We all know what makes a great team: Great people, right?

That’s true—as long as you define “great” correctly. That’s a definition many business owners, and bosses in general, often get wrong.

Years ago I worked in a manufacturing plant where productivity was all-important. We spent significant time and effort working to improve efficiency, reduce waste, reduce downtime... typical improvement initiatives.

As supervisors and managers we also spent a lot of time competing with each other. (Hey, you are what you measure, right?)

One manager decided team performance could be predicted and improved by quantifying the attributes of a great machine operator. He felt that if you could determine the key attributes, and measure potential team members against those attributes, that he could select and create a great team.

I was there when he tried to identify those attributes. During the brainstorming session he filled up 12 easel pad sheets with key skills and attributes.

The problem was, great operators possess a dizzying array of qualities. Many attributes were hard to quantify, like “self starter” and “team player.”

So afterwards he focused on attributes that could be quantified. One was mechanical aptitude. Plenty of tests evaluate and measure mechanical knowledge. And intuitively it made sense: Machine operators run machines, so mechanical knowledge must be important. Off he went, in short order creating a team filled with mechanical aptitude superstars.

Yet my team—most of us with limited mechanical aptitude (based on testing, my mechanical aptitude was the worst)—consistently outran his team by a wide margin.

Where did he go wrong? Faced with too many variables, many of them intangible and hard to quantify, he picked an attribute he could put a number on: mechanical aptitude.

Never mind our plant’s equipment failed less than 4% of the time. Never mind we had skilled machinists who were seconds away if we needed help. Mechanical aptitude could be measured in a way hustle, teamwork, drive, and work ethic could not, even though those qualities were much more important than mechanical aptitude.

So he went with mechanical aptitude because it was something he could “know,” instead of focusing on other qualities that were more difficult to assess.

That’s a simple, and all too common, mistake.

Here's how you can avoid it. The key is to recognize that every employee brings different skills and attitudes, so your goal isn’t to ensure every employee is great; your goal is to ensure that as a team those employees can collectively be great. (There’s a big difference.)

To build a great team:

Decide what key attribute you must have.

Forget about the stereotypically well-rounded employee for a moment. If you could only pick one attribute, what would you choose as the most important skill or quality a great employee needs to have to succeed in the position?

Maybe it's attitude, or interpersonal skills, or teamwork, or a specific skill set... whatever it is, that attribute is the foundation for individual employees and for your team. Training can fill in the gaps, but this is the attribute almost every employee must possess.

Decide what key attribute you can't have.

This one’s easy. Just complete this sentence: "I don't care how great he is, I don’t want him on my team because he…" Typically your answer won’t be skills-based; it will be something like terrible interpersonal skills, a horrible work ethic, or a larger than life ego. Just identify the attribute you can’t live with and make sure it stays off your team.

Determine your threshold point.

You may not be able to build a team where every member possesses your most important attribute. In our case a crew was made up of six operators. We had room for one operator who wasn't quite as fast on job changeovers but was a great leader. (In fact, he could serve as the poster boy for my definition of a remarkable employee.) The rest of us bridged his speed gap and we all benefitted from his leadership skills.

Could we have afforded two operators on the team like him? No, probably not. Decide how many individuals who possess your most important attribute will be enough to make things work. If you can find more, that’s great. If not you’re still okay.

Put together the rest of your puzzle.

Knowing your threshold point frees you up to build a team with complementary skills. You can take on a great team player who is technically weaker, or a loner who is an outstanding problem solver, or a person with limited experience who possesses incredible hustle and drive.

Never assume the only individual attributes that matter are attributes that can be measured. In some cases, when individual contributors work alone and largely outside the scope of a team, quantifiable skills may be all-important.

But where teams are concerned, success is almost always the result of intangible qualities. Focus only on numbers—especially on the wrong numbers—and you build teams that on paper should perform well… but in practice never do.