Employee Engagement

The three words are used interchangeably in most organisations and describe different states.

The distinction has a practical edge. An organisation can raise satisfaction by improving conditions and see no change in performance, because comfortable and indifferent is a stable state. It can also have highly engaged people who are dissatisfied about something specific and will say so loudly, which is a healthier position than it feels.

Engagement is the one most closely tied to what the organisation gets. It describes whether someone gives more than the role strictly requires, which is largely discretionary and therefore responsive to how they are treated.

The evidence across most organisations points at a short list, and it is not the list most engagement programmes address.

• The immediate manager, by a wide margin. Whether they are fair, competent, available and honest determines more than any organisational initiative.
• The work itself: whether it is meaningful to the person doing it, whether they have some control over how it is done, and whether they can see its result.
• Clarity, meaning knowing what is expected and how it will be judged. Ambiguity is corrosive and cheap to fix.
• Being heard, which means the organisation visibly acts on what people raise, at least sometimes.
• Fairness in the things that are compared: pay, progression, workload and how misconduct is handled.

What moves it less than expected: perks, events, communications campaigns and rebranding exercises. These are visible, easy to organise, and largely ineffective against any of the five above. An organisation with a manager problem cannot buy its way out with benefits, and employees read the attempt accurately.

The corollary is uncomfortable. Improving engagement usually means improving management, which is slower and harder than running a programme, and it is the only route that works.

Surveys are the standard instrument and they are useful within limits that are frequently exceeded.

They measure well: change over time in the same population, differences between teams, and which specific factors are rated lowest. They measure badly: absolute levels, comparisons against other organisations, and anything where the sample per team is very small.

The organisation-wide score is the least useful number produced and the one most often reported to a board. Engagement varies far more between teams within an organisation than between organisations, so an average conceals the distribution that contains all the actionable information.

Two design points matter more than the question set. Anonymity has to be real and has to be believed, and in small teams those are different problems, since a team of four cannot be surveyed anonymously whatever the assurance. And free-text responses usually carry more signal than the scores, because they say what the problem is rather than that there is one.

Frequency is a trade-off rather than a virtue. Frequent short pulses track change and produce fatigue if nothing visibly follows; annual surveys allow deeper questions and are too slow to catch a deteriorating team.

Running a survey and doing nothing is worse than not running one.

Employees who report a problem and see no response learn something specific and durable: that raising an issue is not a route to change. The next survey attracts fewer responses and blander ones, which is often misread as improvement.

The pattern is recognisable. Scores drop, the results are discussed at length by a leadership team, an action plan is produced, one or two visible items are addressed, the harder items are not, and nobody tells employees which is which. A year later participation is down and the scores are flat.

What breaks it is closing the loop honestly, including on the things the organisation will not do. Telling people that a request has been considered and declined, with a reason, preserves the credibility of the exercise. Silence does not, and neither does an action plan that quietly omits the largest issue raised.

The practical rule is to survey only as often as the organisation is prepared to respond, which for most is less often than they currently do.

Engagement is most useful when it is not read on its own.

It tends to move before other indicators. A team whose engagement falls will usually show rising unplanned absence some months later, and resignations after that. Reading the three together, by team and by manager, gives a considerably earlier warning than any of them separately.

The pairing worth building is engagement, absence and attrition against manager and tenure. Where all three deteriorate in the same team, the cause is almost never mysterious, and the intervention required is usually a conversation with one person rather than a programme.

It is also worth checking engagement against the things people are comparing themselves on. Where a team's engagement falls after a pay review, or after a promotion round, the cause is likely to be visible in the pay and progression data rather than in the survey.

None of this requires sophisticated analysis. It requires the survey results, the absence data and the exit data to be readable against the same team and manager, which is a systems question more than an analytical one.