Google's People Operations department confronted a paradox. Like most HR organizations, it hired individuals. It developed individuals. It evaluated the performance of individuals.

Yet almost everything Google does, it does in teams.

The group started thinking, "How could HR be different if we reoriented at least some of our services toward the level of the team?" said Brian Welle, Google's director of people analytics, addressing an audience last week at Wharton's People Analytics Conference in Philadelphia. The goal it came up with was characteristically big, hairy and audacious.

The group imagined a future in which People Operations could advise leaders starting projects on exactly what kind of teams they should assemble: both the number and type of people. The advice might go: "You are going to want one extrovert to keep the team excited and motivated," said Welle. "You want two conscientious people to make sure details are attended to. You want three women and two men so you have that diversity represented. You want to be co-located in the first six months, and then you want them distributed in the next six months."

People Operations would then provide the leader with profiles of Google employees who fit the requirements. "I call that the Pokémon approach to team staffing," said Welle. "All we have to do is find the right cards to slot into the team and voilà - an effective team."

Well, that didn't happen. The group studied 180 teams in Google's sales and engineering organizations, and conducted hundreds of interviews. What it found was that the characteristics of individual members don't have a large effect on team performance. What does matter is the dynamics of the team as a whole. Consequently, said Welle, the emphasis should not be on staffing but rather on giving team members insight into how they work together.

Welle described five dynamics that distinguish effective teams at Google. In descending order of importance they are:

1. Psychological safety 

Members feel they can be vulnerable. They know their ideas and opinions will be respected and considered, even when they conflict with those of the rest of the team.

2. Dependability

Members are confident their coworkers will deliver what they are supposed to when they are supposed to.

3. Structure and clarity 

Members understand their roles and the roles of others, and the goals of the team overall.

4. Meaning

Members feel that what they are working on is important to them personally.

5. Impact 

Members believe what they are doing will have a positive effect on the organization and the world.

The data demonstrate psychological safety is critical, said Welle. Google sales teams with the highest level of psychological safety outperformed their revenue targets, on average, by 17%. Those with the lowest psychological safety underperformed, on average, by 19%.

Another takeaway: the effectiveness of teams that were very high in dependability was actually impeded by a lot of structure in terms of role definitions and goals. By contrast, the teams with low dependability benefitted greatly from structure and clarity. That's a useful insight for new teams, where members don't yet know whom they can depend on.

Welle urged audience members to conduct their own in-house experiments. Academic research emerges "from a rather sterile team environment," he says. He would rather see research done "in the messiness of a live organization."