The best way to evaluate top and low performers during annual reviews, according to a Kellogg MBA professor

• Professor George Georgiadis is an associate professor of strategy at the Kellogg School of Management.
• He describes the pros and cons of relative (or stacked) and absolute performance metrics.
• Ultimately, Georgiadis advises leaders to build their incentive structures explicitly around their priorities.

What is the best way to measure an employee's performance?

Performance evaluations are incredibly important for addressing performance issues, rewarding excellence, and encouraging everyone to put in maximal effort. But leaders face a common conundrum: Should they compare members of their team relative to one another or evaluate everyone independently?

Each option — relative (or stacked) and absolute performance metrics — has its advantages in gauging results and improving performance. Each also has some major downsides.

In considering which option could be a better motivator for your team, you will want to distinguish between two layers of motivation, says George Georgiadis, an associate professor of strategy at the Kellogg School. The first is the degree of motivation: How motivated are your employees? The second is whether your employees are motivated to do what your company wants them to actually do.

"When you're deciding how to motivate employees, you have to be careful not to distort incentives," Georgiadis says. "If I create a very competitive environment, for example, I risk killing collaboration. In fact, it might even lead to employees sabotaging each other."

Here, Georgiadis describes several ways a team can benefit from relative performance metrics — as well as when absolute or holistic performance metrics may be a better motivational tool.

What are the advantages of relative rankings?

Relative or stack ranking systems have long been popular. For one, they are often relatively simple to execute: you can stack individuals in similar roles against one another and see who performs best. They also tend to provide stronger incentives for less money. "If I am forcing my team to compete with each other, I'm raising the bar for all of them," Georgiadis says.

Additionally, relative rankings are less risky for employees, particularly when it comes to factors outside of their control. For example, imagine a pandemic hits and your sales team can't make their targets. If they are rewarded based on absolute sales, everyone's compensation will be hit hard for reasons that have little to do with individual effort or talent. "But if they are ranked relative to their peers, this unexpected shock will filter out," Georgiadis says.

Finally, relative rankings provide companies with a useful buffer against uncertainty about what constitutes good performance. Since any business operates under dynamic conditions, it can be difficult to know in advance what might help or hurt employees' ability to execute or how a new product might fare in the market.

"If your salespeople are operating in a new territory or selling a new product, there's a lot of uncertainty and you don't know exactly what to expect," Georgiadis says. "But if you reward people relative to each other, then those who sell the most will earn more."

What are the disadvantages of relative rankings?

Relative ranking systems, which were often based on objective, quantitative criteria such as sales or productivity, had their heyday in the late 20th century at large corporations like GE and Microsoft. Today, however, many organizations are shifting away from this approach in favor of more holistic reviews.

For one, the internal competition associated with relative ranking is not for everyone. It can push out otherwise excellent employees who are averse to the dog-eat-dog culture that can grow out of stack ranking systems. It can also threaten teamwork.

"Microsoft was known for having a stack ranking system and a very, very negative culture," Georgiadis says. "Part of that was due to everyone knowing that if their colleagues did well, it would hurt their own rankings. So they had little incentive to help — and an incentive to sabotage — their coworkers."

This fear of others getting ahead can kill collaboration within the organization, which is what led Microsoft to ditch its stack ranking system in favor of more-nuanced evaluations.

It is possible to take this tendency into account when designing a relative ranking system based on objective criteria. For instance, some companies rate employees on both their own individual performance and their broader team's performance. But this can cause its own problems, leading to unbalanced team compositions where top performers gravitate to one another.

"The problem with this assorted matching is that workers often learn from each other," Georgiadis says. "So if you have all the stars on one team, there's not going to be much learning among the stars because they're well-established, and everyone else will not have the stars to learn from."

How to get the benefits of both (and why that's so hard)

Is it possible to get the best of both worlds: efficiency and competition as well as cooperation and mentorship?

Georgiadis points out that relative and absolute ranking systems can be used effectively in concert. Even in isolation, stack ranking systems don't "necessarily kill collaboration, provided you incorporate collaboration into the criteria," he says.

For example, you could use a relative ranking system to evaluate an individual's objective performance, but also "based on their abilities and skills when working in a team." That is, rather than pitting entire teams against one another, the evaluation would rate individual team members in terms of their contributions to the collective good, including the development of other team members.

"There's this notion that you can have a healthy level of competition and still foster collaboration," Georgiadis says. "Having a more holistic performance evaluation is important."

One complication with this method is that it requires you develop subjective evaluation criteria. Unlike comparing sales numbers between team members, collaboration doesn't have the same objective metrics. So you often have to rely on feedback, peer reviews, and the evaluations and recommendations of managers.

"As soon as you introduce subjective evaluations, you have other issues like trust and influence to worry about," Georgiadis says.

Ultimately, Georgiadis advises leaders to build their incentive structures explicitly around their own priorities.

"Ask yourself what you care about as a manager, what you want your team to do exactly, and then pay them to do that. If I want collaboration, for example, then I should incentivize collaboration. One way to do that is to pay them based on how they are collaborating."