The Wrong Proxy

Here's an exercise.

A company receives a million applications a year. It hires eight to ten thousand.

Only one in six of those hires will reach the top.

So you build a filter. You look for the best grades, the best schools, the most polished credentials. You narrow a million people to 500 pathways that lead to your door.

For twenty years, nobody questions it.

Then someone does. Bob Sternfels, McKinsey's Global Managing Partner, asked his team to use AI to analyze two decades of internal data. The question was simple: what actually predicts who makes partner?

The answer had nothing to do with grades.

Resilience. Recovery from setbacks. The ability to learn new things instead of having mastered familiar ones.

Experience working retail. Playing team sports.

The traits hardest to put on a resume were the ones that mattered most. The traits easiest to measure were the ones they'd been optimizing for. They were the wrong signal.

What matters most resists measurement.

What gets measured gets optimized. The distance between those two truths is where organizations lose what they care about most.

McKinsey had built the world's most sophisticated filter. It filtered out exactly what it was looking for.

Go deeper: Dashboard Blindness