Owner growth · Customer portfolio profitability

Should Your Business Stop Serving Unprofitable Customers?

The short answer: do not end a customer relationship because one month looks inconvenient. First measure revenue minus delivery, support, rework, payment and opportunity costs; then try to repair fit, scope, price or process. Exit respectfully only when the relationship remains structurally unprofitable, damages capacity or cannot meet fair boundaries after a documented correction period.

Editorial illustration of customer relationship streams entering a business capacity reservoir while one cracked route passes through a copper recalibration gate
Protect profitable growth by repairing the relationship before redirecting scarce capacity · Original illustration by ThomPerformance

Revenue does not prove that a customer relationship creates value

A large account can feel indispensable because it produces visible revenue, keeps people busy or carries a recognisable name. Yet the same relationship may require persistent rework, unusual support, senior escalation, discounts, slow payment and work that displaces healthier demand. Conversely, a small customer may look inefficient during onboarding but become simple and valuable after the first delivery cycle.

That is why “fire your worst customers” is a poor operating rule. It encourages an emotional decision before the business has separated a temporary delivery problem from a structurally weak relationship. It can also punish a customer for failures created by unclear scope, inconsistent service or inaccurate pricing.

My verdict is: validate, repair, then decide. Measure contribution across a representative period, establish whether the customer receives fair value, identify the constraint and run one bounded correction plan. Reprice, redesign or exit only when the evidence shows which decision protects both parties.

This intent is distinct from choosing between acquisition and retention, which allocates growth attention across the wider customer base. It also differs from moving upmarket and reducing customer concentration. This guide evaluates the economics and operating fit of an existing relationship.

The Customer Contribution & Fit Gate

Before changing terms or ending a relationship, test six connected conditions. A failed condition identifies work to do; it is not permission to blame the customer.

The output should be a range, not false precision. Direct materials and fulfilment are usually visible. Allocating shared salaries, rent or leadership time requires judgement. Record the method and test the decision under a conservative and an optimistic assumption. If the conclusion changes easily, the evidence is not strong enough for an irreversible action.

Calculate relationship contribution before making the decision

Start with a period that reflects normal delivery. Use collected revenue, then subtract direct delivery, attributable support, rework, refunds, payment costs and any acquisition or onboarding cost the business is deliberately trying to recover. Keep fixed overhead visible but separate so an arbitrary allocation does not turn a useful customer into a supposed loss.

Illustrative example — not client proof
Quarterly relationship viewCustomer ACustomer B
Collected revenue$30,000$18,000
Standard delivery cost$15,000$8,500
Support and rework$10,500$1,500
Payment and credit cost$1,500$300
Relationship contribution shown$3,000$7,700
Delivery capacity used460 hours250 hours

Customer A produces more revenue but less contribution and consumes substantially more delivery capacity in this simplified view. The example excludes tax, shared overhead and future value; hourly allocations are assumptions, not accounting advice. It demonstrates why revenue ranking and customer-profitability ranking can lead to different decisions.

The relationship may still deserve repair. Perhaps the business underpriced a valid scope, failed to train the customer, allowed exceptions without approval or created rework through poor quality. A fair decision separates what the customer controls from what the business must fix. Use a qualified finance adviser where the allocation or tax treatment matters.

The Relationship-to-Capacity Reconciliation Loop

This six-stage loop turns one worrying account into a documented business decision rather than a recurring argument.

The commercial lesson should also return to growth. If unsuitable customers keep entering through paid or organic demand, update the offer, qualification, proof and follow-up—not merely the advertising audience. Feed accepted-customer and contribution evidence into conversion measurement so growth systems optimise towards relationships the business can serve well.

Choose the least destructive valid action

EvidenceVerdictBest next moveAvoid
Value is strong; exceptions and rework cause the lossRepair processClarify workflow, approvals, service levels and exception ownershipRaising price without removing the waste
Scope is valid but price no longer covers deliveryRepriceExplain the changed economics and offer a defined scope or tierSurprise fees or retroactive charges
Only part of the relationship fits profitablyReshapeRetain the valuable core and remove, automate or refer the weak componentProtecting revenue by keeping every exception
Loss is temporary and strategically boundedContinue with a capState the investment, learning goal, maximum cost and review dateCalling an indefinite subsidy “strategic”
Terms, risk or behaviour remain unacceptable after a fair resetExitFollow obligations, communicate directly and provide a controlled transitionBlame, public criticism or abrupt abandonment
Evidence is incomplete or allocations decide the outcomeHoldImprove time, delivery and collection data before actingUsing a spreadsheet estimate as certainty

If price is the main mechanism, use the Price Increase Readiness Gate. If the portfolio is profitable but the team is overloaded, use the Revenue Capacity Release Loop. If one customer creates concentration risk, protect the relationship while building replacement capacity rather than forcing a premature exit.

A 60-day customer correction plan

Days 1–10

Reconcile the facts

Choose the period, define contribution, collect delivery and payment evidence, and test the result under more than one allocation assumption.

Days 11–20

Agree the diagnosis

Separate customer-controlled issues from internal failures. Decide what must change in scope, price, process, payment or ownership.

Days 21–50

Run the reset

Communicate the boundary, document the measure and protect service quality. Track exceptions and customer value during the full test.

Days 51–60

Make and govern the decision

Continue, reprice, reshape or exit. Record transition duties, commercial impact, capacity release and the growth rule that should change.

Sixty days is a governance example, not a universal benchmark. Use a window suited to the contract, buying cycle and delivery model. Urgent safety, legal, ethical or credit risks may require faster professional action; long project and annual relationships may require a longer review.

Review paid digital marketing and AI growth services, AI-assisted customer insight, case-study evidence, evidence standards and Thomas's direct operating model. The purpose is not simply to remove low-contribution customers. It is to improve the offer, acquisition and service system so suitable relationships become easier to win and serve.

Practitioner note: I would not send more demand into a business that cannot explain which customer relationships create value. Before increasing budget, I would ask which customers the team serves well, where capacity leaks, which outcomes sales and delivery accept, and whether the growth system can learn from that evidence.

Sources and evidence notes

Sources and current search results were checked on 19 September 2026. Search prioritisation is qualitative; no unverified search volume, universal profitability threshold or client result is used. The Customer Contribution & Fit Gate, Relationship-to-Capacity Reconciliation Loop, decision matrix and 60-day plan are original ThomPerformance analysis. The worked figures are explicitly illustrative.

  1. Australian Government: set up a profit and loss statement
  2. Australian Government: choose a pricing strategy
  3. Australian Government: payment terms
  4. Business Queensland: plan and conduct customer research
  5. Harvard Business School: measuring and managing customer profitability

Frequently asked questions

What makes a customer unprofitable?

A customer is unprofitable when collected revenue does not cover the costs required to acquire, onboard, deliver, support, collect payment from and retain the relationship over a suitable period. The calculation should also expose unusual rework, discounts, delays and capacity consumed, while keeping shared overhead and one-off investment assumptions explicit.

Should a business fire every unprofitable customer?

No. Some relationships are temporarily unprofitable because of onboarding, a service failure, a growth investment or a fixable scope problem. Start by validating the numbers and the customer value. Then repair process, scope, price, payment or fit. Exit only when the problem is structural, material and unlikely to improve responsibly.

How should a business tell a customer it is ending the relationship?

Follow the contract and applicable obligations, communicate directly, explain the operational decision without blaming the customer, honour paid commitments, protect their data and provide a reasonable transition where possible. The goal is a controlled, respectful conclusion—not a punitive surprise. Obtain appropriate professional advice where regulated, contractual or consumer duties apply.

Can a high-revenue customer still be unprofitable?

Yes. Revenue can hide custom work, frequent escalation, senior attention, discounts, returns, payment delay, risk and displaced capacity. Compare collected revenue with the direct and attributable cost-to-serve over a representative period. A large account may still be valuable, but size alone is not proof of contribution.

How long should a customer correction plan run?

Use a period long enough to observe the behaviour creating the problem. Sixty days can suit recurring service work, but project businesses, annual contracts and seasonal demand may need a different window. Set the measures, owner, review date and exit condition in advance so the decision does not drift indefinitely.

Improve the portfolio, not just the customer list

An unprofitable relationship is evidence about pricing, fit, delivery, payment or process. Treat it as a system signal. Validate the economics, correct what the business controls and give a responsible relationship a fair opportunity to improve. When the structural mismatch remains, a respectful exit can protect service quality for the customers the business should keep.

Which customer relationship would look different if you ranked it by contribution and capacity—not revenue alone?

Turn customer economics into a better growth decision

I can connect customer evidence, acquisition, pricing, delivery capacity and measurement into one practical portfolio diagnosis.

Request a customer-portfolio diagnostic

About the author: Thomas Ho is a Paid Digital Marketing & AI Growth Partner helping business leaders connect acquisition, conversion, customer evidence and practical AI to qualified pipeline and revenue.

Free operating template

Stop reviewing paid ads with screenshots and green arrows.

Use the same weekly review structure I use to connect spend with qualified leads, opportunities, pipeline and decisions.

  • Commercial scorecard
  • Creative test log
  • Decision ownership
Discuss your project