Should you fire your customers? A four-question decision path

The idea arrives in a workshop, about an hour in, from someone who has just seen the customer profitability curve for the first time. A slice of customers at the bottom cost more to serve than they pay. Someone says the words: why don’t we just fire them? Firing customers sounds, in that room, like arithmetic.

Different room, different tone. A support lead comes out of a shift review with an agent who was shouted at for twenty minutes by a customer who has done it before. Someone says the same words, and means something else entirely.

Firing a customer means deliberately ending the relationship from your side: closing the account, declining further orders, or telling the person plainly that you will no longer serve them. The phrase covers two situations that have almost nothing in common, and the right answer for one is the wrong answer for the other. There is also a third customer who gets swept up with them, and firing that one is the most expensive mistake of the three.

Key takeaways

“Fire the customer” describes two unrelated cases, the unprofitable customer and the abusive one, and each needs its own rule.
Most unprofitable customers are unprofitable because of how they are served, and changing the offer, channel, or price fixes the economics without ending the relationship.
An abusive customer should be let go regardless of how much they spend, quickly, in plain language, with staff protected first.
The loud customer is usually the one you failed, and firing them removes the evidence while keeping the problem.
Four questions, asked in order, settle most cases in one meeting: are they harming our people, did we cause this, can the offer, channel, or price fix it, and are they unprofitable at any price and over any horizon.
Firing should be rare; a company that does it often is usually cutting its own process failures out of the customer base one account at a time.

What firing customers means

Firing a customer is a decision, made by you, to stop serving someone who would otherwise keep buying. It is deliberate and it is communicated, which separates it from three things it is often confused with.

It is not churn. A customer who leaves because a competitor was cheaper has fired you, which is a different problem with its own quiet warning signs.

It is not a pricing decision. If you start charging for paper statements and some customers move to a competitor, you did not fire them; you gave them an honest choice and they took it.

And it is not a targeting decision. Choosing not to market to a segment, or to stop offering a discount, changes who arrives. Firing changes who is allowed to stay. The distinction matters because the first two are ordinary commercial life and the third is rare, or should be.

Unprofitable customers: change the offer, channel, or price first

Before you fire the unprofitable customer, look at what “unprofitable” is made of.

In most cost-to-serve analyses I have seen, unprofitable means served the expensive way. The customer calls instead of using the app. They return things. They pay late. They buy small quantities often. None of these is a character flaw. Each is a behavior, and behaviors respond to the offer, the channel, and the price. So change those first, in this order.

Change the offer. A customer who orders small quantities often may be perfectly profitable on a subscription, a bundle, or a minimum order. The unit economics were designed for someone else. Redesign them.
Change the channel. A customer who calls is expensive only if calling is the sole way to get something done. If the self-service option does not exist, or does not work, the cost of that call is yours. Look at what they call about. Often it is the same three things, and fixing those makes the customer cheaper without anyone being fired.
Change the price. Some customers are unprofitable at the price set for the average customer and perfectly fine at a price that reflects what they use. A fee for paper statements, a charge for expedited returns, a tier that matches their volume: honest pricing, not punishment. Some customers will leave. That is not firing them. That is letting them choose.
Check whether you made them unprofitable. The customer who calls five times about one order is not expensive because of who they are. They are expensive because the first four calls did not fix it. Much of the cost-to-serve tail is your own process reflected back at you.
Check the horizon. This is where lifetime value earns its keep. A customer who looks unprofitable this year may be early in a relationship that pays for a decade; another may be a loss forever. If you have wondered whether lifetime value is worth the effort, this is the decision where it changes the answer, and there are ways to use lifetime value even when the business runs on quarters.

Only after all that, if a customer still costs more than they pay at any price and on any channel you can offer, is it time to let them go. You stop the discount, close the tier, and point them to a competitor set up for what they need. Done plainly and politely, it reads less like a dismissal and more like a referral.

A worked example, for illustration

Take an imaginary customer with round numbers. They generate $500 of gross margin a year. They phone eight times a year, and each call costs $25 to handle: $200. They return two orders a year at $50 each to process: $100. Net contribution $200, well below the average customer, and squarely in the tail of the profitability curve.

Now look at the calls. Six of the eight are about delivery status. A tracking link in the shipping confirmation removes most of them; say the customer now calls twice. Call cost drops to $50 and net contribution rises to $350. One of the two returns turns out to be a sizing problem the product page could have prevented; fix the page and the return cost halves. Nobody was fired, nothing was repriced, and a customer who looked like a loss is now ordinary. The numbers are invented, but the shape is what a cost-to-serve tail usually looks like once someone asks why.

When to fire a customer: the abusive case

The other case is not about money, and treating it as if it were does harm.

An abusive customer is one who harms your staff: threats, sustained shouting, slurs, the campaign of complaints designed to get someone in trouble. The profitability of that customer is irrelevant. A large account that abuses your people is a large account you should lose, and your staff will remember which way you went.

Three rules. Protect staff first. The agent can end the call, the store can ask the person to leave, and the manager backs them, without an interrogation afterward. Decide quickly. Abuse cases that sit in a queue for review teach everyone that the policy is theoretical. Say it plainly. A short, calm message that names the behavior, states that the relationship is ending, and explains what happens to any open orders or balances. No lecture, no sarcasm, no invitation to argue.

Firing an abusive customer should be rare. Most anger is not abuse. It is a person who has been failed and has run out of patience, which brings me to the third case.

The loud customer you must never fire

The loud customer is not always the abusive customer. Very often the loud customer is the one you failed, and the volume is the evidence.

They are loud because you gave them a reason: the promise you broke, the fee you did not explain, the fourth transfer between departments. Firing them removes the noise and keeps the problem. The next customer who hits it will be just as loud, or, worse, quiet, having decided it was not worth the argument.

Loud customers are also, inconveniently, information. They tell you exactly where the process breaks, at a level of detail no survey will reach. If some of your best customers are among the complainers, treat that as a listening program you did not have to design. Fix what they are shouting about, tell them you fixed it, and watch what happens to the volume.

Unprofitable vs abusive vs loud: telling them apart

From the outside the three look alike: a customer who takes a lot of time and is not pleasant about it. The cause is the difference, and the cause decides the response.

Customer	What you see	What is usually going on	The right response
Unprofitable	High cost to serve, low margin, the wrong end of the curve	Served the expensive way; economics designed for someone else	Change the offer, channel, or price; check the horizon; let go only as a last resort
Abusive	Threats, sustained shouting, slurs, complaints aimed at individuals	Behavior that no fix on your side will change	Protect staff, decide today, say it plainly; money is irrelevant
Loud	Repeated complaints, escalations, public reviews	You broke something and they have not given up on you yet	Fix the cause, tell them, keep them; treat the complaint as research
Loud and unprofitable	Both of the above at once	Your process failure is also what makes them expensive	Fix the failure first; the economics usually follow

The last row is the common one. A customer who costs a lot and complains a lot is very often complaining about the thing that makes them cost a lot. Looking at the person behind the profitability number before acting on the number is the whole discipline here.

A decision path in four questions

When someone says “fire this customer,” walk through these in order and stop at the first yes.

Are they harming our people? If yes, protect staff, decide today, and say it plainly. Money does not come into it.
Did we cause this? If the cost or the anger comes from our own friction, fix the friction, apologize, and keep the customer.
Can the offer, channel, or price make them profitable? If yes, change it, tell them why, and let them decide whether to stay. Some will leave, and that is fine.
Are they still unprofitable at any price, on any channel, over any horizon we believe in? If yes, let them go politely, with a referral if you have one. If you cannot say yes with confidence, you are not ready to fire anyone. You are ready to do the analysis.

The next time the words come up in a meeting, ask which of the four you are on. In my experience the honest answer is usually the second, and the room goes quiet for a moment, because fixing your own friction is more work than closing an account.

What goes wrong when firing customers becomes a habit

A company that fires customers regularly has usually stopped asking the second question. Four things follow.

The tail grows back. Every profitability curve has a bottom slice. Remove it and the customers just above it become the new bottom, with the same process failures making them expensive. A company can fire its way down the curve indefinitely and never fix the thing that made the curve steep, which is the opposite of the work that reduces attrition.

The frontline learns that difficult means disposable. If the easiest way to end a hard conversation is to escalate it toward account closure, hard conversations stop being resolved. The staff who should be telling you where the process breaks become the people who route the evidence out of the building.

“Unprofitable” drifts into a description of a kind of person. Cost to serve can track age, income, language, and comfort with technology. A rule that started as “customers who call too often” can quietly become “customers who are not like our best ones,” and some industries have obligations to serve that make this more than a reputational risk.

And “unprofitable” is often a product problem in disguise. Customers who use a product on the wrong channel at the wrong price are frequently people the product was never designed for, so check whether your product invites a relationship at all before deciding that the people who use it are the problem.

Where to start

Pull the bottom slice of the cost-to-serve curve and label each customer with the main reason they are there: calls, returns, late payment, small orders. Most tails have three reasons, not thirty.
For the biggest reason, estimate how much is your own process. Read a sample of the calls or the return notes. Count how many are about something the company got wrong first.
Write the abuse policy on one page, with the three rules and the name of the manager who backs the frontline, and tell every agent it exists.
Make one change to the offer, channel, or price aimed at the biggest cost pattern, and tell the affected customers why before it lands.
Decide the horizon. Agree, in writing, how many years of a customer’s value the company believes in, so that “unprofitable” is measured against something everyone accepts.
Revisit the list in a quarter. Whoever is still in the tail, at the new price, on the new channel, over the agreed horizon, is a candidate for a polite referral elsewhere. There will be fewer than the first meeting expected.

FAQ

When should you fire a customer?

Fire a customer when they harm your staff, or when they remain unprofitable at any price you can offer, on any channel you can provide, and over any horizon you believe in. The first case is decided in a day and money plays no part. The second is decided only after you have changed the offer, channel, and price, and checked whether your own process created the cost.

How do you fire a customer politely?

Say it in a short, calm message that names the reason, states that the relationship is ending, and explains what happens to open orders, balances, or data. Where the reason is economic rather than behavioral, point the customer to a provider set up for what they need, so the message reads as a referral rather than a dismissal. Do not lecture, do not argue, and do not leave it to the frontline to improvise.

What makes a customer unprofitable?

A customer is unprofitable when the cost of serving them exceeds the margin they generate over the period you measure. In practice the cost usually comes from behaviors such as frequent calls, returns, late payment, and small orders, and those behaviors respond to the offer, the channel, and the price. Many customers who look unprofitable in a year are profitable over the life of the relationship, which is why the horizon matters.

Should you fire a customer who is rude to your staff?

Rudeness and abuse are different, and the answer depends on which it is. A customer who is angry because the company failed them should be listened to and the failure fixed. A customer who threatens, uses slurs, shouts at people repeatedly, or runs a campaign against an individual employee should be let go, quickly and regardless of how much they spend, with the staff member backed publicly.

Can an unprofitable customer become profitable?

Very often, yes. Moving them to a self-service channel, offering a subscription or minimum order, charging for the costly extras they use, or fixing the process failure that makes them call are all ways to change the economics without ending the relationship. Fixing what the company got wrong first is the most common route, because much of the cost-to-serve tail is the company’s own friction reflected back.

Why should you not fire complaining customers?

Complaining customers are usually the ones you failed, and their complaints describe where the process breaks in more detail than any survey will. Firing them removes the noise and keeps the problem, so the next customer who hits it will complain too, or, worse, leave quietly. Fix what they complain about, tell them, and the volume tends to fall on its own.