Data doesn't work in companies, I think everyone feels this on some level. One reason I've heard repeated is that it's a people problem, a lack of data culture and data literacy. Companies spend millions on training programs, hire Chief Data Officers, bring in consultants to build their "data culture." And then they wonder why nobody's actually using the data.
You can't become data literate by taking courses or reading documentation. That's like trying to learn to swim by watching YouTube videos on your couch.
When you look closer, it's easy to see something interesting. These same people who supposedly lack data literacy are out there learning TikTok algorithms, figuring out Excel formulas for their personal budgets, tracking their fantasy football stats with frightening precision.
The problem isn't that people can't learn data. It's that we never actually let them touch it.
How do most companies handle data access? Everything goes through the data team. Want to see how customers are using a feature? Submit a ticket. Curious about conversion rates? Wait three days for someone to run a query. Have a follow-up question? Get back in line.
We've built these elaborate temples around our data, with guardians at every gate. And then we're shocked when people don't develop data skills.
A few years back at one of my startups, someone joined us from a business background. Really smart, but zero technical training. At some point she moved into product work and wanted to understand our data better.
My first instinct, honestly, was to say we were too busy to support her. We were a small team, moving fast, and hand-holding someone through data exploration felt like a luxury we couldn't afford.
But then I made a different call. We spun up a read replica of our database just for her. The instructions were simple: play around, you can't break anything, we can't really support you, and if something goes wrong we'll just delete it and start over.
You know what happened? She taught herself SQL. Not perfectly at first. Not elegantly. But she learned. She started understanding our data models, how the product actually worked under the hood, what the numbers really meant.
Within months, she was pulling insights that surprised all of us. She became an incredibly effective product manager, and I'd like to think a big part of that was because she could actually see what was happening in the data.
The lesson here isn't about that one person. It's about what becomes possible when you let people actually engage with data instead of just reading about it.
Real learning happens through iteration. You try something, it doesn't quite work, you try again. You discover limitations and figure out what they mean. You build intuition about what questions are answerable and which ones aren't.
We learned this lesson with software development decades ago. Agile taught us that you can't spec out the perfect solution upfront. You need tight iteration loops with actual users. You need to build, show, learn, adjust.
The same principle applies to data literacy. You can't predict what insights people will need or what questions they'll ask. The only way to find out what's valuable is to let them explore.
But our current data infrastructure makes exploration terrifying or impossible. Access is locked down tight because, well, people might break something or see something they shouldn't. We pre-model everything for specific use cases, which means people can only answer the questions we already thought to ask.
That's not a path to data literacy. That's a path to dependency.
If we can make data fearless – if we can create environments where people can safely play, experiment, and learn – I think we'd see an explosion of actual data literacy. Not the kind you get from a training certificate, but the kind that comes from really understanding what the numbers mean and what they don't.
You can't learn to swim without getting in the water. And you can't become data literate without actually working with data.