I build high-performing teams without star performers by moving performance out of people and into the system: a named owner for every critical decision, an operating rhythm the whole team can see, shared measures and norms the team enforces on itself. A team is high-performing when it still delivers in the week its best person is away.
This is written for CEOs, CHROs and boards of companies with a few hundred people and a leadership team that has grown faster than its habits. In those companies the star is usually real. The fixer who rescues every escalation, the executive who knows where everything is buried, the director everybody asks. I am not arguing that stars do not exist. Aguinis and O'Boyle argue that a small number of people contribute a disproportionate share of output, and that individual performance follows a power law rather than a bell curve. My argument is about dependence. When one or two people carry the team, the team has borrowed its strength rather than built it.
Why high-performing teams without star performers are more resilient, not weaker
The strongest evidence I know of comes from a study of star security analysts. Groysberg, Lee and Nanda tracked star analysts who changed employers and found an immediate drop in performance that persisted for at least five years, most pronounced among those who moved alone or to a firm with weaker capabilities. Analysts who moved to firms with better capabilities, or who moved together with other members of their team, showed no significant decline. The authors conclude that firm-specific skills and the capabilities of the firm both play an important part in a star's results. In plain terms, a good share of what looked like individual brilliance lived in the system around the person.
A second study points the same way from a different angle. Oettl looked at what happened to the co-authors of scientists who died. When a highly helpful scientist died, the co-authors' output quality fell. When a highly productive but unhelpful scientist died, their co-authors' output did not change. The star that matters, on this evidence, is the person who makes the people around them better. The loudest problem-solver is not that person, and a team that rewards the loudest problem-solver is training everyone else to wait.
A client puts the distinction plainly on my About page. Chris Cunniffe, who led TMNZ as CEO, wrote: "Stuart moved us from a group of high-performing individuals to a high-performing team." Individuals and teams are different assets. A board can value the first on a résumé. The second can only be seen in what happens when pressure arrives and the individuals are busy elsewhere.
Audit the dependency before you touch anything else
Start with a list of the names people rely on most. For each name, ask what happens if they are absent for two weeks. If the honest answer is panic, you have found a structural weakness rather than a people problem. I use my Leadership Capability Stack to read the result. Team performance is Culture Output, the top layer, and you do not improve it by pulling on culture directly. You improve it by repairing the layers beneath: who has authority to decide (Decision Systems) and what good looks like at each level (Capability Standards).
| Question | A star-dependent team | A system-led team |
|---|---|---|
| Who decides? | Whoever is loudest or most senior in the room | One named owner per critical decision, recorded |
| What happens when the best person is away? | Work waits for their return | The next owner decides; the rhythm continues |
| How is performance seen? | Through the reputation of two or three people | Through shared scorecards covering every role |
| How is conflict handled? | Avoided, then escalated late | Rehearsed in low-stakes settings and resolved in the room |
| What does the team review? | Outcomes after the fact | Its own behaviour: decision speed, follow-through, meeting discipline |
Build the operating system, not the hero
The work is architectural. It covers decision rights, meeting cadence, escalation paths and accountability for outcomes. McKinsey's research on decisions gives a useful frame: it separates big-bet decisions made by top leaders, cross-cutting decisions made in cross-functional forums, and delegated decisions made by the person or team closest to the work. In a survey of more than 1,200 business leaders, fewer than half said decisions in their organisation were timely, and only about a quarter said their organisations made high-quality, speedy delegated decisions (McKinsey, 2023). Those delegated decisions are the most frequent ones. Unclear ownership there taxes every week of a star-dependent team.
McKinsey's own advice is to clarify who has a vote and who has a voice. I would add a rule I give leadership teams: every critical decision has exactly one owner, and every unresolved issue has a named escalation owner within a day. Not a committee, not a shared maybe. If you want the mechanics, I cover them in how to improve executive team decision making and, for AI-assisted calls, in decision rights for AI-assisted decisions. If decisions all drift to one person, the Decision Bottleneck Challenge shows where.
- Audit dependency: List the people the team relies on most and ask what stops if each is absent for two weeks. A panicked answer marks a structural weakness.
- Clarify ownership: Give every recurring issue one owner who knows they are accountable for moving it forward and closing the loop.
- Make norms explicit: Write down how the team works together: how decisions are made, how conflict is handled, what gets escalated. Keep it short, use it in meetings and challenge it when behaviour drifts.
- Rehearse pressure: Use pre-mortems, hard-case reviews and simulations while the cost of getting it wrong is still low.
- Inspect the system: Review the operating rhythm every quarter. If meetings, decisions and escalations are still messy, the team is not scaling. It is only getting busier.
Make performance visible across the whole team
If only the top two people are measured, the rest of the team learns to stay invisible, and mediocrity survives in plain sight. Stronger teams build shared scorecards, peer feedback loops and cross-functional measures that force collaboration, so that a result belongs to a group of named owners rather than to whoever took the credit. Measure outcomes by role, not by reputation. Look at who owns which result, where the handoffs break, and whether the team can deliver without the loudest voice in the room.
Some leaders resist this because they think removing dependence on stars lowers standards. It does the opposite. It raises the standard so the whole team has to meet it. Hybrid and distributed teams feel the pressure more, because visibility cannot rely on proximity. My notes on trust in remote and hybrid settings cover the practical side.
Team norms and the cost of unclear rules
Strategy rarely kills a leadership team. Unwritten rules do. If people do not know how decisions get made, how conflict is handled or what good looks like in the room, they fill the vacuum with politics, and politics is expensive. The fix is a short set of explicit norms that the team lives by, challenges and reviews. A values poster on the wall does none of those three things.
- Name how decisions are made, not only who attends the meeting
- Set response expectations for urgent issues and for routine follow-up
- Agree what gets escalated and what the team solves itself
- Define what respectful challenge looks like in the room
- Review the norms every quarter, not once a year
- Call out a breach quickly, before it becomes the culture
Amy Edmondson's study of 51 work teams in a manufacturing company is the research I return to here. It found that team psychological safety, a shared belief that the team is safe for interpersonal risk-taking, was associated with learning behaviour, and that learning behaviour sat between psychological safety and team performance (Edmondson, 1999). Safe teams admit errors and ask questions, so they correct course faster. A team whose only gear is agreement is practising avoidance, and avoidance always comes due. If you want to go further on this, psychological safety for C-suite teams is the companion piece.
Design for tension before the pressure arrives
Pressure does not create team problems. It exposes them. I watched an $8bn organization fracture from the inside when a restructure promised efficiency but delivered confusion to high-performing teams. Three waves of senior leaders exited within twelve months, not because they lacked capability, but because the cultural trust had eroded. A team that cannot disagree cleanly in a low-stakes meeting will not handle a customer failure, a restructure or a board challenge without damage, so rehearse the tension while the rehearsal is cheap. I worked with a CHRO who introduced pre-mortem sessions before major change programs. Ask the team to name the ways the plan could fail, who will resist it and where the handoffs will break. The conversation that follows is usually more honest than the planning meeting that came before it, and when real pressure lands the team has already walked the hard ground.
Where team interventions go wrong
Offsites that produce good conversation and no change in behaviour are the usual culprit. People talk about trust and alignment and then walk straight back into the same routines. If you want change, alter the conditions that shape behaviour: the meeting structure, the decision rights and what gets rewarded in the room.
The second mistake is treating a system problem as a personality problem. Take an illustrative case, not a client story. The most visible executive on a team is also the least collaborative. Everyone knows it, and nothing changes because the team keeps praising speed over shared delivery. Replace the personal appraisal with a scorecard that measures cross-functional outcomes and the behaviour usually follows, because the incentive moved. The point holds in reverse too. If the same issue keeps returning in different meetings, treat it as a design problem before anyone calls it a communication problem.
Discipline without losing pace
Some leaders hear the word discipline and picture bureaucracy. In my experience discipline is what stops pace turning into expensive rework. McKinsey's guidance is to convene only the meetings that are necessary and to give those meetings short, well-prepared pre-reads. I would add a hard stop once the decision is made. Fewer, sharper meetings with named owners beat a calendar full of updates.
- Keep meetings outcome-based rather than update-based
- Agree a decision rule in advance so debate does not drift
- Separate strategic discussion from operational noise
- Make follow-through visible within the same week
- Protect time for the work that moves the business
The review habit that changes the team
The habit I care about most is regular, honest review of how the team works. I mean a short monthly look at team behaviour, separate from the annual performance review: decision speed, follow-through, quality of conflict and meeting discipline. What is sticking? What is slowing us down? Where are we pretending something is fine? Leaders avoid it because it is uncomfortable, but discomfort is cheaper than drift. One blunt conversation costs less than six months of silent underperformance.
What gets reviewed gets improved. What gets ignored becomes culture.
A star raises a team's ceiling. A system raises its floor. At the scale where a board is watching, the floor is what you are accountable for, and it is the only part of performance you can hand to someone else. If you want that built into the way your leadership team works rather than into one person, start with a conversation, look at the Architecture Accelerator or read about CapabilityAI, which follows the same rhythm as the sessions.