Data analytics: Going for the gold
Batting average is fairly easy to calculate: hits divided by at bats. Earned run average is a tiny bit more complicated: runs allowed divided by innings pitched, multiplied by nine, so that all pitchers could be measured off of the runs allowed each nine innings.
For the first half of my life, and for most of the first century of baseball in America, those were the ways you measured the value of a player. Then baseball data got more sophisticated. Smart people realized that players get walks, and those walks turn into scoring opportunities in the same way hits do. Since walks aren’t calculated into batting average, on-based percentage (OBP) was created, which is calculated as walks plus hits plus hit-by-pitch divided by at bats.
The effectiveness of pitchers also was measured by how many base runners they allowed per inning, which is called walks and hits per innings pitched (WHIP). That calculation is made exactly that way.
The proliferation of sports data has overtaken the game itself in many ways. We now can measure the speed of a wide receiver running down the football field or how many miles (and by what routes) a basketball player runs during the course of a game.
We measure our races in hundredths of a second, and we use stop-action digital photography to determine whether a ball is caught, whether a long-jumper stepped over the line or whether a tennis ball was in or out. The final score almost is anticlimactic. Almost. We still care about the final score, of course. We care about how far they jumped or how fast they skated, but in the end, we still measure everything in wins and losses, in gold, silver and bronze.
The recent Rio Olympics was a lesson in data management we all should pay attention to in manufacturing. If you watched any of the Games of the XXXI Olympiad, you saw not just speed and strength and skill and grace on display but also the amazing new ways we have to measure and manage all of it. We can do the same each day in our own plants.
Last month, I suggested that we not be consumed by data; that we use our native intelligence to evaluate what the data is showing us. That message was for those who believe data has all the answers. Data isn’t an answer in itself, but what it does do is provide us an unfiltered view of our world each day, hour, minute or nanosecond or whatever measure you decided to benchmark from. Data is objective. It doesn’t have an ego, and it doesn’t have any experience. It simply shows you what is happening at any point in time.
String enough of those points together and you have a story. You can see into the immediate past, and it may allow you to look into the future if you care to look ahead. It is this promise of prescriptive maintenance that has all of the manufacturing marketing and sales folks and all of us pundits all atwitter. If we can see the future, we can change it, and that holds great promise for manufacturing.
This manufacturing agility is as daunting as tumbling and flipping along a 4-inch wide balance beam, yet as Simone Biles showed during the Rio Olympics, it can be done with grace, style and power. Our goal is to make it look just that easy. Easy is not, of course, just that easy. It takes practice and thought, and it takes an imagination of the possibilities. We cannot image a different future for manufacturing if we do not embrace the opportunities for change before us.
Data itself does not offer us answers, because there often is no one right answer. Data helps us ask better questions. How we analyze that data, and the new ways offered to help us do this, will point us toward the gold.