How to Improve Teamwork — and Which Lever to Pull First

To improve teamwork, work on five things in roughly the order they depend on each other: clear roles, trust, an agreed definition of what success looks like, disagreement that actually gets said out loud, and people who hold each other to what they promised. Most attempts fail because they start at the end of that list.

You can usually feel the problem before you can name it. The meeting where nobody objects and then nothing moves. The handoff that lands three days late with no explanation, and nobody mentions it. Almost all of the advice you find first is written for whoever runs the team — so the more useful question is which of these five you can change when that isn’t you.

The five parts of teamwork, and why the order matters

Teamwork isn’t a single quality a group either has or lacks. It’s five separable behaviors, and the research keeps landing on roughly the same set. Google’s Project Aristotle, the company’s internal study of its own teams, identified five factors behind the effective ones — psychological safety, dependability, structure and clarity, meaning, and impact — with psychological safety the most important by a clear margin. The academic model from Salas and colleagues, known as the “Big Five” of teamwork, cuts it slightly differently: team leadership, mutual performance monitoring, backup behavior, adaptability, and team orientation, held together by shared mental models, mutual trust, and closed-loop communication.

The most useful version for someone trying to fix a team is Patrick Lencioni’s, because it’s the one that puts the parts in sequence. In his model the failures stack: an absence of trust makes people avoid conflict; avoided conflict produces decisions nobody genuinely committed to; uncommitted people don’t hold each other accountable; and a team that won’t hold itself accountable stops paying attention to results at all. Run that forward and it reads as instructions. It also explains why the single most repeated tip on this topic — hold each other accountable — so often bounces off. Applied to a team with no trust, accountability doesn’t land as accountability. It lands as an attack.

Clear roles and complementary contribution

Knowing exactly what you own, what other people own, and where the team currently needs you. It’s the only one of the five that is structural rather than relational, which makes it the cheapest to fix: one explicit conversation about who does what can settle it, and nobody has to feel differently about anyone afterward. Setting clear expectations and defining the outcomes the team is aiming at is the most repeated non-trust recommendation across the popular guides, and it matches Project Aristotle’s “structure and clarity” factor — two independent lines pointing at the same lever.

Gallup adds the part most teams skip: fit. When people work in roles matched to their actual strengths, engagement and confidence rise, and the effect spreads through the group. You don’t need authority to use that. Noticing out loud what a teammate is genuinely good at, and steering the relevant work toward them, is available to any member.

Trust and psychological safety

The state where someone can admit an error, ask the obvious question, or raise a concern without first calculating the cost. It’s the factor Project Aristotle ranked first and the base of Lencioni’s stack, which is why so little else holds without it.

It’s also widely misread. Psychological safety is not comfort, niceness, or the absence of conflict — the sources on it go out of their way to say so, because high standards and high safety are meant to coexist. The difference is that accountability happens inside a culture of trust rather than one of fear. A team where nobody argues isn’t necessarily safe. It may just be quiet.

And it isn’t built by announcing it. It comes from ordinary consistency: meaning what you say, doing what you committed to by the time you said, and letting other people’s mistakes go rather than filing them away. All three are available to anyone, starting immediately.

A shared purpose that outranks individual agendas

Orienting around what the group is collectively trying to achieve instead of optimizing your own slice of it: coordinating rather than working in parallel, sharing what you know rather than holding it, giving credit to the team rather than collecting it. Project Aristotle’s last two factors, meaning and impact, are the two halves of this — whether the work matters to you personally, and whether it matters to anyone outside the room.

This is the dimension that erodes silently. Nobody argues against a common purpose; it just loses to individual incentives, one reasonable decision at a time. The symptom is a team where everyone is busy, everyone is technically doing their job, and the actual goal is nobody’s problem.

Disagreement that gets raised instead of managed

Engaging the disagreements the work genuinely requires, keeping them on the topic rather than the person, and then committing to whatever the team decides — including when you argued the other way.

This is the dimension most often mistaken for the opposite of teamwork, and the mistake is expensive. Both Lencioni’s model and the psychological-safety literature treat the absence of conflict as the failure mode, not its presence. If your team never disagrees, that usually isn’t alignment; it’s an earlier problem showing up as silence. And backing a decision you lost is not a compromise of principle — it’s the mechanism that lets a team decide anything at all.

Holding each other to commitments

Saying something when a teammate misses what they promised: directly, quickly, specifically — and being someone others can say it to.

It comes last because it depends on the other four being at least partly in place. Attempted without trust, it reads as an accusation. Attempted without clear roles, it’s arguable, because nobody ever agreed whose job it was. Timing does most of the work here. “You said Thursday and it’s Monday — what happened?” is a small, answerable question at three days and a grievance at three weeks.

Why team-building events rarely improve teamwork on their own

Escape rooms, quizzes, seminars, afterwork meetups: team-building activities are among the most prominent recommendations in the commercial guides on this topic, and none of the ones making the recommendation ties it to a measured, durable change in how a team functions. That doesn’t make them worthless — people who have enjoyed an afternoon together do tend to find the next awkward conversation slightly easier. But the mechanism is thin, and part of why the advice is everywhere is that it’s the one intervention that can be purchased.

The five dimensions above all describe what happens on an ordinary Tuesday: who owns what, whether the difficult thing gets said, whether Thursday’s commitment arrives on Thursday. An event doesn’t reach any of that by Monday morning.

Where to start when you don’t run the team

Belbin’s guidance is the rare advice on this topic that begins with the individual rather than the group: evaluate what you personally bring, ask other people for feedback on it, then watch your teammates’ strengths and development areas and work out which gap you’re positioned to plug. It’s an unglamorous sequence, and it’s close to the only one available to someone with no authority to redesign anything.

If you want a single place to begin, begin with dependability. It was Project Aristotle’s second factor and it’s the entry-level behavior across every individually focused source: do what you said by when you said, and when you drop something, say so early and move on. It’s the fastest signal to send, and everything else gets built on top of it.

The hard part is that you can’t see your own contribution from the inside — which is exactly why Belbin’s sequence starts with asking. If opening with a round of feedback requests feels like a lot, checking your own collaboration habits gives you a similar read in private first, and what shows up weak is a habit rather than a fixed trait, so it’s practiceable.

The skills a functioning team actually runs on

Look back at those five dimensions and notice what none of them is: a property of a team. Each one is a behavior, performed by a person, that a group afterward gets described as having. Three of those behaviors carry most of the load.

Teamwork is the direct one. It’s the habit of putting the team’s common purpose ahead of your own scoreboard — coordinating rather than optimizing your own piece, sharing what you know, being excellent at your role while respecting other people’s. It’s where trust actually gets produced, in the specific form of being sincere, being reliable, and forgiving mistakes rather than remembering them. And it covers the two hardest moves on this page: disagreeing when the work requires it without making it personal, and holding a teammate to a commitment without delaying until it’s a grievance.

Communication is what makes any of that visible to other people. Not the written mechanics, but the choices — stating your main point first, being clear and brief, actually listening rather than waiting, moderating how long you talk, and picking a conversation over a written thread when the subject is sensitive or needs working through. It also covers the specific moments a struggling team keeps failing at: expressing disagreement, giving feedback, and apologizing cleanly when you’ve dropped something.

Influence is the one that answers the question the rest of this article keeps running into. Improving a team you don’t lead is a problem of getting and applying influence, and it has a method: build a reputation by delivering consistently, take responsibility for more than your own role, and take the initiative by setting the agenda — volunteering to prepare, write, or organize the thing nobody has picked up. It also supplies the preparation step, working out what’s in it for the other person before you propose a change they didn’t ask for.

None of the three is a personality setting; each is behavior, and a gap in any of them is something not practiced yet rather than something you are. The free Job Skills Test scores where each of yours currently sits, along with the other nine skills it measures, and that’s worth doing before picking which dimension to work on — because the bottleneck usually turns out to be your weakest of the three rather than anything about your team.

Where your own habits actually sit

Some of this probably describes things you already do — flagging a slip before anyone has to ask about it, or saying the awkward thing in the room rather than in the corridor afterward. The rest is practice. None of it asks you to become a different person at work; a gap here is simply something not learned yet, and the ones worth naming first are whichever matter most for the team you’re on right now. It’s also the kind of ability that gets heavier as you go: the more your work depends on other people, the more of your week is spent doing exactly this. And you’ve already done the part most people skip, which is reading about what you contribute instead of assembling the case against your teammates.

Which leaves one thing to settle. The free Job Skills Test is a short self-assessment of your work skills: it scores you across all twelve, including the teamwork, communication, and influence behind everything above, and shows which of them would make the biggest difference to the team you’re in now. It takes about seven minutes, and you finish with a specific place to start rather than a general intention to be a better teammate. You can see where yours stand right now.