Connection is the engagement domain most often confused with sociability. It is not how much people like each other — it is whether they can rely on each other, and whether they believe they are accepted as they are.
Connection and belonging is whether people experience genuine teamwork and inclusion, and feel part of something larger than their own task list. Two components again: reliance — whether colleagues can be depended on — and acceptance, whether a person can be themselves without cost.
Both are required and they are not the same. A high-performing team can be entirely reliable and quietly exclusionary; a friendly team can be warm and useless under pressure. Measuring only one produces a misleading picture of the other.
It is also not the same as social activity. Organisations frequently respond to low connection with events, which addresses proximity rather than reliance or acceptance.
Connection is what makes difficulty survivable in a team rather than individually. Shared load is one of the most robust moderators of strain in the occupational health literature, and it operates through exactly this domain.
It also governs whether problems surface. In a low-connection team people conceal difficulty — not from malice but from reasonable self-protection — which means the organisation discovers issues later and more expensively than it needs to.
Connection is scored 0–100 from behavioural statements bounded to a recent window, and it asks about both components separately — reliance and acceptance — because a single score conflates two different problems with different remedies.
The statements describe events rather than feelings. “I can raise concerns without worrying it will be held against me” is answerable from the past fortnight; “I feel I belong” asks for a self-characterisation that many people will supply generously regardless of experience.
Statement design, cadence and reporting thresholds: how the measurement works.
Managers shape this domain mostly through what they permit rather than what they organise. How the first admitted mistake is handled sets the safety level for everyone watching, and it is usually set within weeks of a team forming.
The practical levers are designing genuine interdependence into work rather than parcelling it into individual streams, and distributing information deliberately so access does not track proximity to the manager. Team events are the weakest lever available and the most frequently used.
Structural conditions determine how much any manager can build.
Incentive design decides whether cooperation is rewarded or personally expensive. Working-pattern policy decides whether distributed teams get a deliberate substitute for incidental contact or simply lose it. Team stability decides whether reliance has time to form. Information architecture decides whether inclusion is systemic or discretionary.
An organisation with purely individual incentives will show low connection no matter how collaborative its managers try to be.
Connection relates to retention strongly and to performance indirectly — through whether problems surface early, whether load is shared under pressure, and whether people stay through a difficult period they would otherwise leave during.
In our validation study the Engagement Index moved 48.0 to 97.5 between the bottom and top decile of the Thrive Index, and advocacy moved from 9% to 94% promoters — a 155-point eNPS swing on an independently measured variable.
Its interaction with capacity is worth naming: connection determines whether depletion is shared or hidden. A well-connected team under strain surfaces it; a disconnected one absorbs it individually until someone resigns.
Team research consistently identifies psychological safety as a predictor of error reporting and learning behaviour, and finds it to be a property of teams rather than of individuals. Occupational health research identifies social support as one of the more robust moderators between demand and strain. Inclusion research distinguishes access to information and decisions from representation.
Our domain draws on that literature — 140+ peer-reviewed papers underpin the eight domains. The mechanisms are established; the magnitude in a given team is not, which is what measurement supplies.
Engagement = Purpose + Leadership + Connection + Growth: what is employee engagement, and how the Engagement Index is scored.
Index figures come from thrive at work’s own validation study (n=10,000+, 2025–26), fielded across the US and Europe. Comparisons are bottom decile versus top decile of the Thrive Index unless stated otherwise.
The research described here draws on the published literature in organisational psychology, team research and occupational health. We summarise mechanisms that are well established in that literature rather than claiming novel findings, and the size of any effect in a specific workplace is not established by it. No score is shown for any group of fewer than five people.
A 30-minute call with the research team covers what a first measurement would surface, and how your sector compares.