Unlocking employee commitment and performance
The link between employee engagement and business outcomes is one of the best-evidenced findings in applied research — and one of the most misapplied.
Employee engagement research has an unusual property for a management idea: the evidence base is enormous and the effect is consistent. The meta-analytic work now spans on the order of 183,000 business units drawn from several hundred separate studies, and the relationship between team-level engagement and performance holds across industries and geographies.
The outcomes it relates to are the ones that appear in an operating review — customer ratings of service, productivity, retention, safety and profitability. Business units in the top quartile on engagement have been found to achieve around 23% higher profit than those in the bottom quartile.
So the finding is robust. The misapplications are where the damage happens.
Misapplication one: measuring individuals
The relationship is established at business-unit level. It is a property of teams, not a score to be attached to a person. Engagement data used in individual performance management stops being a measurement almost immediately, because respondents work out what the data is for and answer accordingly. Confidentiality is not an ethical nicety here; it is the precondition for the instrument working at all.
Misapplication two: treating the score as the intervention
Engagement is an outcome of how a team is managed. The meta-analytic work attributes the large majority of the variance in team engagement to management quality — on the order of 70%. That locates the lever precisely: it is the manager, not the survey, not the intranet, and not the benefits package.
An organisation that runs a survey, publishes a score and asks managers to improve it has skipped the part that works. The programmes that move are the ones where each manager gets their own team’s results, has a structured conversation with the team about them, and commits to a small number of changes that team can see.
Misapplication three: surveying more often than you act
Annual, biannual, pulse — the cadence matters less than the ratio of measurement to response. Asking people what is wrong and then not changing anything is worse than not asking, because it converts a diffuse frustration into a specific piece of evidence that nobody listened.
What to do with it
Measure at team level. Report to the team. Give the manager the result before their director sees it, so the conversation is about improvement rather than defence. Choose few enough items that the survey can be repeated. And treat a low-scoring unit as a management development question rather than a communications one, because that is what the evidence says it is.
