The Cost of Not Listening

What the room says, and what it costs. An article from the Listen series.

I have been in the room.

Board meetings. Faculty meetings. Dorm hallways. Zoom squares. I have been the person raising something, and I have been the person on the other side of the table when someone else raised it. Over twenty years working in schools, I have heard the same small set of sentences used to end a conversation that should have started one. If you have spent any time inside an organization, you have sat in those rooms too, on one side of the table or the other.

None of them sound like dismissal. That is what makes them effective.

What the Room Says

"You're just one person, and that's your opinion."

"That's not something that really hits the mark of what we can address right now."

"As we reflect on this, the majority of the organization has not expressed this."

"Thank you for sharing. What I hear you saying is..."

"Let's take that conversation offline."

"There is not enough data to support that."

"You need to move on from that and consider the bigger picture."

Read them again. Not one is hostile. Several are things I have said myself.

I did not invent them. I was taught them. Heads of school and managers coached me toward this language over the years. When I handled a moment badly, someone more senior showed me a better way to say it, and the better way was almost always one of these. They de-escalate. They protect the meeting. They make you sound composed. I got better at them, and getting better at them was treated as growth.

The irony took me years to see. Every one of those coaching conversations was itself an act of not listening. Nobody ever asked what I had actually heard, or why I would not let it go, or whether the thing I was raising might be true. The correction was always about my delivery. Never once about the content.

The pressure ran both directions. Using this language was rewarded. Not using it had a cost. Staying in a hard conversation past the point where the room wanted it closed got me described as someone who could not read the moment, who was too invested, who needed to work on executive presence. There is a real professional price for being the person who will not let a thing go, and most of us learn to pay attention to that price before we learn anything else.

In a company that price has a procedure attached to it. An employee raises something formally, the documentation routes to their manager, and the person who spoke up is now identified to the one relationship that sets their reviews and their assignments. The process is designed to be fair and it is still an act of exposure. Most people run that arithmetic once.

That is how this actually transmits.

Nobody trains you to stop listening. They train you in a vocabulary that ends conversations and call it leadership.

You learn to sound like a leader before you learn to lead, and the sounding-like gets rewarded first.

They do four different jobs.

They discredit the source. One person. An opinion. The signal is reclassified as a personality.

They defer. Offline. Later. Let me consider that. These are the most damaging because they are the most courteous. The person leaves feeling heard. Nothing was recorded, assigned, or scheduled. Six months on there is no trace the conversation happened, so when the same thing surfaces again it registers as new rather than as a pattern. That is how an institution loses its own memory.

They invoke the majority. Most people have not said this. There is not enough data. Both sound rigorous. Both are wrong in a specific way I will come back to.

They reframe the speaker as the problem. Move on. Consider the bigger picture. Now the person who noticed something is the one being asked to grow.

I want to be fair about something. Some of these are sometimes true. There are conversations a room genuinely cannot hold that day. There are problems an institution cannot afford to solve this year. There are moments when taking it offline is the right call, because the topic deserves more than the eleven minutes left on the agenda. Leaders sequence. That is the job.

The difference is what happens next. A legitimate deferral has a date on it. Somebody writes it down, somebody owns it, and the person who raised it hears back. What I am describing is the version with no next. The sentence that sounds like sequencing and functions as an ending.

The Part That Is Actually a Math Error

The majority has not expressed this. There is not enough data to support that.

I want to take those two seriously, because they are the ones offered in good faith by people who think of themselves as evidence-driven. They rest on a mistake about what kind of information a complaint is.

Decades of customer research, most of it traceable to the work of TARP, the Technical Assistance Research Programs Institute, found that a typical organization hears from about 4% of its dissatisfied customers. For every one person who says something, roughly twenty-six say nothing. Of that silent 96%, about 91% never come back. And they are not quiet everywhere. Each dissatisfied person tells nine to fifteen others, and roughly 13% tell more than twenty.

The same body of research found the reverse also holds. About 70% of people who raised a problem and were satisfied with the response stayed. More than 95% stayed when the response came quickly. Which means the person who speaks up is not your problem. That person is the one you can still keep.

So when one family raises something, that is not one family. Under those numbers it is closer to twenty-six, most of whom you will never hear from, some of whom are already deciding, and all of whom are talking to people you are trying to enroll.

The error is treating a complaint as an estimate when it is an indicator.

An estimate answers how much. An indicator answers whether. Complaint data was never built to measure a proportion, because the people who complain were never a random sample of anyone. Its job is to tell you that something exists.

Nobody asks whether smoke is statistically significant.

You do not need a representative sample of smoke. One column of it tells you something is burning, and the only useful question is where. A complaint works the same way. The response that matters is not counting how many people reported it. It is going to find what is producing it.

Quality engineering has a precise version of this distinction. Common cause variation is the ordinary scatter a stable system produces on its own: some weeks run better than others and nothing has actually changed. Special cause variation is different in kind. It means something in the system itself has shifted, and it almost always shows up first as a small number of readings that do not belong. Telling those two apart is the entire discipline of process control, and the reason it exists is that an organization which files every special cause as noise fails slowly and then all at once.

There is a second problem with the data most organizations collect, and it is worse. Schools survey the families who are still enrolled. Companies survey the employees who are still on payroll. We measure the students who stayed and the people who did not quit. The ones who could tell us the most are already gone, and they are structurally absent from every dataset we use to reassure ourselves. Survivorship bias is not a flaw in our instruments. It is the design of them.

Two more sit on top of it.

Proximity bias is what happens because leaders hear most from whoever is nearest. The cabinet, the department heads, the families who email. Distance from the person in charge is not randomly distributed, and neither is who feels safe closing it.

Recency bias is what happens because the last thing said outweighs the pattern. A complaint in March is a crisis. The same complaint in each of the four previous Octobers is history nobody assembled.

Survivorship decides who is in the data. Proximity decides who reaches you. Recency decides what you act on. An organization can be running all three at once and still believe it is listening.

What I Learned Doing the Interviews

I did not learn any of this from the research. I found the research later, because I was looking for language to explain something I had already watched happen.

Over the years I have sat in roughly a hundred and twenty conversations with graduating students, another hundred and twenty with families who were leaving, a hundred with mentees, the ninety day check-ins with new hires, and the alumni calls on top of those. Some of it was my job. Some of it became BLAC. All of it was the same act: asking someone what the experience had actually been.

That is the data TARP says organizations do not have. I had it. I was the person collecting it.

At one school, year over year attrition ran around 10%. That number is not scandalous. It is normal, which is exactly the problem. A tenth of the community turns over, the budget absorbs it, and the figure gets reported as a rate rather than as a hundred and twenty separate conversations that somebody could have had earlier.

What recurred was almost never about a person. Very few people left because of a bad teacher or a bad manager, and when they did, they usually said so early and it usually got handled.

What surfaced again and again was a promise.

Every organization has a value proposition. It is on the website. It is in the viewbook. Sometimes it is painted on a wall. It says what this place will be for the person who chooses it, and people choose it because they believe the sentence. What kept coming up in those conversations was the distance between that promise and what the person actually experienced, and underneath that, the sense that nobody was checking.

Institutions set the proposition once and then rest on it. On the mural. On the stated morals. On the fact that it was true when it was written and nobody has said otherwise loudly enough. What almost none of them do is validate it, which would mean asking the people currently living inside that promise whether it still describes their experience, and then changing either the promise or the practice when the answer is no.

Status quo is not neutral. It is a decision that gets made over and over by not being examined. And the people who leave are usually the ones who noticed the gap first.

Here is the finding that has stayed with me longest, and it is the one I did not expect.

The ninety day conversations and the exit conversations were about the same things.

Not the same person. That is the part that matters. A new hire at ninety days would name something small: a gap between what they were told during recruitment and what they found, a process that did not work the way it was explained, a stated value they had not yet seen practiced. They raised it gently, as an observation, usually assuming someone would want to know.

Then, in a different conversation, sometimes in the same season, someone on their way out after five or eight years would describe the same gap. Except now it was not an observation. It was the reason they were leaving, with years of accumulated evidence attached.

One person handed me the seed. Another handed me the tree.

Same gap, both ends of the lifecycle, two different people who had never spoken to each other.

And we held both documents. The ninety day notes and the exit notes lived in the same organization, often in the same office. Nobody ever put them side by side. If anyone had, the pattern would have been obvious, and the exit interview would have read like a prediction that had already been filed and ignored.

That is a cost curve, drawn from my own notes rather than a textbook. The cheap version of that information was on the table at day ninety, when someone volunteered it without being asked twice. The expensive version got paid at the exit interview, in a search, an interim, and a year of getting someone new up to speed.

We had the signal. We had it early. We had it in writing. What we did not have was any mechanism that treated a ninety day observation as something that would still matter in year three, or that connected what one person noticed on the way in to what another person concluded on the way out.

And here is what I need to say plainly, because it is the part that took me longest to accept. The information was not missing. I gathered it. I wrote it up. I presented it. I sat in rooms and delivered what departing families and graduating students and outgoing employees had told me.

Then it was received. And that was the end of it.

Nobody suppressed anything. There was no bad actor. The reports were read, thanked, filed, and referenced occasionally in the way institutions reference documents that have already been absorbed into the past. The listening had been performed. The deliverable had been delivered. The thing the deliverable was supposed to change went on unchanged.

That is when I stopped believing the problem was collection. Organizations are not short on information about how people experience them. They are short on the willingness to let that information cost them something.

What It Costs

Quality management has a rule of thumb, the 1-10-100 rule, developed to describe manufacturing defects. A problem costs about one dollar to prevent at the source, about ten to correct once it is inside your process, and about one hundred once it reaches the customer. Every stage a problem survives multiplies what it will eventually take to resolve.

That is the curve I had already drawn from my own notes, without knowing it had a name.

Translate that into a school, or any organization where the relationship is the business and the people are the whole of it.

Prevention is a question. Someone asks a student how the year is actually going and hears the answer without flinching. A manager asks a new hire what has been harder than expected and does not rush to fix it. Someone follows up with a family after a difficult week. Cost: a conversation.

Correction is a meeting. Now there are three adults, a counselor, a division head, a scheduling problem, and a parent who has already told several other parents. In a company it is HR, a manager, a skip level, and a person who has already started preparing to leave. Cost: hours, plus trust that has to be rebuilt rather than maintained.

Failure is someone who leaves. A family, an employee, a customer, a client. And it is never one departure. A family leaving is the remaining years of tuition, the sibling who now goes elsewhere, the referral that never happens, the nine to fifteen people who hear the story, and the alumnus twenty years later who does not give, does not return, and never tells anyone why. An employee leaving is the search, the interim coverage, the institutional knowledge that walks out with them, the year it takes their replacement to become useful, and the colleagues who quietly start looking because they watched what happened.

That last one is the part institutions never count, because it does not arrive as a line item. It arrives as an absence.

What I Finally Understood

What I finally understood is that not listening is not free. It is financed. The few who speak up are the tip of an iceberg, and ignoring them does not make the iceberg go away. It moves the cost downstream, into students who leave, employees you replace, and a ledger that never names what it is actually paying for.

There is one more thing about the timing, and it is the reason I keep pressing on this.

The cost of resolving something scales with how many people it has touched. When a signal comes from a few, the response is small: a policy adjusted, a practice changed, a person heard. When the same signal finally comes from enough people that leadership feels safe acting, the response has to be large: a restructuring, an outside consultant, a communications strategy, a search to replace the people who left. The bill arrives eventually. Waiting only determines the size of it.

Institutions do not usually decide not to listen. They decide to wait until the number is big enough to be comfortable. And by then they are not preventing anything. They are recovering.

The Ledger

If you run something, you already have the data on the expensive end of this.

You know your attrition rate. You know what re-enrollment looked like this year and last. You know how many searches you ran, how many good people left, how much you spent on the consultant who came in after the thing that could not be ignored anymore. Those numbers are in your budget right now, under labels like turnover, attrition, and professional services.

What is not in your budget is a line for the conversations that would have prevented them. There is no account for the question nobody asked, no cost code for the thank you for sharing that went nowhere, no way to attribute this year's departure to the meeting three years ago where somebody said something and the room decided it was one person's opinion.

That is the real problem with the cost of not listening. It is not that institutions refuse to pay it.

It is that they pay it without ever knowing what they bought.

The first pillar of the way I work is Listen, and this is why it comes first rather than somewhere polite in the middle. Listening is not a soft skill sitting next to the real work. It is the earliest and cheapest point of intervention any organization has, and every stage past it costs more. A question is a dollar. A meeting is ten. A departure is a hundred, and it does not arrive with a label explaining what it was.

Everything I build sits on that order. Listen, then learn, then lead. I was taught the reverse, and I taught it to other people before I understood what it cost.

If this sounds familiar

You already have the numbers on the expensive end of this. Attrition, re-enrollment, the searches you ran, the consultant who came in after the thing that could not be ignored. What no budget carries is a line for the conversations that would have prevented them.

If you want to talk about what your own ninety day notes and exit notes would say placed side by side, my calendar is open.

A note on sources

TARP refers to the Technical Assistance Research Programs Institute, a Washington, D.C. consultancy, not the 2008 financial program of the same acronym. Its landmark study, Consumer Complaint Handling in America, was based on research conducted between 1974 and 1979, initiated by the White House Office of Consumer Affairs, published in 1979 and updated in 1986. John A. Goodman and Steve Newman of TARP summarized much of this work in "Understand Customer Behavior and Complaints," Quality Progress, January 2003. TARP later became CX Solutions.

The loyalty findings cited here, that roughly 70% of customers who complained and were satisfied remained loyal and more than 95% of those satisfied quickly remained loyal, come from TARP's original research. The estimate that twenty-six customers stay silent for every one who complains is widely circulated in customer service literature and is commonly attributed to Lee Resource Inc. rather than to TARP. The figures on word of mouth, that a dissatisfied customer tells nine to fifteen others and about 13% tell more than twenty, are attributed to the White House Office of Consumer Affairs. I have separated these deliberately, because they are frequently blended together and credited to a single source.

The 1-10-100 rule was codified by Labovitz, Chang, and Rosansky in Making Quality Work (1992), with a related rule of ten for software defects demonstrated by Boehm in 1976. Common cause and special cause variation come from statistical process control as developed by Shewhart and Deming.

All of these figures come from commercial contexts. I use them here as a way of thinking about magnitude and timing, not as precise predictors of school outcomes.

The sentences quoted at the top of this article are ones I have heard in schools over two decades, in board meetings, faculty meetings, dorms, and on video calls. They are reproduced here as language, not as attributions to any particular institution or person.

Jonathan E. Lee is the founder of Bridging Legacies Across Campuses (BLAC), where he collects the stories of how people find their way through educational and professional life and studies why those journeys succeed. This article explores the first pillar of The BLAC Framework™: Listen.