Bad Metrics

Businesses trying to step up their game need measures of success or failure, or "metrics."  Getting the metrics right is important (though as always, don't go overboard looking for perfection).

McKinsey Quarterly just published a new article proposing profit per employee as more appropriate for knowledge-based companies than traditional return-on-capital measures.  Sounds fair, but these ratios can produce truly stupid actions.  And believe me, I've seen companies do pretty stupid things because of bad metrics.

Let's say that a division is evaluated on its profit per employee.  It's 25 employees generated about $500,000 in profits, or $20,000 per per person.  This is about twice the national average.  Now suppose that the team leaders point out that if they could offload some clerical tasks to a new, entry level employee, they could bring in an extra $10,000 in profits, even considering the new employee's salary, benefits, equipment, training, and occupancy costs.  An extra $10,000 is profits is good.  But what happens to profit per employee?  Total profits are $510,000, headcount is 26, profit per employee is $19,615.  Profit per employee has fallen.

Is the company better or worse off with the new employee?  More profits are better, certainly.  But if company is managing to maximize profit per employee, it fore-goes a perfectly good opportunity.

This metric may have some value in understanding how two different companies in an industry differ in their strategies.  But don't manage to it.