students can analyse raw data from ship manifests, census records, or coal production charts to see the human impact of economic shifts. For example, in Literature, teachers can help students understand linguistic shifts and cultural evolution over time through sentiment analysis and frequency of word usage. This encourages the students to analyse texts as data sets to find themes that aren’t obvious through traditional reading.
Strong Penalties
Corporations have long used incentive mechanisms for change management. Despite this, only 42% of global enterprises offer foundational data training to all employees. We need to flip the narrative. My 137th article was entitled “Use AI or Get Fired!”. As the title suggests, it was an argument for strong corporate penalties to enforce adoption of new capabilities. Because inertia is the greatest enemy. And self-motivation only applies to no more than 15% of the population. There should be a corporate agenda for employees to be data literate, barring which they would get fired or career-limited. A simple approach could be to require all employees to document their decision-making or proposal-making process with data (evidence), and to have this incorporated as a standing pillar in their performance appraisal. If an employee is unable to support their arguments with clear data evidence, their proposals are immediately returned for rework or even outright rejected. Of course, for this to work, it must be role-modelled from the top. Alignment must be cascaded. If senior management is not data literate, then it would be terribly unfair to expect it of their subordinates.
Conclusion
In this Age of AI, jobs without decision-making or judgement-making will cease to exist. Data literacy is a necessary cognitive ability to survive being gainfully employed. Data literacy is the foundational skill to critically evaluate AI outputs and understand algorithmic bias. We need to start enabling it much sooner during the formal academic phases and punitively enforce it in the corporate phase.
