Counting versus listening

Your engagement survey isn’t listening. It’s counting.

The difference between the two, why organisations keep choosing the wrong one, and what a survey actually promises There’s a meeting that happens in almost every large organisation, usually about six weeks after survey results land, and I’ve sat in more versions of it than I can count.

The engagement score has dipped. Not collapsed, dipped. A 7.2 has become a 6.8. The heat map has gone amber in two divisions. Leadership is concerned, genuinely concerned, and a working group has been formed.

And after the deck has been presented and the drivers have been discussed, the working group converges, with remarkable consistency, on a conclusion: the problem might be the survey.

Maybe the questions are outdated. Maybe the scale is confusing people. Maybe a different provider would give richer insight. Within a quarter, there’s a procurement process. Within a year, there’s a new instrument, a new baseline, and, conveniently, no way to compare anything with the years before.

The score dipped, so the organisation changed the ruler.

I want to be fair: nobody in that room believes they’re dodging. Every step feels reasonable from the inside. But step back and the pattern is unmistakable, and it points at a misconception buried so deep in people strategy that we’ve stopped noticing it’s there.

The misconception: measurement is listening.

It isn’t. Counting and listening are two different acts. They answer different questions, they require different postures, and they produce different obligations. Most organisations do a great deal of the first, very little of the second, and sincerely believe they’re doing both.

What counting can and can’t do Let me be precise about what a score is, because I’m not here to sneer at measurement. I’ve spent years building listening architectures that depend on good quantitative instruments, and I’ll defend them in a moment.

A score is a compression. Thousands of individual experiences, each with its own texture and history, compressed into a number so that patterns become visible. That compression is genuinely useful. It tells you that something moved, roughly where, and roughly how much. It lets you spot the division that’s drifting, the demographic that’s diverging, the trend that’s building. Counting is a superb detection system.

But everything a score detects, it detects by throwing away the detail. That’s not a flaw, it’s the mechanism. A 6.8 is legible precisely because it has discarded every story that went into it.

Which means a score can never tell you the only things you can actually act on.

It can tell you “flexibility” dropped 0.6 in one division. It cannot tell you that flexibility there was never formally withdrawn, just quietly made expensive: the raised eyebrow at the 3pm school run, the important meeting that keeps landing at 3:30, the promotion that went, again, to someone with no pickup to do. Nobody changed the policy. Everybody changed their behaviour. The score registered the seismic activity. The cause lives entirely in details the score was built to discard.

It can tell you a frontline team’s engagement fell four points. It cannot tell you about last Tuesday: one person short again, the workarounds, the client who bore the brunt, and what hour ten of that shift felt like. One honest verbatim from that Tuesday will reorganise a leader’s understanding faster than the whole dashboard, because it carries the thing a number cannot: the reason.

Counting tells you something moved. Listening tells you what moved, and why. Detection and diagnosis. You need both, and only one of them can tell you what to do on Monday.

Why counting keeps winning If listening is where the actionable truth lives, why does organisational energy keep flowing to counting? Three honest reasons.

Counting is comfortable. A dashboard is a controlled environment. The data arrives cleaned, aggregated and anonymised, which is another way of saying it arrives with the humanity removed. Listening is not comfortable. A verbatim can be angry, unfair, or painfully accurate.

It might mention a named leader. Sitting with raw accounts of your organisation’s lived experience is emotionally expensive work, and dashboards let you avoid the expense while feeling diligent.

Counting is benchmarkable. Scores can be compared: against last year, against the industry, against a global norm. Benchmarks give leaders a place to stand (“we’re two points above sector”) and a language boards understand. Stories don’t benchmark. There is no percentile for last Tuesday. So the instrument that supports comparison gets the executive airtime, and the material that supports understanding gets an appendix.

Counting offers deniability. This is the uncomfortable one. A score creates distance between leadership and any specific human account. “Engagement declined 0.4” obliges no one.

“Rosters in the northern team have been short-staffed since March and people are breaking”

obliges everyone who hears it. Aggregation launders obligation out of information. I don’t think most organisations do this consciously. I think the system does it for them, and everyone quietly accepts the comfort.

Notice that all three forces are real and none of them has anything to do with understanding employees better. That’s how you end up with organisations that measure more than ever and know less than ever.

A survey is a promise Here’s the part that connects this essay to everything else I write, and it’s the part I most want to leave with you.

Every survey is a promise.

The moment you ask people how they feel, you have implied, whether you meant to or not, that the answer will change something. That implication is heard loudly. People don’t fill in surveys for fun; they fill them in as a bid, a small act of faith that honesty travels somewhere and returns as change.

Which means the survey itself sits inside the Promise Gap.

Ask and act, and something compounds: people learn that speaking honestly here is worth the effort, and every cycle the signal gets richer, because trust is what buys you the truth.

Ask and file, and something corrodes: people learn the annual ritual is theatre, and they respond as people always respond to theatre, by performing. Scores drift toward the safe middle. Verbatims go bland or go silent. The instrument keeps running, and it measures less every year, not because the tool got worse but because the trust that powered it got spent.

This is why “survey fatigue” is mostly misdiagnosed. People aren’t tired of being asked.

They’re tired of being asked by organisations that don’t answer. Nobody has ever been fatigued by being listened to.

An unanswered survey doesn’t just waste a data cycle. It widens the exact gap it was meant to measure.

The counterargument: in defence of the score

Now let me argue against myself, because there’s a version of this essay that overcorrects into “burn the surveys,” and that version is wrong.

Scores let organisations see at scale, and scale is not optional. A leader of twelve can listen to everyone; a leader of twelve thousand cannot, and pretending otherwise produces its own failure: anecdote-driven leadership, where the most recent or loudest story sets the agenda and quiet suffering stays invisible. Verbatims have their own biases. The people most willing to write three paragraphs are not a representative sample of your workforce. And longitudinal comparison, which only numbers can give you, is how you distinguish a bad quarter from a structural decline.

So keep the survey. Genuinely. The argument of this essay is not counting versus listening.

It’s about sequence and weight.

Count to detect. Listen to diagnose. Act on what you hear. Then count again to see if the action worked. That’s a loop, and the score plays a vital role in it. What the score can’t do is substitute for the middle steps, and the middle steps are exactly what the pattern I opened with skips. Changing the survey when the score dips is running the detection step twice and calling it a response.

What listening actually looks like If your organisation counts well and wants to start listening, the shifts are practical and none of them require new software.

Verbatims reach decision makers raw. Not themed, sentiment-scored and summarised into a word cloud. A standing agenda item where leaders read a rotating selection of unedited comments from their own area. Twenty minutes a month. It changes rooms. A word cloud can show what is common. It cannot show what the experience cost the person who wrote it.

Someone follows the score down. When a number moves, the response is a person going to where it moved and asking, face to face: “the survey says something changed here, tell me about it.” The score is the map; the conversation is the territory. Organisations that institutionalise this one habit, score-triggered listening visits, close more gaps in a quarter than survey redesigns close in a decade.

The loop closes visibly and honestly. “You said, we did” is the standard version, and done glibly it makes things worse. The honest version has three parts: here’s what we heard, here’s what we’re changing, and here’s what we heard but can’t change, and why. That third part is the one organisations skip and the one that builds the most trust, because it’s the part that treats people as adults.

Endings get listened to hardest. The richest, least defended truth in your organisation walks out the door with every departure, and most exit processes are designed to file it rather than hear it. The last conversation deserves to be your best-designed listening moment, not your most perfunctory. That thought is getting its own essay soon.

What to do with this on Monday

Before you change anything about your survey, pull the last cycle’s verbatims for one team whose score moved, and read fifty of them yourself, unfiltered. Diagnose before you re-instrument.

Ask your leadership team one question: “What did we change because of the last survey?” If the answer takes longer than thirty seconds, the organisation is counting, not listening, and everyone below already knows it.

Pick the single lowest-scoring item in one division and send a person, not a pulse, to find the story behind it. Report the story, not the number, at the next executive meeting. Watch what happens to the room.

Add the honest third line to your next results communication: what we heard and won’t be changing, and why. It will be the most-read sentence in the document.

The sentence to keep

The score tells you something moved. Only a story can tell you what.

Organisations don’t have an information problem. The truth about the experience they’re providing is available, every day, from the people living it, at the price of being genuinely heard. Counting was never the failure. Stopping there was.

Next week: back to the promise itself, and a test anyone can run in ten minutes. Read your careers page. Then read your reviews. The distance between them is your real employer brand.

If this one hit a nerve, subscribe and tell me: what’s the verbatim you’ve never forgotten?

The ones that stay with us are usually the ones that were true.

Shehzad Bhanji writes The Promise Gap, a weekly perspective on the relationship between organisational promises and lived experiences. Across a 25-year international career spanning marketing, customer experience, employer brand, HR technology and people experience, he has worked across Australia, Asia, Europe, the Middle East and Africa.