Talent Assessments

Consulting

Talent Development

 

Tower and Company

Tower and Company collaborates with you and your team to get people aligned with business objectives. We help diagnose causes of poor performance, and implement relevant, creative, and practical solutions that get results.

Contact Us

It’s Time to Accept the Truth that Most People Are Below Average (Part 2)

Note: This is Part 2 of a 3-part series on rethinking performance in organizations. If you haven’t read Part 1, I suggest that you read that first. Subscribe to our newsletter at the bottom of the post to get Part 3 delivered to your inbox.
sad face on a graph
Sad, but true: employee performance is not normally distributed
 

In Part 1, we discussed how employee performance is best represented by a power distribution, resulting in two general groups: a small number of hyperperformers, and everyone else. However, most companies are still treating employee performance as if it’s normally distributed, and applying some version of Jack Welch’s famous forced rankings performance management system.

In Part 2, we’ll discuss some of the implications of applying a performance management strategy based on the assumption of normal distribution on an employee population whose performance actually follows a power law distribution.


  Why It Matters How We Conceptualize Performance  

Let’s take a look at some of the strategies that managers employ when they assume that performance is normally distributed, and see how they hold up in a power distribution:

 

Strategy: Cater to the Middle

Normal Distribution

Power Distribution

Under a normal distribution, 68% of the people fall within one standard deviation of the mean. In other words, more than two-thirds of your workforce is normal, average, or pretty close. It makes sense to cater to this crowd rather than top performers or underperformers, since they are the bulk of the employee population. So, we’ll cater our performance management systems, compensation, hiring practices, etc. to please and benefit this group.   In a power distribution, catering to the middle means focusing on 80% of people, but it also means focusing on the group that produces only 20% of the results. When we cater to this group by designing policies and structures that appeal to this group, we’re actually rewarding mediocre, below average performers.  
 

A common example of catering to the bottom 80% is rewarding employees for their seniority. School teachers and other unionized employees often get pay raises based on their tenure. Even non-unionized employees are often given preference based on their seniority for things like promotions or scheduling vacations. High performers are demotivated and frustrated by systems that don’t recognize their contributions.

Catering to the bottom 80% also means investing resources like training and coaching in people who are performing below or at standard. There is often an assumption that the high performers are smart enough to figure things out on their own, and besides, it’s not fair to give a leg up to those who are already doing so well by providing additional support and resources. However, from an organizational standpoint, this is a poor allocation of resources. Companies get a better return on their investment when they further develop “A” players over the “C” players.

 

Strategy: “Rank and Yank”

Normal Distribution

Power Distribution

Let’s identify the underperformers and get rid of them. They are a drag on organizational resources and clearly not pulling their weight.   In a power distribution, the majority of employees are “C” players. If the majority of employees are underperforming, it’s harder to figure out who should get the boot. It often ends up to be a political decision.  
 

Forced rankings, where employees’ performance is rated and the bottom 10% must shape up or ship out, is a very difficult strategy to apply in a power distribution because the bottom 10% is often indistinguishable from the bottom 80% or so. Even in jobs in which performance is easy to measure, (for example, sales numbers for salespeople), the difference between the bottom 10% and the next 10% is often so small it’s hard to tell if it is due to real performance differences, or just chance.

It becomes even more problematic to use forced rankings when jobs have more ambiguous standards for performance. It’s easy for the trampoline park to say “performance is defined by getting more memberships” but for more complex jobs, it’s not so easy to define good performance.

Without objective performance criteria, managers end up relying on their “gut feeling” when rating workers, which often doesn’t turn out to be representative of an individual’s performance. And where there’s “gut” decisions in HR, you can be sure there will be lawsuits to follow – a $10.5 million-lesson that Ford Motor Company learned when they settled a class-action lawsuit alleging age discrimination using the forced ranking system.

 

Strategy: Leave the High Performers Alone

Normal Distribution

Power Distribution

Under a normal distribution, hyperperformers are considered outliers, instead of normal, predictable outcomes. The trampoline park employee calls his hyperperforming coworker “a freak of nature,” rather than investigating what she might be doing differently that makes her productivity so much higher than everyone else. She’s effectively written off.     Managers often spend most of their time and energy on problem employees and mediocre performers. Managers tend to be reactive, and these are the squeaky wheels. As a result, they often don’t have time to invest in the high performers. Sometimes, managers don’t even notice very high performers, because things are always running smoothly when they are around.
 

Ever heard the phrase, “it’s lonely at the top?” Hyperperformers at companies that view performance as a normal distribution probably feel this way. They likely receive little to no attention and support from their managers, and may be frustrated that their contributions don’t seem to be recognized, let alone rewarded.    

Clearly, when we assume that performance is normally distributed, it causes problems. Mediocrity is rewarded, some individuals are targeted for political, or at best, arbitrary reasons, and top performers are lonely and frustrated. All in all, it’s a recipe for poor organizational performance.  

In Part 3 we will discuss what to do about it – how to manage performance under a power distribution. Subscribe to the newsletter below to get it delivered to your inbox.  

It’s Time To Accept the Truth that Most People Are Below Average (Part 1)

Note: This is Part 1 of a 3-part series on rethinking performance in organizations. Subscribe to the newsletter to get Parts 2 and 3 delivered to your inbox.

I hope you are sitting down for this, because I have some bad news. We are thinking about employee performance all wrong. And, it’s causing a ton of problems in how we manage people. Bear with me as it gets a little nerdy.

 

The Myth of the Normal Distribution: Most People Are Average Performers

For decades, most people leaders assumed that employee performance is best represented by a normal distribution (or Bell curve), like below:
Most people are below average - bell curve
A Normal Distribution, or Bell Curve
Under this conceptualization, a company has a few under-performers, a few top performers doing a better-than-average job, and mostly a bunch of “normal” performers doing an average job. There’s a few outliers, but most people are average performers. Under the assumption that people’s job performance is distributed normally, we might reasonably take the following actions to manage talent in the organization:
  • Cater to the middle. Under a normal distribution, 68% of the people fall within one standard deviation of the mean. In other words, more than two-thirds of your workforce is normal, average, or pretty close. It makes sense to cater to this crowd rather than top performers or underperformers, since they are the bulk of the employee population. So, we’ll cater our performance management systems, compensation, hiring practices, etc. to please and benefit this group.

  • “Rank and yank.” Let’s identify the underperformers and get rid of them. They are a drag on organizational resources and clearly not pulling their weight.

  • Leave the high performers alone. It’s not fair to give this group additional training, coaching, and other investments, because they’re smart enough and talented enough to figure things out on their own. In addition, these employees are so rare we can’t expect to plan our talent management strategy around them.
In fact, this is precisely the strategy that General Electric pursued when Jack Welch instituted what is arguably the most famous performance management system of all time: forced ranking.

In a forced rankings system, the top 20% are designated as “A” players, the middle 70% are defined as “B” players, and the bottom 10% are “C” players. Once employees are grouped into the buckets, the “A” players are showered with perks, recognition and pay increases, the “B” players are given coaching, training and other investments in the hopes that they too can become “A” players, and the “C” players are, in one way or another, shown the door.

The assumption that most people are average performers is alive and well. Many Fortune 500 clients that I have worked with are still using some form of forced rankings to this day.


The Power Distribution Truth: Most People Are BELOW Average Performers

But does performance really play out in a normal distribution in the real world? Researchers Ernest O’Boyle Jr. and Herman Aguinis set out to answer this question by looking at performance across multiple studies. From their abstract:

“We conducted 5 studies involving 198 samples including 633,263 researchers, entertainers, politicians, and amateur and professional athletes. Results are remarkably consistent across industries, types of jobs, types of performance measures, and time frames and indicate that individual performance is not normally distributed—instead, it follows a Paretian (power law) distribution.” (emphasis mine)

These results eviscerate the model of normal distribution: in over 90% of the samples in the study, performance was best described with a power curve (drawn below) rather than a normal distribution.
Most people are below average - power curve
A Power Distribution
In a power distribution, there is a small group of hyperperformers who account for an incredibly outsized portion of the results, and then there’s everyone else. Hyperperformers don’t just produce a little more than their peers, their productivity is 2X, 10X, or even 100X or more than the typical performer. It’s commonly known as the “80/20 Rule,” where 20% of the people are responsible for 80% of the results.

O’Boyle Jr. and Aguinis refer to these two groups as “the best and the rest.” Rather than three groups (high, average, and low performers), there are actually two groups: the hyperperformers, and everyone else.

Which brings us to a somewhat paradoxical truth about performance: most employees are below average. Sometimes, significantly below average.

Example #1: Sales at the “Shoes of Best Fit” shoe store

Let’s say that you own an imaginary shoe store called “Shoes of Best Fit” and employ five salespeople. Four of them sell $10,000 of shoes, but you have one superstar salesperson who sells $100,000 of shoes.

Production Per Salesperson at “Shoes of Best Fit”

Salesperson A $10,000
Salesperson B $10,000
Salesperson C $10,000
Salesperson D $10,000
Salesperson E $100,000
Average Sales Per Salesperson $28,000

Here’s the same information in a graphical format:

Most people are below average - shoe sales
Production Per Salesperson at “Shoes of Best Fit”
When we calculate the average sales per salesperson ($28,000), we can see that four out of five salespeople are below average performers. The superstar shoe salesperson has destroyed the notion of “normal” sales for a salesperson, and indeed, the majority of salespeople are performing significantly below average.


Example #2: Marathon Race Times

Earlier this month, I was a spectator at the Detroit Free Press race, an event with more than 26,000 registered participants. As an average (read: below average) runner myself, it’s a humbling experience to watch such a power distribution in action:
  1. The first runners finish with plenty of elbow room. Often, one runner dominates the others. (In this year’s Detroit Free Press Marathon, winner Jonathan Mott was nearly five minutes faster than the second-place finisher.)
  2. As time progresses, more and more runners finish, and the course gets more and more crowded.
  3. When the recreational runners are finishing, there are so many people crossing the finish line at the same time that they’re practically shoulder to shoulder.
If, like me, you’ve ever been a spectator at a marathon, you’ve probably come to the conclusion that most people are terrible runners.


Example #3: The trampoline park membership content

Just yesterday, I ran into yet another example of the power distribution in real life. I took my daughter to an indoor trampoline park and chatted with an employee:

TPE (Trampoline Park Employee): Would you like to sign up for our free membership program?

Me: Actually, one of your colleagues already told me about it.

TPE: Oh, ok. We are having a contest to see who can get the most people to sign up. The winner gets $100. But I guess it doesn’t really matter anyway, because the person who is winning has 56 new memberships and the person who is in second place only has 14.

Me: That’s interesting. What do you think the difference is?

TPE: I have no idea how she does it, she’s just way ahead of everyone else. She’s like a freak of nature!

The HPTPE (hyperperforming trampoline park employee) produced four times as many memberships as the next-highest performer. Her performance is a strong outlier, and the performance of the team is clearly best represented by a power distribution.

So what’s the big deal? Why worry too much about replacing one curve for another?  The truth is, we are harming our people, and organizational performance, when we assume that employee performance is normally distributed when it’s actually not.

We’ll discuss the implications of this misconception in Part 2, and what to do about the performance of your team in Part 3. To get these delivered to your inbox, subscribe to the newsletter.