When Every Department Optimizes Its Part but the Whole Process Gets Worse

Observable condition

Sales improves response targets. Finance tightens controls. Operations maximises utilisation. Procurement batches work for efficiency. Support protects its queue. Each function can show a reasonable local decision, yet customers wait longer, handoff queues grow or the overall process needs more coordination than before.

The monthly reviews look healthy. Each department reports its own measures, most of them green. Yet the customer's experience of getting an order fulfilled, a problem resolved or an invoice corrected has not improved, and in some cases has worsened. When leaders ask why, each function can explain convincingly that the delay is happening somewhere else.

Realisation

A company is not experienced by customers as separate departments. Work moves horizontally across those departments from demand to fulfilment, service, billing and resolution. If every function optimises only its own measure, the result can be a collection of locally efficient steps connected by increasingly inefficient waits and handoffs.

The point that is easy to miss is that the departmental scorecards may all be accurate. Nobody is misreporting. The scorecards simply do not measure the time work spends between departments, which is often where most of the elapsed time sits.

Realisation prompt: for your most important customer journey, does anyone measure the total time from request to completed outcome—and who would be accountable if it rose?

Identification

This is a local-optimisation and end-to-end operating-model problem.

Diagnosis

The mechanism is usually structural: departmental metrics reward local utilisation or risk reduction, queues are invisible outside the function, handoff costs belong to nobody and there is no accountable owner for end-to-end flow. Lean value-stream thinking explicitly distinguishes improving the whole flow from merely optimising individual parts. The operating problem is not that departments are behaving irrationally; it is that the system can reward rational local behaviour that produces a weaker total outcome.

The Lean Enterprise Institute's description of value-stream mapping puts this as improving the whole, not just optimising the parts, and stresses that the map shows how information flow and material flow link together. In a separate Lean Enterprise Institute article on business and IT value streams, Steve Bell describes functions and departments seeking to fully utilise their assets and optimise their own performance, often at the expense of the value stream.

How local improvements make the whole worse

Batching for efficiency: a team that processes requests once a day reduces its own switching cost but adds up to a day of waiting for everyone downstream.
Utilisation targets: a team kept fully busy has no slack, so any variation in arrivals becomes a queue.
Risk controls without flow ownership: an extra approval reduces one function's exposure but adds a handoff, and the delay it causes is nobody's metric.
Speed targets at the front: a faster sales response can push more incomplete orders into fulfilment, which then spends time chasing missing information.

An illustrative example

Illustrative only, not a client case. A business sets sales a target to confirm orders within an hour. Sales meets it by confirming before all delivery details are collected. Operations, measured on utilisation, groups orders into scheduled runs. Finance, measured on credit risk, holds orders above a threshold for review once a day. Each target is met. The customer waits for three separate queues, and the operations team spends time phoning customers for the details that sales did not collect.

Warning signs

Every department hits its targets while customer complaints about delay rise.
Escalations are resolved by senior managers negotiating between functions.
Work-in-progress is visible within teams but not across them.
Improvement projects report local efficiency gains with no change in total cycle time.
Nobody can say how long a typical case spends waiting rather than being worked on.

Commercial impact

The value wrapper is cycle time, margin, customer experience and strategic capacity. Waiting accumulates between teams. Managers spend time negotiating priorities. More work-in-progress can exist without more completed value. Technology investments may accelerate individual steps while leaving the cross-functional constraint untouched.

There is a cash dimension where the flow ends in billing: work that waits longer between departments is also invoiced later. And management attention spent mediating handoffs is attention not spent on growth.

Common misidentification

The usual response is to ask each department to “work better together”. Collaboration matters, but it cannot compensate indefinitely for incentives, measures and workflows that optimise different objectives. If the system rewards local success, recurring cross-functional friction is a design signal.

A second misidentification is to buy a tool for the department with the loudest complaint. If that department is not the constraint in the whole flow, its improvement may simply move the queue elsewhere.

Possibility

A stronger model adds an end-to-end view alongside functional accountability. The business can define the value stream, measure total cycle time and completion, make queues visible, assign ownership across handoffs and evaluate local improvements by their effect on the whole process.

This does not mean abolishing departments or their measures. Functional expertise and controls remain necessary. The change is that a flow-level measure sits above them, and local decisions are tested against it.

Intervention

Choose one important customer or revenue journey and map it from trigger to completed outcome. Record actual processing time, waiting, handoffs, queues, rework and ownership. Then compare departmental metrics with what the end-to-end customer or commercial outcome requires. Improvement priorities should follow the constraint in the whole flow, not simply the loudest local pain.

Map the current state as it actually runs, including information flow between teams, not only the steps within them.
Measure waiting separately from working time at each handoff.
Name an owner for the end-to-end flow with authority to resolve cross-functional trade-offs.
Add one shared flow measure, such as total cycle time, to every participating department's review.
Review each proposed local change against its effect on that shared measure before approving it.

Trade-offs

End-to-end ownership can conflict with existing reporting lines, and a flow owner without authority becomes another coordinator. Shared measures can also dilute accountability if every department can blame the others. The practical answer is to start with one value stream, give its owner clear decision rights over handoff rules, and keep departmental measures that remain useful while removing those that actively work against flow.

Practical diagnostic questions

Who owns the complete process across departmental boundaries?
Which local metrics can improve while total cycle time worsens?
Where does work wait between otherwise efficient teams?
Can leadership see work-in-progress and queues across the whole flow?
Are improvement projects measured by local output or completed customer value?
Which recent local improvement had no visible effect on the customer?

Bottom line

If every department can prove it is improving while the overall process gets worse, the missing layer may be end-to-end operating ownership. Making the whole flow visible and accountable lets the business judge local decisions by what they do for the customer and the margin, not only by what they do for the department.

Sources and further reading

Value Stream Mapping Overview - Lean Enterprise Institute (last verified 10 October 2026)
Lean Business-IT Integration, Part Two: Obstacles to Value-Stream Transformation - Lean Enterprise Institute (last verified 10 October 2026)