"People don't know the right questions to ask from their data" is something I've heard countless "data people" tell me.
Yet, I've never found it to be true.
Sure, the non-technical people on your team might not know the exact right question to ask to immediately unlock massive insights. But neither does the data team!
Working with data is an exploratory process. It's one that requires playing around and trying different things. It's iterative.
For data scientists or analysts working in tools like Jupyter notebooks, their first "question" is almost always something like df.head(), to just get a sense for what they're even looking at.
Through iteration, they build up an intuition about what's going on – in the data, and in their business. They bring in different datapoints, and they combine stuff. They backtrack, and they change course. They write hack-y Python and throw most of it away. Sometimes, magic happens – and they stumble on something insightful.
But we expect their non-technical friends to just blurt out perfect questions. They have one shot at this, and they'll get something back several days or weeks later – whenever the data team gets around to it. They aren't granted access to the magic of iteration & experimentation.
They're just forced to play on hard mode.
So when someone claims that their "non-technical colleagues wouldn't be able to ask the right questions anyways" – I think they're being unfair and wrong.
Worst of all, they're hurting their companies' odds of uncovering actual nuggets of gold in the data.
Shameless plug: I founded Supersimple (a self-service business intelligence tool) a few years ago because of this. For everyone to be able to get in on this magic. To give everyone the data and facts they need to do their best work.
