When Should You Stop Saying Yes to Custom Work?

Custom work is not automatically bad. The problem is accepting exceptions without knowing what they cost, what they teach you, or what business they are pushing you to build.

For a service business, agency, consultancy, freelancer, or productized service, the useful question is not simply “Can we do this?” It is “Should this business do this kind of work again?”

A request can be profitable once and still be strategically wrong. Another can look inconvenient today but reveal a repeatable offer worth building. The right response is usually one of four things: standardize it, re-scope it, re-price it, or decline it.

Custom work creates variability. Every exception can change what has to be sold, staffed, explained, delivered, reviewed, supported, and repeated.

McKinsey's research on project-based services highlights accurate scope and pricing as major determinants of project margin and notes that customer changes become difficult to execute profitably when change-management processes are weak.

The cost of a request is therefore rarely the request itself. A “small” addition can create another approval cycle, revision, management task, specialist dependency, support obligation, schedule disruption, or precedent for the next customer.

Ask:

Would we want five more customers asking for this?

If the answer is no, pause.

Good custom work usually:

• serves your target customer
• strengthens an existing capability
• improves the core offer
• creates reusable knowledge or assets
• has economics you understand

Dangerous custom work often:

• requires a new capability for one account
• creates permanent support work
• pulls the team away from the core offer
• changes the business model for one customer
• must be rebuilt from zero every time

Revenue is one input. Strategic fit is another.

Do not count only delivery hours.

A useful internal model is:

Incremental cost = delivery time + management time + specialist cost + rework risk + support burden + opportunity cost

Opportunity cost matters when the request consumes a scarce founder, designer, engineer, strategist, or salesperson.

Customer-profitability research also shows why revenue is not the same thing as economic value: customers and orders can require very different levels of effort and resources to serve.

The better question is not “How much are they paying?” but “What does this account consume to produce that revenue?”

Custom work becomes more attractive when the exception teaches you something reusable.

Suppose a client asks for a reporting dashboard. You might discover:

1. it is a one-off nobody else needs;
2. several target customers have the same problem; or
3. the capability belongs in your standard offer.

The second and third cases may justify the work because they create future leverage.

But do not call every unusual request “R&D.” Look for evidence that another target customer would pay for the capability.

A healthy engagement makes the boundary visible:

Outcome → Deliverables → Assumptions → Exclusions → Timeline → Price → Change process

If you cannot explain what is included, what is excluded, how changes are handled, and what additional work costs, you are not really accepting a defined piece of work.

Scope changes are normal. The objective is not to prevent them. The objective is to make their economic consequence visible before the work is performed.

Absorb a small request when most of these are true:

• it is genuinely small;
• it does not materially change the promised outcome;
• it fits the existing process;
• it does not delay important work;
• it improves the customer experience;
• it does not create a harmful precedent;
• the account remains economically healthy.

A five-minute favor can be harmless. Ten five-minute favors can become a new unpaid service.

Re-scope when the customer wants a legitimate outcome but the original agreement no longer describes the work.

Re-price when the work is valuable but the original economics no longer support it.

This matters especially for:

• senior-level involvement
• accelerated delivery
• additional revisions
• new integrations
• additional research
• new deliverables
• ongoing support
• unusual operational requirements

The commercial conversation should be explicit rather than emotional: the request changes the scope, so the timeline, deliverables, or price changes too.

Decline when several tests fail at once:

Poor strategic fit + weak economics + high complexity + low repeatability + high opportunity cost.

At that point, accepting the work can make the company busier without making it better.

Harvard Business Review's research on unprofitable customers adds an important caution: low profitability alone does not automatically mean “fire the customer.” Strategic value, the possibility of renegotiation, and the broader consequences should also be considered.

The same principle applies to custom work.

Do not confuse unusual with valuable.

You do not need a mathematical score. The purpose is to move the conversation from “Can we please the client?” to “Is this a good business decision?”

One custom request is an event.

Five similar requests are data.

Repeated exceptions may indicate:

• a missing standard package
• an unclear offer
• a pricing problem
• a customer-segment problem
• a product opportunity
• a delivery-process weakness
• an expectation that was not defined during the sale

The repeated request should trigger research, not automatic acceptance.

Large customers create another risk. A business may accept unreasonable customization because an account represents substantial revenue. That can make the company more dependent on the customer while simultaneously making delivery more expensive.

Customer profitability research explains why revenue alone is an incomplete measure of customer value: different customers can consume very different levels of effort and resources.

The danger becomes greater when one customer's requirements start shaping the entire roadmap or delivery system.

At that point, the customer is no longer simply buying the business's offer. The business is reorganizing itself around the customer.

Every custom request should become one of four states:

1. Standard - included because it fits the existing offer.
2. Change - separately scoped and priced.
3. Experiment - accepted deliberately to test a repeatable opportunity.
4. Decline - rejected because it does not fit the business.

This creates organizational memory. Later, you can see which exceptions became valuable products and which simply consumed capacity.

The mature question is not “Should we eliminate custom work?”

It is:

Which exceptions increase the strength of the business, and which exceptions quietly weaken it?

Before the next “Can you just add one more thing?” request, ask:

Does it fit? Does it pay? Can it repeat? Can we define the boundary?

If the answers are clear, make the commercial decision explicit. If they are not, do not let urgency make the decision for you.

This article synthesizes research on customer profitability, project-based service economics, scope changes, and customer-specific resource allocation. It does not treat any single source or case as universal proof.

Key sources consulted:

• Harvard Business Review - “The Right Way to Manage Unprofitable Customers” (2008)
• Harvard Business Review - “How Valuable Are Your Customers?” (2014)
• McKinsey - “The winning moves in project-based services”
• Management Science - “The Strategic Value of High-Cost Customers” (2013)
• Journal of Marketing Research - “Linking Customer Management Effort to Customer Profitability in Business Markets” (2004)

This is general business information, not individualized financial, legal, or professional advice.

This article should be read alongside the nearest SOE Knowledge Objects covering the same problem, stage, and next decision. Use those related pages to compare evidence, context, and implementation choices rather than treating this page as a standalone rule.

Related reading: \2 · \2