Measuring team cohesion: the signals to track before things break

A team almost never falls apart in one go. Before your best person resigns or two departments end up in open conflict, there have been months of quieter meetings, decisions sidestepped and small arrangements to avoid working together. The problem is not a lack of signals. It is that nobody tracks them.

Why measure team cohesion when you can feel it?

Most managers say they can “sense” the mood in their team. That instinct is valuable, but it has two flaws. It comes late, because employees often shield their manager from bad news. And it is skewed by proximity: you see the people you have lunch with clearly, but you see far less of the Rabat branch when you are based in Casablanca, or of the field team that rarely comes to head office.

Measuring does not mean reducing everything to a number. When Google studied the effectiveness of its teams in Project Aristotle, the researchers deliberately measured team effectiveness in four different ways: the executives’ evaluation of the team, the team leader’s evaluation, the team members’ own evaluation, and sales performance against targets. Each view captured a different reality. The same logic applies to cohesion. No single indicator is enough: a diagnosis comes from several indicators pointing the same way.

Staff surveys, workplace climate and employee engagement are related ideas, but they are measured across the whole company, once a year. Cohesion plays out in a team of 5 to 15 people, and it moves faster. In the Moroccan subsidiary of an international group, the global survey often arrives too late, and is too aggregated, to help the manager of a single team.

The 4 pillars of team cohesion according to Carron’s model

Search for “the 4 pillars of team cohesion” and you will find lists that change from one site to the next and rarely cite a source. Yet there is a published, measurable framework. In 1985, Albert Carron, Neil Widmeyer and Lawrence Brawley published the Group Environment Questionnaire (GEQ). Their 18-question version proved consistent and reliable from one study to the next, and it rests on four scales.

The tool was designed for sports teams, so it should not be copied as it stands into an office. Its framework, however, transfers very well to a work team. It crosses two axes: what the group shares (integration) and what each member gets out of it (individual attraction), on both the task side and the social side.

This framework avoids the most common misreading: assuming that a team that gets on well is a close-knit team. A team can be strong on the social side and weak on the task side, warm and friendly with goals that nobody owns. That is exactly the situation where a festive team building event changes nothing, as we explain in our article on team spirit at work.

The HR indicators you already have

Start with what costs nothing: the data already sitting in your HR information system or dashboards. These are lagging indicators. They confirm a problem more than they predict it, but they are factual.

• Voluntary departures (turnover). Track them by team, not only at company level. Two departures in a team of eight in one year do not mean the same thing as two departures on a floor of a hundred. Look at who is leaving too: the strongest people often go first, because they have the most options.
• Short-term absence. Repeated absences of one or two days say more about the climate than long periods of sick leave. Compare the team with itself over the last twelve months rather than with a national average.
• Internal transfer requests. A transfer request is sometimes a healthy career move. Several requests from the same team within a few months are often a polite way of leaving a group.
• Escalated conflicts. The number of issues settled by the manager’s own manager or by HR, complaints, reported tensions. A cohesive team resolves most of its disagreements internally.
• Meeting behaviour, where it can be counted. Team meetings cancelled or postponed, attendance rates, chronic lateness, people who attend without ever speaking. These are simple data to note down for a month.

A word on method: never draw a conclusion from a single indicator. A departure may be explained by a house move, a spike in absences by a seasonal epidemic. The signal becomes reliable when three or four indicators move in the same direction.

Weak signals: what shows up in meetings before the numbers

Leading indicators are behavioural. They appear weeks, sometimes months, before the HR figures. To spot them, Patrick Lencioni’s five dysfunctions model is useful because it describes a progression: absence of trust, fear of conflict, lack of commitment, avoidance of accountability and inattention to results. Each level shows up as behaviour you can observe.

• Absence of trust. Nobody says “I don’t know” or “I got it wrong”. Mistakes are discovered late, often by someone outside the team.
• Fear of conflict. Meetings are calm, too calm. The real disagreements get settled afterwards, on the side or in private messages.
• Lack of commitment. Decisions taken in a meeting are reopened the following week. Everyone leaves with their own interpretation.
• Avoidance of accountability. Nobody follows up with a colleague who fails to keep a commitment. Everything goes up to the manager.
• Inattention to results. Individual or departmental indicators take priority over the shared goal.

There is an important nuance in the Moroccan context. In organisations with high power distance, notably some family businesses or very young teams working under an experienced manager, silence in meetings can be a cultural norm rather than a symptom. What matters then is change: a team that used to debate and no longer does, an employee who used to make suggestions and has stopped. Compare the team with itself.

These behaviours often affect alignment too: people can like one another and still pull in different directions. If that is the dominant symptom, our article on how to align your team sets out the approach in detail.

A 10-question team cohesion survey

The third pillar of the diagnosis is the team’s own voice. You do not need to buy a complex tool to get started. A short, anonymous survey scored from 1 (strongly disagree) to 5 (strongly agree) is enough to give you a useful snapshot, provided you repeat it unchanged every three months.

The questions below draw on the four GEQ dimensions and on Amy Edmondson’s work on psychological safety. In her study of 51 work teams in a manufacturing company, she defines it as “a shared belief held by members of a team that the team is safe for interpersonal risk taking”. Google’s Project Aristotle went on to rank it first among the factors behind team effectiveness. The statements below are adaptations, not the original items.

A few rules to get honest answers:

1. Guarantee real anonymity. With fewer than five respondents, do not share detailed results: everyone would work out who answered what. The answers are still personal data, so handle them in line with loi 09-08, Morocco’s personal data protection law.
2. Reverse the scores for questions 4 and 7 before calculating averages (a 5 becomes a 1, a 4 becomes a 2).
3. Announce in advance what you will do with the results. Nothing damages trust more than a survey that nobody hears back about.
4. Have a third party run the survey (HR, a consultant) if the relationship with the manager is part of the problem.

Reading the results: green, amber or red

A score only means something if you know in advance what will trigger action. The table below suggests practical thresholds. They are rules of thumb drawn from experience, not scientific standards: adjust them to your sector, the size of the team and its history.

Two readings deserve particular attention. The first is the spread: an average of 3.8 can hide a team where half answer 5 and the other half answer 2. That is often the sign of two sub-groups, split by site, length of service or job. The second is the gap between dimensions: a team scoring 4.5 on the social side and 2.8 on the task side does not need more socialising. It needs clarity.

To put your results in perspective, there is a regional benchmark for engagement, a concept close to cohesion. According to Gallup, only 14% of employees in the Middle East and North Africa region were engaged in 2025. If your team sits well above that level of commitment, treat it as an asset to protect, not something you can take for granted.

When the signal is red: what to do within 30 days

A red signal is not a sanction against the manager. It is information, and it has arrived before the next resignation. Here is a realistic sequence.

• Share the results with the team without interpreting them on its behalf. Present the figures, then ask: “What do you read in this?” The team often knows the cause better than the table does.
• Hold short one-to-one interviews with each member, especially if the spread is wide. An hour per person is enough to understand what the survey does not say.
• Identify which dimension is slipping in Carron’s framework. That is what determines the action, not the instinct of the moment.
• Choose the right lever. If the problem is the task, clarify priorities, roles and decision rules. If it is the social side, create new opportunities to work and talk together, especially across sites. If it comes from management or a reorganisation that has not been properly absorbed, look at profiles and skills, with nothing off limits.
• Measure again three months later, with the same questions. Without a second measurement, you will never know whether the action worked.

What to avoid: organising a festive outing to “lift morale” when the red is on the task dimension. The team will have a good day, and the same decisions will be reopened the following Monday.

From measurement to action

When the diagnosis points to priorities and to the way the team works together, a cohesion workshop targeted at the gaps you measured lets you work on both at once: the bond between people and a shared reflection on the company’s challenges. When it points to management, a reorganisation or a key profile, an objective assessment of skills and behaviours brings data rather than impressions. That is what we offer with our partner Mercuri Urval and its MU Leader Selection Science® method, certified to ISO 10667.

Measuring cohesion is pointless if the results end up in a drawer. If your first figures worry you, or you simply want an outside view to help you read them, the TakeUp Consulting team can help you make the diagnosis and choose the right lever as part of our support for teams. A first conversation is often enough to see things clearly: let’s talk about your situation.