The Case for Stack Ranking of Employees

I’ve just read a number of posts about how stack ranking of employees has been ruining Microsoft’s ability to innovate. Stack ranking is one tool that Microsoft is using—perhaps misusing—to identify high and low performers.  Stack ranking means that employees are ranked from best to worst, forcing their manager to make hard choices.  From reading the comments, everyone hates it, including some Microsoft employees.  Read Fred Allen’s The Terrible Management Technique That Cost Microsoft Its Creativity, and Erika Andersen’s The Management Approach Guaranteed to Wreck your Best People.
Absolutely no one is talking about the need to accurately evaluate performance, to reward the high performers and force the low performers to get better or move out. As a veteran CEO, I found that if you don’t shed low performers and encourage high performers, you’ll lose your best and most creative people.
My experience has been with middle market companies ($10 million to $300 million revenues), which are many-fold smaller than Microsoft. In these companies, I most often see the exact opposite problem that’s purportedly plaguing Microsoft:  Employees (including management) aren’t evaluated with sufficient rigor, and low performers are tolerated. These low performers destroy the performance environment by encouraging average performers to slack off. The high performers either leave — they prefer working with other high performers — or become arrogant and hard to manage.
One key to increasing performance is to reduce the range between the lowest and highest performers in a given population. A tight range produces higher performance. For example, Olympic athletes compete to be the best in the world. The range between the gold medal and no medal at the Olympics is very small, and we see amazing performance. This is true in business as well. Much research has been done in this area by Elkiem, a human high performance research firm with which I’m quite familiar.
In a high-performance environment, the lowest performers are made to feel uncomfortable. They get counseling, encouragement, and tough evaluations. This nudges them to perform at higher levels or encourages them to leave. Those who don’t respond well and get fired or quit don’t feel great about their ex-company, and they might write disparaging comments about it. But the company is left with higher average performers and a tighter range.
With the lowest performers gone or improved, new people—who used to be average performers—are now on the low end of the scale, and are encouraged to perform at still higher levels. This will seem harsh to people who don’t want to continuously improve themselves. They should not work for any firm that strives for a high-performance work environment. Likewise, CEOs dedicated to high-performance results should avoid hiring these people.
There are a few caveats.  A company that applies too much pressure will lose good people.  Further, if the metrics for performance aren’t understood and accepted by all, the environment will suffer badly and great people will leave. The reason is that it’s not clear to them what they must do to be successful. The path to winning becomes political—to make friends with the right people. Performance is bound to suffer.
From the posts I read, the stack ranking at Microsoft is political and not based on valid accepted metrics that define performance. But I’m inclined to fault the measurement system more than stack ranking.
So what can middle market companies learn from Microsoft? What these companies need is a firmer, more formal approach to identifying high and low performers, and must treat each appropriately. Because they are much smaller than Microsoft, these companies typically have only 2-4 layers of management. They have less need for big-company, machine-like systems that are overly rigid.
In the end, a CEO who is doing her job will set the example for her management team: that high performance — not politics — is elemental. In turn, they will have a similar caliber of vice presidents and other managers below them. If management keeps its own house in order, it will greatly increase the odds for building a great performance environment that attracts and keeps the best talent.

In my consulting work, I have some CEOs who like everyone in their leadership team and rarely reprimand or fire poor-performing leaders. In these cases, I will occasionally force a CEO to stack rank his leadership team. It forces critical thinking. It forces inquiry about performance and using evaluation tools like 360-degree performance reviews. It can be quite useful.
In my own company (I was CEO for 23 years), executives would at times want to give equal raises to their whole team. I would often force them to stack rank their team, then based on that, try and justify the pay levels of each person. It forced a harder comparative look, and drove better decision-making. Stack ranking is at its heart, an ordering of items, based on a set of criteria. From what I read, at Microsoft the blame may fall more on the absence of clear, accepted, visible criteria.
Mid-market company leaders must unfailingly evaluate people based on clear criteria. They then must reward the high performers and get the low performers to improve or leave.