Low complaint volumes do not mean satisfied customers.

They may mean customers who have decided that complaining is not worth their time.

Those two conditions produce identical complaint data — and very different retention outcomes.

Your complaint handling process is probably working. Complaints are logged. Responses are sent. Resolution times are tracked. Escalations are managed. The dashboard looks clean.

And still — the same problems keep generating the same complaints, month after month, in the same parts of the customer journey.

The issue is not your complaint handling process. The issue is what you are using it for.

The Complaint Iceberg

The fraction of dissatisfaction that chose to surface

I call this the Complaint Iceberg: the structural condition in which an organisation mistakes the complaints it receives for a representative picture of the dissatisfaction its customers experience.

It is not.

No more than 25% of customers raise a serious problem with frontline staff. For an irritating one, fewer than 5% do. (John A. Goodman, TARP / CCMC)

Dissatisfied customers experience the problem. They decide the effort of complaining is not worth it. And they leave — quietly, without notice, without giving the organisation the information it would need to prevent the next customer from doing the same thing.

The complaints that reach your manager are the visible tip. The dissatisfaction that drives customer exit is the mass beneath the waterline — invisible, unmeasured and, in most organisations, structurally inaccessible.

The organisation that manages its complaint volume is not managing its customer dissatisfaction. It is managing the fraction of dissatisfaction that chose to surface.

Sources: John A. Goodman, co-founder of TARP (1971) and vice chairman of Customer Care Measurement & Consulting. His complaint-rate figures draw on more than 1,000 customer service studies, including TARP's White House–sponsored evaluation of complaint handling in government and business. On non-complainers specifically, see Voorhees, Brady & Horowitz, "A Voice From the Silent Masses: An Exploratory and Comparative Analysis of Noncomplainers", Journal of the Academy of Marketing Science 34(4), 2006; and Chebat, Davidow & Codjovi, "Silent Voices: Why Some Dissatisfied Consumers Fail to Complain", Journal of Service Research 7(4), 2005. A separate Netigate analysis (2025) found that 85% of customers who left a provider say they would have stayed if their problem had been addressed in time. The widely circulated claim that 91% of unhappy customers leave without complaining traces to marketing material rather than to published research, and is not used here.

What the research shows

Complaint volume is a measure of complaint-worthiness, not dissatisfaction

The threshold for formal complaint is high. Customers complain when they believe the effort is worth it — which means they complain about significant failures in high-stakes situations. For everyday friction, poor interactions and incremental disappointments, they stay silent.

The implication is direct: complaint volume is not a reliable measure of customer dissatisfaction. It is a measure of the dissatisfaction that crossed the threshold of complaint-worthiness — which is a small and systematically unrepresentative subset of the whole.

An organisation with low complaint volumes does not have satisfied customers. It may have customers who have decided that complaining is not worth their time.

Those two conditions produce identical complaint data. They produce very different retention outcomes.

Why organisations misread the signal

The Signal Inversion Problem

Most organisations treat complaint volume as an inverse measure of customer satisfaction. Fewer complaints means better service. Rising complaints means deteriorating service.

This is wrong in both directions.

A sudden increase in complaint volume often does not mean service has got worse. It means something has changed — a channel has made it easier to complain, a communication has prompted customers to surface existing frustrations, a process change has lowered the effort of complaining. The underlying dissatisfaction may have existed for months before it became visible in the complaint data.

Conversely, stable or declining complaint volumes tell you almost nothing about the direction of customer satisfaction. They tell you that the threshold for formal complaint has not been crossed more often. What is happening beneath that threshold — in the everyday friction, the unacknowledged errors, the interactions that did not meet expectation but not badly enough to justify the effort of complaining — is invisible to any system that measures only what surfaces.

I call this the Signal Inversion Problem: the tendency to read complaint volume as a measure of service quality, when it is actually a measure of complaint-worthiness — a function of customer effort, expectation and the perceived value of voicing dissatisfaction.

The organisations that understand this distinction manage very differently from those that do not. They do not celebrate low complaint volumes. They ask what those low volumes are concealing.

The governance failure underneath the iceberg

Two failures before the complaint arrives

When a complaint reaches a manager, two things have already failed.

The first is the service delivery system — whatever process, interaction or decision generated the dissatisfaction that produced the complaint.

The second is the early warning system — the organisation's ability to detect and respond to dissatisfaction before it becomes a formal complaint, a churn event or a negative review.

Most organisations invest heavily in the complaint resolution process — the response time, the resolution quality, the recovery experience. This is not wrong. How an organisation responds to a complaint matters significantly for whether the complaining customer stays or leaves.

But it addresses the wrong problem. It addresses what happens after the complaint arrives. It does not address why the complaint arrived — and it does nothing about the much larger population of customers who experienced the same dissatisfaction and did not complain.

The organisation has built an excellent system for managing the visible portion of its customer dissatisfaction. It has no system — or an inadequate one — for understanding, measuring and acting on the portion it cannot see.

The result is a permanent cycle: complaints arrive, are resolved, and are generated again by the same underlying causes — because those causes are never addressed structurally. The complaint handling process becomes a cost of doing business rather than a source of diagnostic intelligence.

Four signals beneath the waterline

What to measure when complaint volume is not enough

If formal complaint volume is an unreliable measure of customer dissatisfaction, what replaces it? Four signals that exist beneath the formal complaint threshold — and that most organisations are not systematically measuring.

Signal 1: Silent churn

Customers who leave without explanation, without a formal complaint, without a single interaction that registered as a dissatisfaction signal in the internal system. Consider the customer who experienced a failed onboarding, never contacted the service team, renewed once more because switching felt effortful — and then quietly left at the next renewal point. They do not appear in complaint analytics. They appear only in retention data as a number, stripped of the diagnostic information that would have allowed the organisation to prevent their departure.

The diagnostic gap: Exit surveys, conducted at the point of churn, are the only mechanism that recovers this information after the fact. Most organisations do not run them consistently. Those that do rarely connect the findings to the governance processes that would need to change.

Signal 2: Contact without complaint

The customer who calls to ask a question that should have been answered in the onboarding material. The customer who requests a callback because the digital journey broke down at a specific step. The customer who contacts the service team three times to resolve a problem that should have been resolved once. None of these interactions register as complaints. All of them represent dissatisfaction — with the information architecture, the channel design, the process reliability.

The diagnostic gap: Most organisations track contact volume. Few track contact reason at sufficient granularity to identify the underlying journey failures generating unnecessary contacts. Contact reason analysis — a systematic review of why customers contact the service team, disaggregated by journey stage — reveals dissatisfaction signals that never appear in complaint data.

Signal 3: Friction without feedback

Digital journey analytics — abandonment rates, retry rates, error frequencies, time-on-task variance — reveal where customers are experiencing friction that never becomes a formal complaint or a service contact. A customer who abandons a renewal process three times before completing it on the fourth attempt has experienced significant friction. They have not complained. They appear in the analytics as a completed transaction.

The diagnostic gap: The friction is real. The dissatisfaction is real. The signal exists — but only in data that most organisations are not connecting to their customer experience management processes.

Signal 4: Social and informal expression

Reviews, social posts, informal conversations — the expression of customer dissatisfaction that reaches the market rather than the organisation. By the time dissatisfaction appears in a public review, the organisation has already lost the opportunity to recover the relationship. But the diagnostic content — what failed, where, and why — is often more specific and more honest than formal complaint data, precisely because the customer is not constrained by the organisation's complaint process.

The diagnostic gap: Most organisations monitor this signal reactively, as a reputation management function. Few use it systematically as a source of diagnostic intelligence about the experience failures occurring beneath the formal complaint threshold.

What a below-waterline measurement system looks like

Three components — and the governance decision that makes them matter

Managing the Complaint Iceberg requires a deliberate decision to measure dissatisfaction rather than complaints. These are different things — and the systems required to measure them are different.

A proactive listening architecture

A Voice of Customer system designed to capture customer sentiment at multiple points in the journey — not just at moments of formal complaint. Transactional surveys at key journey stages, relationship surveys at defined intervals, contact reason analysis embedded in service operations. The goal is not to generate more data. It is to generate earlier signals — to move the detection point upstream, before dissatisfaction crosses the complaint threshold or the churn threshold.

A silent churn diagnostic

A structured process for recovering diagnostic information from customers who leave without complaint. Exit interviews or surveys conducted within a defined window of churn, designed to surface the dissatisfaction that was present but never expressed. The findings should feed directly into the governance process — not as a retention report, but as a root cause analysis that identifies the journey failures, process gaps and decision points that drove the departure.

A contact reason taxonomy

A classification system for service contacts that identifies the underlying journey failure generating each contact type — not just the channel or the query category, but the specific process failure, information gap or system error that caused the customer to need to make contact. Applied consistently, contact reason analysis becomes one of the most powerful diagnostic tools available to a CX leader — because it captures dissatisfaction at the moment it surfaces, before the customer decides whether to complain, stay or leave.

None of these three components are primarily technical. They are governance decisions — about what the organisation will measure, who is accountable for acting on what it finds, and what will change when the findings point to a structural cause.

Without those governance decisions, the measurement systems produce data that informs without obliging. The Complaint Iceberg remains. The same problems continue generating the same complaints. The resolution process continues to address symptoms while the causes remain unmanaged.

The question before your next complaint review

Your complaint handling process is managing the tip of the iceberg.

Before your next complaint review, ask one question:

The diagnostic question

What proportion of our customer dissatisfaction last quarter was never expressed as a formal complaint — and do we have any mechanism to understand what it was, where it occurred, and what caused it?

If the answer is no — you do not have a complaint handling problem. You have a measurement problem. And the complaints your manager receives are not a picture of your customer dissatisfaction. They are the fraction of it that chose to surface.

The rest is beneath the waterline. And in most organisations, it stays there — until it appears in a churn report that nobody can fully explain.

If this is relevant to your organisation — share it with the person who presents your complaint data to the leadership team.

Frequently Asked Questions

The Complaint Iceberg is the structural condition in which an organisation mistakes the complaints it receives for a representative picture of the dissatisfaction its customers experience. Research by John A. Goodman, co-founder of TARP, finds that no more than 25% of customers raise even a serious problem with frontline staff, and fewer than 5% raise an irritating one — roughly 20 silent customers for every complaint. The visible tip is a small and systematically unrepresentative subset of the whole, and its size varies with how serious the problem is.
The Signal Inversion Problem is the tendency to read complaint volume as a measure of service quality, when it is actually a measure of complaint-worthiness — a function of customer effort, expectation and the perceived value of voicing dissatisfaction. An organisation with low complaint volumes may have satisfied customers — or it may have customers who have decided that complaining is not worth their time. Those two conditions produce identical complaint data and very different retention outcomes.
Because it addresses what happens after the complaint arrives — not why the complaint arrived, and not the much larger population of customers who experienced the same dissatisfaction and did not complain. The complaint handling process becomes a cost of doing business rather than a source of diagnostic intelligence when it is not connected to a system for measuring and acting on the dissatisfaction that never surfaces as a formal complaint.
Silent churn — customers who leave without complaint or explanation; contact without complaint — service contacts that represent dissatisfaction but are not logged as complaints; friction without feedback — digital journey failures that appear in analytics but not in complaint data; and social and informal expression — dissatisfaction that reaches the market rather than the organisation.
Three components: a proactive listening architecture that captures customer sentiment at multiple journey points before dissatisfaction reaches the complaint threshold; a silent churn diagnostic that recovers information from customers who leave without complaint; and a contact reason taxonomy that identifies the underlying journey failures generating unnecessary service contacts. All three require governance decisions — about accountability, measurement and what changes when the findings point to a structural cause.
A sudden increase in complaint volume does not necessarily mean service has deteriorated. It often means the threshold for complaint has been lowered — by a new channel, a prompted communication or a process change that reduced the effort of complaining. The underlying dissatisfaction may have existed for months. Conversely, stable complaint volumes do not confirm stable satisfaction. They confirm only that the complaint threshold has not been crossed more often than before.