The best data for decision making is data that REJECTS ❌ one or more hypotheses.

We all have biases and we especially have biases when things aren’t going smoothly with something.

The best data forces us to say “well, that can’t be the explanation here, because if that were the explanation the data wouldn’t look like this”

My favorite chart 📊...

Every day, we compute the predicted profitability of each project at its completion. We do this based on the effort to date compared to the progress toward completion to date. If nothing changes about our productivity, then the chart will be a boring flat line. When it goes up, it means some new information has come to light that has allowed us to make progress with less effort or make more progress with the same effort. When it goes down, it means the opposite.

If yesterday’s data was valid, then the up/down today can’t be explained by anything we knew the day before. It must be something new we learned, and I want to know what that is.

👍 This prevents us from attributing it to a lot of factors we would otherwise be tempted to attribute it to.