I guess they measure some KPIs and observe big difference between peers.
So yeah if MSFT can make 3 billion dollars with 1000 engineers, and Google makes 1 billion dollars with 1000 engineers, then 1 MSFT eng is worth 3 of Googles (simplified - obv business involves sales, marketing, etc)
The company with (let's say) identically-skilled employees putting in twice as many hours probably won't be the more profitable of the two.
Replace "mow lawns" with "write pointless, doomed-from-the-start messaging apps" and the actual problem starts to become clear.