"The difference between a .300 hitter and a guy at .275 is one hit every two weeks."
“It might be that a reporter, seeing every game that the team plays, could sense that difference over the course of the year if no records were kept, but I doubt it.”
There's a well earned distrust of analytics in lean manufacturing circles, but it would be a mistake to write them off completely. The problem is manufacturing analytics and software that lock us into processes, an idea that's toxic to the continuous improvement ethos behind "lean".
Tools like MRP, finite scheduling, and OEE can have their place, but like any other tool, applied blindly they can lock you into a process you don't fully understand.
Because of that, analytics has gotten a bad name in the lean/Toyota Production System community. We need to think with more nuance though. Analytics can be useful if they help you understand and improve your process, instead of locking you into one.
Here are some examples:
We know from Eli Goldratt's work on the Theory of Constraints that finding your bottleneck is key to making improvements. To oversimplify: improvements to a non-bottlenecked process step have no effect on the improvement of the overall process.
The problem is, spotting the bottleneck isn't easy, especially early in your lean journey. Excess WIP and poor part flow can be incredibly effective at hiding a bottleneck. This is where having good data helps. Seeing where your parts have been queuing up the longest can help you spot the bottleneck, and focus your improvement efforts on the most impactful area.
This one is simple, but often overlooked. How do we know our improvements are working? Can we easily spot if lead times are decreasing or output is increasing?
Most companies have a vague sense of where they are, and certainly know enough to quote a lead time to a customer. But you shouldn't have to scramble to find the data or manually tabulate it. You should be able to see whether your continuous improvement efforts are working at a glance.
As Bill James noted, spotting the better hitter isn't always as easy as it seems. With better analytics in our factories, we have a real opportunity to improve our process, not lock ourselves into them.