How to build a workplace culture people do not leave

MIT Sloan's analysis of millions of employee reviews during the Great Resignation found that a toxic corporate culture was around ten times more predictive of attrition than compensation. Not slightly more. Ten times.

The organisations losing people are rarely the ones without table tennis tables. They are the ones where disrespect goes unaddressed, where the same manager keeps producing departures, and where saying something honest in a meeting carries a cost. No benefit package compensates for a place where you cannot speak.

1. 1

   Psychological safety

   Can people disagree, admit a mistake or ask an obvious question without paying for it? Amy Edmondson's research makes this the strongest single predictor of team performance and of whether problems surface early.

2. 2

   Procedural fairness

   Do people understand how decisions about pay, promotion and workload get made, and do they believe the same rules apply to everyone? Perceived favouritism corrodes faster than bad news does.

3. 3

   Clarity

   Does each person know what good looks like in their role this quarter? Ambiguity is not neutral. It is the environment in which blame and politics grow.

4. 4

   Sustainable workload

   Is the expected output achievable in contracted hours by a competent person? If the answer is no, wellbeing provision is a plaster over a wound you are still making.

5. 5

   Leaders who model it

   Do senior people take leave, admit errors, and act on the behaviour of high performers who break the rules? Staff read behaviour, not statements, and they read it accurately.

Most engagement surveys fail for the same three reasons: they are too long to repeat often, too identifiable to answer honestly, and never closed off with visible action. Fix those three and measurement starts working.

• Short and repeated beats long and annual. A focused set of questions every quarter gives you a trend, and a trend is the only thing you can manage.
• Genuinely anonymous, with results never reported for groups small enough to identify individuals. If people suspect otherwise, you are measuring caution, not culture.
• Cover all five markers, not just happiness. Happiness tells you the weather. Safety, fairness, clarity, workload and leadership tell you the climate.
• Triangulate with hard signals: regretted attrition by manager, absence, grievance volume, internal mobility and exit interview themes. Where the survey and the data disagree, believe the data.
• Close the loop within thirty days. Publish what you heard, what you will change, and what you will not change and why. Silence after a survey is worse than never running one.

The HeartBrain Workplace Culture Questionnaire is built for exactly this: an anonymous, repeatable measure across the areas that predict whether people stay, with results presented so leaders can act on them rather than admire them.

Whether bullying is challenged is the single clearest read on a culture, because it puts values into direct conflict with convenience. Everyone watches what happens next, and they draw the right conclusion.

Four questions worth answering honestly at board level:

1. Can a member of staff name the route for raising bullying without looking it up?
2. Has anyone senior been held to account for their behaviour in the last two years?
3. Do complaints get resolved, or only closed?
4. Does anyone who raises a concern still get promoted afterwards?

If those answers are uncomfortable, the good news is that this is the fastest lever available to you. Acting on one visible case changes how a whole organisation reads its own values, and it does it in weeks rather than years.