Your Engagement Survey Is Measuring the Wrong Thing

Every People function runs the survey, watches the score, and treats the direction of that number as the verdict on the year. Up is good. Down triggers a workstream. It is the closest thing HR has to a P&L. And that is exactly the problem, because a single composite score can drift upward while the thing you actually care about, whether your best people stay and perform, quietly moves the other way. When engagement and regretted attrition rise in the same year, the survey isn't lying. It is answering a different question than the one you're asking it.

Engagement measures how people feel about being here. Retention of your strongest people depends on something the feelings score barely touches: whether they are growing, whether their manager is any good at the hard conversations, whether they are becoming more capable or just more comfortable. A team can be pleasant, well-fed, and pleasant to survey, and still be losing its sharpest members to boredom.

Engagement tells you how people feel. It does not tell you whether they are getting better. Your best people leave over the second question, and your survey never asked it.

The number is under real pressure right now

This matters more than usual because the headline metric itself is wobbling, which tempts leaders to pour budget into lifting a score rather than fixing what it misses.

Gallup's State of the Global Workplace 2025 reported that global employee engagement fell to 21%, only the second decline in twelve years, a drop Gallup linked to falling manager engagement in particular, and estimated the lost productivity at roughly US$438 billion. Pair that with Gallup's long-standing finding that managers account for around 70% of the variance in team engagement, and the picture sharpens: the score is falling, and it is falling from the manager layer, which is precisely where capability, not sentiment, is decided.

Read together, those two facts undercut the reflex to treat engagement as a thing you boost with perks and comms. If the score is set at the manager level, then the lever is manager capability, and the survey's job is to point you there, not to become the target you optimise for its own sake.

The two questions, on two axes

The way out is to stop reading engagement as a single line and start reading it against a second axis: capability. Put feeling on one axis and growth on the other, and four very different situations appear. High engagement and high capability is what everyone assumes a good score means. Low on both is the obvious fire, and at least it is visible. The dangerous quadrant is the quiet one: high engagement, low capability. People are content, the survey looks healthy, and your most ambitious performers are the first to notice they have stopped growing. That is your silent flight risk, and a composite engagement score is designed not to see it.

What to do with a score that looks fine

None of this means the survey is worthless. It means the survey is an input, not a verdict, and it is blind on the axis that predicts whether your best people stay. The fix is not another pulse survey. It is to measure capability and its growth alongside engagement, and to treat the manager layer as the variable that moves both. A pleasant, plateaued team is not a success just because it scored well. It is a flight risk wearing a good number.

Your engagement dashboard will keep going up and down, and leadership will keep reading the direction as the story. The more useful question is quieter: are your people getting better, and does the manager above them know how to make that happen. Answer that, and the score takes care of itself.

Sources

• Gallup, State of the Global Workplace 2025: global engagement fell to 21%, roughly US$438 billion in lost productivity, decline linked to manager engagement (gallup.com).

• Gallup, "Managers Account for 70% of Variance in Employee Engagement" (news.gallup.com).

Written by Mahesh Tharani. Connect on LinkedIn.