Across industries, organizations are investing heavily in data infrastructure and analytics based on AI. Yet, while tools have become more advanced, human capacity to leverage them has lagged behind. Digital maturity essentially depends on building teams with the capacity to think empirically, the capacity to question, interpret, and act on insights. Cultivating such literacy is a problem modern organizations face today.
The question on everyone’s mind among tech leaders today is how to build product teams structurally so that everyone can meaningfully engage with the data that drives their decisions, with or without a technical background. To discuss how product organizations can cultivate these capabilities, we invited Balbodh Chauhan, Senior Product Manager at Smartsheet. Having previously worked on data-driven products at BigTech companies and led transformation initiatives at leading global consulting firms, Chauhan brings a practitioner’s perspective on building data literacy within product teams.
What Data Literacy Really Means in Product Teams
Data literacy for PMs is not strictly SQL or statistical modeling. While technical skills like these are the basic building block of data literacy and are indeed quite helpful, the main point is actually focused on empirical thinking: knowing how to pose answerable questions, read experiment outcomes critically and back up assumptions with quantitative data.
“I consider data literacy in four stages,” explains Chauhan. “First is descriptive literacy: the ability to read dashboards fluently, to spot anomalies in metric trends and to understand what your numbers actually tell you. This is basic, yet I’ve witnessed leaders misread retention curves, confuse correlation and causation, or ignore obvious seasonal trends.”
The second layer is diagnostic literacy: asking why a metric changed and knowing how to proceed methodically. When conversion drops, a data-literate PM does not automatically blame the latest release. They slice by user type, search for external factors, examine the whole funnel and formulate testable hypotheses. Root Cause Analysis – slow, steady, and painful.
There’s also predictive literacy or understanding what can be predicted and what cannot. “At McKinsey’s Digital Practice, where I led transformation initiatives, I saw clients over-investing in overcomplicated predictive models (when will my industrial operation halt) to achieve uncertain prospects and under-investing in simple predictions (how much chemical does this process actually need) that would have generated some real value,” Chauhan notes.
The fourth level, which is prescriptive literacy, is knowing the key metrics to act on. Not all of them are equally important. And they differ based on context. “When working at Smartsheet it was daily audit usage and at Amazon Prime Video it was timeliness of payments. I realized that the most valuable skill is being able to see which metrics really reflect success from the product’s point of view,” Chauhan continues.
These layers translate into tangible advantages. Teams with strong data literacy operate tighter feedback loops because PMs can interpret results without waiting for analyst translations. They test hypotheses faster because they design experiments with clear measurement strategies from the start. Communication will become even more efficient because all product managers, engineers, and data scientists will use a common analytical language. If all members of each product team speak about statistical significance, sampling bias, average versus median, and other analytical concepts, communication will be more efficient.
Developing a Data-Literate Culture
Personal skill development is necessary, but lasting change is achieved through systems design. Successful organizations that develop data literacy incorporate empirical thinking into daily workflow, not solely in one-off training.
A best practice is to hand over metric ownership directly to product managers. Instead of analysts presenting weekly reviews, PMs do it themselves by walking through the numbers, diagnosing changes and interpreting what they’re seeing.
Including data quality checks as part of sprint rituals creates another critical touchpoint. Teams need to check if instrumentation is correct, if metrics align with the hypothesis under test and if the data will actually answer questions they care about before releasing any feature.
Experiment design is another place where ownership can be shifted. PMs must own experimental design completely: defining metrics for success, estimating sample size requirements, detailing possible confounds and determining the analysis plan. “When I was at Prime Video, where we worked with content partners, this shift transformed how teams ran experiments. PMs who authored their own experiments developed much stronger intuitions for statistical power and effect sizes,” Chauhan recalls.
Peer learning mechanisms accelerate literacy within teams. Data office hours, internal guilds for measurement practices and postmortems reviewing analytical reasoning create safe spaces to learn. Leaders who model curiosity and admit their own knowledge gaps create environments where literacy can grow.
From Literacy to Leadership
As AI-driven analytics tools grow more sophisticated, some wonder whether data literacy is still necessary. If natural language interfaces can convert to SQL queries and large language models can interpret charts, why go to the trouble of building human analytical capabilities?
“The answer is that literacy is more important, not less,” Chauhan emphasizes. “Data-literate product managers are strategic multipliers in an AI-augmented world precisely because they can critically evaluate machine-generated insights, guide AI-driven tools to pose relevant questions and interpret between business intuition and algorithmic output. When a generative BI tool spits out an analysis, someone still needs to decide whether it addresses the right question, whether assumptions informing the analysis hold and whether conclusions actually buttress the proposed action.”
From his experience building products across a range of industries, from oil and gas engineering software to digital media sites to enterprise collaboration tools, Chauhan has found that effective product leaders share one trait: they are empirical thinkers. They instinctively convert fuzzy product intuitions into testable hypotheses. They are skeptical of anecdotes not supported by data. They hold both quantitative rigor and qualitative sensitivity in their minds simultaneously, using each to inform the other.
Data-driven leaders can effectively operate when markets shift or competitive contexts change because they know how to separate important data from the outside noise.
The goal of building data literacy is not to make everyone an analyst. It’s to create leaders who think critically, question systematically and guide teams through ambiguity using data as a shared language. When product teams are capable in this manner as a whole, they move faster, decide more confidently and build products that actually solve customer problems rather than optimizing for arbitrary metrics.
The product-data chasm is not an inevitability. It is a business choice that forward-thinking firms are actively reversing. The winning organizations of the future won’t be the ones with the most elaborate analytics infrastructure or largest data science groups. They will be the ones where each product leader can have a meaningful dialogue with the empirical evidence that should guide every strategic decision.
