Most businesses have some way of hearing from customers. Complaints are logged and resolved. Praise is collected and shared. Perhaps a satisfaction score is reported each month. All of this may be working well, and yet a large group of customers appears in none of it, because they have nothing strong enough to say.
They are not delighted and not upset. They reorder without enthusiasm and would switch without drama if something slightly better came along. They generate no complaint, no testimonial and no survey response. Businesses act on things that arrive on someone’s desk. Where nothing arrives, there is no owner, and where there is no owner, there is no action. Nobody decided to ignore these customers. Nothing ever required a decision.
This article explains why customer feedback systems hear mostly from the extremes, what silence can actually mean, why a single score can hide very different situations, how to read customer behaviour instead of waiting for opinions, and how to make sure someone is responsible for the quiet middle.
Feedback is filtered by intensity
Every feedback channel has a threshold. A complaint requires enough irritation to make the effort worthwhile. A testimonial requires enough enthusiasm to lend your name to someone else’s marketing. A survey response requires enough of either to bother answering. All three thresholds filter in the same direction: towards the extremes. A business’s knowledge of its customers is therefore built from a sample selected on the very thing it is trying to measure.
What silence can mean
Silence is often read as satisfaction. It can also mean:
- Indifference: the relationship works but is shallow. There is no attachment to lose because none formed.
- Resignation: the customer decided long ago that saying something changes nothing. In the data, they look identical to contented customers. Under competitive pressure, they behave very differently.
- Normalisation: an irritation is common across the industry, so customers no longer notice it, until a competitor removes it.
- Departure already decided: the customer has chosen to leave and sees no reason to explain.
Non-response is information
Survey response rates are often treated as a data quality issue. They are also evidence. People who do not respond differ from those who do, so the quiet middle is under-represented in exactly the data meant to describe it. A rising response rate is not necessarily good news. People tend to answer when they have something to say.
One score can hide two very different situations
The Net Promoter Score, a widely used loyalty measure developed by Fred Reichheld, is calculated by subtracting the share of customers who are unlikely to recommend you from the share who are very likely to. Customers in between count only in the total.
That design has a consequence. Two businesses with identical scores can have opposite patterns. One is polarised, with strong advocates and active critics. The other is uniformly lukewarm, with almost nobody at either end. The score cannot tell them apart, and they need nearly opposite responses: the first has a service problem to fix, while the second has no strong relationship at all. A headline score is useful, but publish the full distribution beside it. Also check that the definitions and survey questions behind your score have not changed over time, because a quiet change can break a trend without anyone noticing.
The signals you already own
Behaviour is a signal that does not depend on anyone feeling strongly. Useful behavioural indicators include:
- Purchase frequency and how it is trending for each customer.
- Order size and the range of products or services bought.
- Time to reorder or renew, compared with each customer’s usual pattern.
- Response time to your communications.
- Channel shift, such as moving from personal contact to the cheapest self-service channel.
- Payment behaviour, such as paying later than usual.
These indicators usually already sit in accounting, ordering or customer management systems. They are rarely read as customer intelligence, because they arrive as operational data owned by people who are not looking for it.
Make the middle somebody’s job
Build a simple signal map. For each important customer group, record what signal it generates, who receives it, what triggers action and what a reliable signal would cost to obtain. Then apply one rule: every important customer group has a named owner and a defined trigger for action, or is explicitly declared unmanaged, with a date to review that decision.
Choosing not to manage a group can be reasonable, because attention is limited and signals from the middle cost effort to obtain. Leaving a group unmanaged without ever deciding to is not a decision at all.
Three tests
- The intensity test: what share of last quarter’s customer actions were triggered by a customer contacting you, rather than by someone analysing data? If almost all of them, the business has a complaints process, not a customer strategy.
- The non-response test: who did not answer your last survey, and are they different from those who did? Finding out requires deliberate follow-up with a sample of non-respondents.
- The silent loss test: of the customers you lost last year, how many had no complaint, support request or survey response in the year before they left? That number measures the size of the blind spot.
Build a simple behavioural report
A useful behavioural report can be built from data most businesses already have:
- Export each customer’s order history for the past one to two years from your accounting or ordering system.
- Calculate each customer’s usual pattern, such as average orders per month and average order value over the past year.
- Compare the most recent period with that pattern.
- Flag significant changes, such as ordering frequency falling more than 30% below the customer’s own average, order value dropping sharply, or a customer no longer buying a product line they used to buy regularly.
- Assign each flagged customer to a named person with a deadline for contact.
- Record what is learned and whether any action followed.
Comparing each customer with their own history matters, because a small customer ordering less than usual may be more significant than a large one ordering at its normal level.
Contact the middle deliberately
Once behavioural indicators flag a customer, someone needs to act. Useful approaches include a short call from an account manager, a visit for larger customers, or a brief, specific question by email. The aim is not to sell but to understand: how is the relationship working, what has changed in their business, and is there anything that would make dealing with you easier? Many quiet customers will tell you something useful when asked directly, even though they would never volunteer it.
What to ask
A short conversation with a quiet customer works best with open, specific questions:
- How have things changed in your business over the past few months?
- How well is our product or service working for you at the moment?
- Is there anything about dealing with us that takes more effort than it should?
- Who else are you talking to or buying from for these needs?
- Is there anything you need that we do not currently offer?
Listen more than you talk, avoid turning the conversation into a sales pitch and thank the customer for their time. Many will mention something they would never have raised in a complaint or survey.
Turn what you learn into action
Individual conversations are useful, but patterns are more valuable. Record the reasons customers give in a few simple categories, such as delivery, price, product range, service, competitor activity or changes in the customer’s own business. Review the totals each month or quarter and route them to the people who can act: operations for delivery problems, product decisions for range gaps, pricing for competitive pressure. A short summary of what quiet customers are saying, shared with the leadership team, often reveals issues no complaint ever raised.
Keep your measures stable
The value of customer measures grows with consistency. A business that uses the same behavioural definitions and survey questions for several years can tell whether its quiet middle is gradually drifting away or strengthening. Frequent changes to questions, scales, timing or definitions destroy that ability, often without anyone noticing. Choose simple, durable measures, record any changes and their dates, and resist redesigning the approach every year.
Value the middle by business model
How much the quiet middle matters depends on how the business makes money. In a repeat-purchase business, the middle represents tenure at risk and revenue already assumed in forecasts. In a referral-driven business, it represents customers who would recommend you if asked but are never asked. Either way, it is usually the largest group of customers, and small changes in its behaviour have large effects.
A worked example
This is an illustration. A supplier of industrial consumables has about 400 business customers. It receives roughly 20 complaints a year and reports a stable satisfaction score each quarter. The owner assumes most customers are happy.
The owner runs the silent loss test. Of the 48 customers lost last year, 35, about 73%, had made no complaint, support request or survey response in the twelve months before they stopped ordering. Looking back at their order history, most had reduced their ordering frequency three to six months before stopping.
The business introduces a monthly report flagging any customer whose ordering frequency has dropped more than 30% below their own twelve-month average. Each flagged customer is assigned to an account manager, who calls within two weeks. In the first year, about 60 customers are flagged. Around 22 mention specific issues, such as inconvenient delivery windows, a new competitor’s representative visiting regularly, or a change in their own staff who did not know the products. Many issues are easily fixed.
The following year, customer losses fall from 48 to 31. The satisfaction score barely moves, but the business now knows far more about its largest group of customers than the score ever told it.
How this applies to a small Australian business
Small businesses often know their loudest customers well and their quiet majority hardly at all. Practical steps:
- Run the silent loss test on last year’s lost customers.
- Set up a simple behavioural flag, such as a drop in ordering frequency.
- Assign the middle to a named person with a clear trigger for action.
- Follow up a sample of survey non-respondents.
- Report the distribution of scores, not just the headline.
- Keep survey definitions stable so trends remain meaningful.
- Respect privacy and marketing rules: if you collect or use customer data, check your obligations under the Privacy Act, and follow the Spam Act for marketing emails. The Office of the Australian Information Commissioner and the Australian Communications and Media Authority publish guidance.
The articles on customer lifetime value, acquisition cost and retention and customers are more than revenue cover related ideas.
Signals worth watching
- Customer losses with no prior contact of any kind.
- Response rates falling or changing by customer group.
- Almost all customer actions triggered by inbound contact.
- A stable score over a widening spread of opinions.
- Falling order frequency, narrowing range or a shift to self-service among middle customers.
- Changes to survey wording or timing that nobody recorded.
Common mistakes
- Reading silence as satisfaction.
- Relying on complaints and praise as the whole picture.
- Reporting a single score without its distribution.
- Treating non-response as a nuisance rather than evidence.
- Ignoring behavioural data already in your systems.
- Leaving the middle without an owner.
Frequently asked questions
How often should we check behaviour? Monthly is usually enough for most business-to-business suppliers. Faster-moving consumer businesses may need weekly checks.
Does this apply to consumer businesses? Yes, though the methods differ. Purchase frequency, basket size, loyalty program activity and email engagement can all show quiet customers drifting away, and short, optional check-ins can gather their views.
Will customers mind being contacted? Most business customers appreciate a short, genuine check-in, especially if it leads to something being fixed. Keep it brief and focused on them, not on selling.
What if we have too many customers to call? Prioritise by value and by the size of the behavioural change, and use short written check-ins for smaller accounts.
Who should own the quiet middle? Usually someone in sales or customer service with time allocated for proactive contact, supported by whoever can produce the behavioural report. The important thing is a named person and a clear trigger, not a particular job title.
Should we stop using satisfaction scores? No. Keep them as a headline, but add the distribution and behavioural indicators so the score is not the only lens.
Questions to ask
- What share of our customers gave us any signal last quarter?
- Who is responsible for customers who said nothing?
- Would we notice if our score held steady while opinions polarised?
- How many lost customers gave us a warning we could have read?
- What would a reliable signal from the middle cost, and what would we do with it?
- Which of our measures depend on strong feelings, and which on behaviour?
Bringing it together
A satisfaction score describes the difference between two extremes. Most of the business sits in between. Feedback systems that rely on customers volunteering will always hear mostly from the loudest. Read behaviour, test who is not responding, measure silent losses, publish distributions and give the quiet middle a named owner with clear triggers for action. The answers will be less dramatic than the extremes, and often more valuable.
Source: KEVOS notes. The Net Promoter Score was developed by Fred Reichheld. Examples and figures in this article are illustrations. This article is general information, not legal advice.
