Information you can use is better than data that you don't

2 years ago, stuck at home mid-pandemic and a couple kilos heavier than I’m happy with, I started working out and focusing on my health. Fast forward to today: I’m making a peanut butter sandwich and a guilty thought pops into my head - why do I keep choosing the peanut butter with sugar? Ultimately: It’s tasty, quick, and has a high protein content. No sugar? Not as tasty. I wouldn’t actually want to eat it and it would ultimately lead to me missing my protein-target more frequently. I compromise because it ultimately means I’ll stick to the plan.

We’re building a data product at All Gravy. The data we’re putting in front of people in our first version is pretty basic - and data they can get today. But they don’t: it’s a CSV exported from a workforce system, a janky Excel sheet only one person knows how to use to get the desired output. Maybe they do it twice a year. Across stores it might be multiple systems with incompatible schemas.

I think back to my high-school IT class: “data” gets processed to become “information”. Sure, they have the data, but they’re no better informed.

Does it cover all the nice-to-haves? No. Is it a sensible set of numbers they can easily understand, updated often? Yes. Are you more likely to hit a goal if you can see your performance indicators changing in real time? Of course.

A common sight is too much data, little ability to get the information that matters in front of the right people. Stakeholders have analysis paralysis: if it’s not possible to extract 100% of the value from this data and get all of the answers, it’s not worth doing.

You hear the phrase often in UX and product: perfect is the enemy of good.

It’s no different when it comes to data. Perfect is the enemy of good.

Information you can use is better than data you don’t.