Put your seven best people on a problem and you do not automatically get your best team. Sometimes you get seven individuals being polite, hoarding doubts, and quietly waiting for someone else to take the risk. The difference between a working group and a high-performing team is not the calibre of the members, it is the quality of the dynamics between them. The good news is that those dynamics are observable, predictable, and improvable.

The quick version

  • Team dynamics are the patterns of behaviour between people on a team, how they handle conflict, trust, accountability and decisions, and they predict performance better than the individual talent on the roster.
  • Tuckman's model (forming, storming, norming, performing) describes the stages a team moves through over time. The "storming" stage is normal, not a failure.
  • Lencioni's five dysfunctions (absence of trust, fear of conflict, lack of commitment, avoidance of accountability, inattention to results) describe the faults that stall a team, and they stack, so you fix them from the bottom up.
  • Google's large internal study found the single biggest differentiator of effective teams was psychological safety, feeling safe to take an interpersonal risk. That is the foundation both older models were circling.

The idea in depth: stages, faults, and the foundation underneath both

The most-taught map of how teams develop is Bruce Tuckman's, from his 1965 review article "Developmental Sequence in Small Groups" (Psychological Bulletin, Vol. 63, pp. 384–399). Reviewing the literature on small groups, Tuckman proposed four stages: forming (polite, dependent, finding their feet), storming (conflict surfaces, over roles, approach, status), norming (the team settles into shared ways of working), and performing (energy goes to the task, not the friction). He added a fifth, adjourning, with Mary Ann Jensen in 1977 for teams that disband.

The most useful thing this model does is reframe conflict. Storming is not a sign your team is broken; it is a sign your team is forming. So when the disagreement shows up a few weeks into a new team, the job is not to suppress it but to name it: "We're in the messy bit where we work out how we operate, that's expected, let's get the disagreements on the table." A manager who treats storming as failure papers over it and produces a fake calm. A manager who treats it as a stage works through it.

flowchart LR
  A(["Forming
polite, dependent"]) --> B(["Storming
conflict over roles"])
  B --> C(["Norming
shared ways of working"])
  C --> D(["Performing
energy on the task"])
  D -.-> E(["Adjourning
team disbands"])
  C -.->|"new member,
new goal, reorg"| B
Tuckman's stages, and the dotted line back to storming that a reorg or a new joiner can trigger. Leaders Loop

An honest limitation. Tuckman's stages are intuitive and have stuck for sixty years, but the original was a literature synthesis, not a tested longitudinal study, and decades of later work show real teams rarely climb the four steps in a clean line. A historical review by Denise Bonebright, "40 years of storming" (Human Resource Development International, 2010), traces how the model spread among practitioners faster than the evidence justified. Teams loop back to storming when a member leaves or the goal changes; some never reach performing. Use the stages as a vocabulary for what you are seeing, not as a timetable a team is obliged to follow.

If Tuckman maps the stages, Patrick Lencioni maps the faults. In The Five Dysfunctions of a Team (Jossey-Bass, 2002), he stacks five failures into a pyramid, because each one makes the next inevitable. At the base sits absence of trust, people unwilling to be vulnerable, to admit a mistake or ask for help. Without that trust you get fear of conflict (artificial harmony instead of honest debate), which produces lack of commitment (people who never truly bought in), then avoidance of accountability (nobody calls out a peer), and at the top inattention to results (status and ego over the scoreboard). It is, importantly, practitioner wisdom rather than peer-reviewed research, but it is a sharp diagnostic.

flowchart TD
  A(["1 · Absence of trust
won't be vulnerable"]) --> B(["2 · Fear of conflict
artificial harmony"])
  B --> C(["3 · Lack of commitment
no real buy-in"])
  C --> D(["4 · Avoidance of accountability
nobody calls peers out"])
  D --> E(["5 · Inattention to results
ego over the scoreboard"])
Lencioni's pyramid, read bottom-up: each dysfunction rests on the one below it. Fix trust first. Leaders Loop

Because the dysfunctions stack, the order of repair matters, and the temptation is always to start at the top. Resist it. A team missing its numbers (dysfunction five) is rarely fixed by a results dashboard; the rot usually starts at the base, where nobody trusts anyone enough to disagree out loud. Build the trust that makes honest conflict possible, and the upper layers become reachable. Lencioni's own first move is leader-led vulnerability, admitting your own mistake first, because the team takes its permission from you.

Storming is not a sign your team is broken. It is a sign your team is forming.

What the evidence actually points to: psychological safety

Both models are circling the same foundation, and a large modern study put a name on it. Google's Project Aristotle studied 180 of its own teams to find what separated the effective ones. The surprise was that who was on a team mattered far less than how the team worked together. Five dynamics stood out, psychological safety, dependability, structure and clarity, meaning, and impact, and one sat clearly above the rest: psychological safety, the shared belief that the team is safe for interpersonal risk-taking.

That term comes from Harvard's Amy Edmondson, whose research and book The Fearless Organization (Wiley, 2019) defines psychological safety as the belief that you won't be punished or humiliated for speaking up with ideas, questions, concerns or mistakes. Two independent strands, Edmondson's academic work and Google's internal data, landing on the same factor is what makes this load-bearing rather than a single study's quirk. Notice that it is precisely the base of Lencioni's pyramid (vulnerability-based trust) and the thing that lets a team survive Tuckman's storming without splintering.

What to do about it is concrete and small: change how you respond to bad news and dumb questions, because the team reads your reaction far more than your stated values. When someone admits a mistake or asks the "obvious" question, your first job is to make that feel safe, thank them, get curious, never make it a cost. Model it yourself: say "I don't know" and "I got that wrong" out loud. Safety is built in those micro-moments, not at an offsite.

One caution, because this is where the idea gets misused. Safety is not softness. Edmondson is explicit that psychological safety is not about lowering standards or being nice, it pairs with high accountability. A team that is safe but unaccountable is a comfort zone; a team that is accountable but unsafe is an anxiety zone. High performance lives where both are high, and getting only one of them is the common failure.

A worked example

Take a newly merged engineering team, call it the Atlas team, six people from two formerly separate groups. (Illustrative scenario; not a real team.) For the first month it is all surface politeness: meetings end early, everyone agrees, nothing ships late, and nothing hard gets said. A manager reading the room through these models sees it clearly: this is not performing, it is forming, and the agreeable calm is Lencioni's "artificial harmony", fear of conflict resting on an absence of trust between two camps that don't yet know each other.

The wrong move is to celebrate the quiet. The manager instead does three small things. She names the stage out loud ("we're a new team, the friction is coming and that's fine"). She goes first on vulnerability, opening the next retro by owning a call she got wrong, to lower the cost of admitting fault. And she puts the real disagreement on the agenda: which of the two teams' deployment processes Atlas will adopt, a decision both camps had been avoiding. The meeting gets uncomfortable. That discomfort is the storming, finally surfaced where it can be resolved instead of festering.

flowchart TD
  A(["Month 1: surface calm
nothing hard gets said"]) --> B{"Read it right?"}
  B -->|"Mistaken for performing"| C(["Celebrate the quiet
→ harmony stays fake"])
  B -->|"Seen as forming +
fear of conflict"| D(["Name the stage · go first on
vulnerability · table the real decision"])
  D --> E(["Productive storming
→ norming → performing"])
The same quiet team, read two ways, the diagnosis decides the move. Leaders Loop

Three months on, Atlas argues openly in design reviews and ships faster for it, not despite the conflict, but because the conflict now happens in the room instead of in side-channels. The takeaway you can repeat to a colleague: a quiet team is not always a healthy team, and a leader's job is often to make the disagreement safe enough to have.

Frequently asked questions

Is "storming" a bad sign I should stop?

No, it is the opposite. In Tuckman's model storming is a normal, necessary stage where a team works out roles, approaches and pecking order. Conflict suppressed at this point doesn't disappear; it goes underground and resurfaces later as passive resistance. The leadership task is not to prevent storming but to keep it about the work, not the people, and to move the team through it rather than getting stuck.

Which is better, Tuckman or Lencioni?

They answer different questions, so use both. Tuckman tells you where a team is in time, is this a new team finding its feet, or an established one? Lencioni tells you what's broken right now, is the problem trust, conflict, commitment, accountability or focus? Diagnose the stage with Tuckman, diagnose the fault with Lencioni, and fix the fault from the bottom of the pyramid up.

What exactly is psychological safety, isn't it just being nice?

No, and the distinction is the whole point. Psychological safety, as Amy Edmondson defines it, is the belief you won't be punished for speaking up, admitting a mistake, asking a question, challenging a decision. It is about candour, not comfort. It pairs with high standards: the goal is a team where people can say hard things and be held to account, not one where everyone is pleasant and nothing improves.

How do I actually build trust on a team?

Start with how you respond, not with a trust exercise. Trust in Lencioni's sense is vulnerability-based, and it is contagious from the top: when the leader admits a mistake or says "I don't know," the team learns that doing so is safe. Practically, change your reaction to bad news and questions so neither carries a penalty, and make space for the team to know each other as people, not just roles.

My team is small and seems fine, do I need any of this?

"Seems fine" is worth checking, because the most common dysfunction looks like calm. Artificial harmony, everyone agreeing, no hard conversations, reads as healthy but often means people have stopped raising concerns. A quick test: when did someone last openly disagree with you, or admit a mistake without prompting? If you can't remember, the quiet may be costing you more than the occasional argument would.

Related in the Toolkit

How you read a team connects directly to how you choose to lead it: the right approach for a forming team differs from a performing one, which is the heart of leadership styles & models, and the everyday mechanics of running the team, meetings, one-to-ones, decisions, are where these dynamics are actually built or broken, covered in day-to-day people & team management.

Where to go next