Law 6: Local Optimisation Breaks Global Performance

The system does not care how efficient each part becomes. It only rewards the extent to which inputs convert into outcomes that customers value and the business can monetise.

In almost every struggling organisation, the same pattern eventually reveals itself. Individual departments improve their own metrics while the company as a whole slows down.Engineering increases velocity. Product expands the roadmap. QA raises test coverage. Finance reduces costs. HR introduces new processes and governance layers.

Each function can present evidence of progress. Dashboards look healthier, reports grow more sophisticated, and managers defend their local achievements with confidence.

Yet customers wait longer, teams grow more frustrated, and competitive advantage erodes.

This paradox sits at the heart of the sixth law of systems:

Local optimisation breaks global performance.

Among all principles of systems thinking, few carry greater practical importance.

The first five laws prepare the ground for this one.

• Law 1 reminds us that unused work has no value.
• Law 2 warns that every addition increases entropy.
• Law 3 shows that dependencies tax flow.
• Law 4 demonstrates that delayed feedback compounds failure.
• Law 5 explains that systems optimise for what they reward and tolerate.

Law 6 brings these principles together.

When teams optimise locally, they often add complexity, create new dependencies, delay learning, reward the wrong behaviours, and produce outputs that never translate into meaningful use.

Organisations rarely fail because people work too little. More often, they fail because every part improves independently while the system itself loses coherence.

The Seductive Logic of Local Success

Local optimisation feels sensible, even responsible.

Each team receives objectives, KPIs, and incentives. Managers naturally concentrate on the indicators they can influence directly. Success appears measurable and accountability seems clear.

The flaw lies in a deeper misunderstanding. Systems do not behave as a simple sum of their components.

A racing car equipped with the best engine, brakes, suspension, and tyres does not automatically become the fastest machine on the track. Performance emerges from the interaction and tuning of parts, not from isolated excellence.

Organisations follow the same rule. When departments pursue their own targets without sufficient regard for the whole, interfaces deteriorate, coordination costs increase, and throughput declines.

The Motorway Analogy

Imagine widening one section of motorway from two lanes to six.

Traffic accelerates briefly through that segment, creating the impression of improvement. A few kilometres later, however, the road narrows back to two lanes.

The result does not include higher overall throughput. Instead, a larger queue forms at the next constraint. The local enhancement simply transfers and amplifies the bottleneck.

Companies behave in much the same way. Accelerating coding without strengthening requirements, testing capacity, deployment automation, or decision-making merely shifts congestion downstream. Work accumulates, lead times expand, and frustration intensifies.

Goldratt and the Theory of Constraints

Eliyahu Goldratt built an entire management philosophy around this principle in his influential book The Goal.

A system's throughput depends on its principal constraint.

Improving activities that do not constrain the system may generate visible effort, but rarely changes the outcome in any meaningful way. This point carries profound implications for technology organisations.

Imagine a company investing heavily to improve engineering productivity. Development environments become faster, deployment pipelines shrink from hours to minutes, and teams double their output.

At the same time, marketing identifies strong customer signals, but product management takes months to translate those signals into prioritised and actionable work.

In such a situation, engineering may operate with remarkable efficiency, yet overall Time to Market barely improves.

Time to Market encompasses far more than software delivery. It includes the time required to detect customer pull, convert it into decisions, deliver a solution, and learn from real-world usage.

The true constraint does not reside in software delivery. It resides in the organisation's ability to convert market insight into decisions.

Optimising beyond the constraint produces diminishing returns. And Goldratt proposed a disciplined sequence:

1. Identify the constraint.
2. Exploit it fully.
3. Subordinate the rest of the system.
4. Elevate the constraint.
5. Repeat.

Most organisations do the opposite. Every department pursues optimisation regardless of where the true limitation resides. The outcome looks impressive from a distance, but amounts to motion without progress.

Toyota Understood This Decades Ago

Toyota never treated departmental efficiency as the ultimate objective.

The Toyota Production System focused on end-to-end flow, systemic stability, and customer value.

A machine operating at full utilisation meant little if downstream processes could not absorb its output. In fact, overproduction ranked among the most dangerous forms of waste because it concealed underlying problems and increased inventory.

This philosophy reached far beyond the factory floor. Engineering, procurement, logistics, and product planning operated as interconnected parts of a single system designed to deliver value to the customer with minimal waste.

A similar lesson emerged at Amazon during its formative years. Teams did not optimise for the number of services launched or lines of code written. They worked backwards from the customer and measured success by faster delivery, better reliability, and improved customer experience.

The same principle explains why many European manufacturers lost ground in the automotive transition. Several companies invested heavily in electric vehicle programmes, battery plants, and public messaging while neglecting the broader system of customer demand, charging infrastructure, production economics, and supply chain resilience. Local investments looked impressive, yet global competitiveness weakened.

Many technology companies continue to repeat mistakes that industrial engineering solved decades ago.

The Feature Factory Illusion

This law explains why so many software organisations appear extraordinarily busy while delivering surprisingly little.

Product teams maximise feature count. Engineering teams maximise story points. QA teams maximise defect detection. Platform teams maximise ticket closure. Executives maximise reporting activity.

Every function reports progress.

Customers, however, do not care about internal metrics, story points, or the apparent efficiency of individual departments. They care about how quickly an organisation detects meaningful signals, converts them into decisions, and delivers solutions that address real problems.

This is where the notion of customer pull becomes essential.

Traditional value stream mapping offers a useful representation of how work flows through an organisation, but it often begins too late in the process. The true impulse does not originate in backlog refinement or software delivery. It begins with customer needs, market signals, and emerging opportunities.

The most effective organisations therefore optimise not only the delivery stream, but the entire sensing and response system. They seek to capture external signals rapidly, interpret them accurately, and convert them into action with minimal delay.

Who would not want to improve their ability to detect customer demand and respond more effectively than competitors?

This broader perspective explains why engineering productivity alone rarely guarantees faster Time to Market. If product management, decision-making, or strategic prioritisation cannot translate customer pull into executable work, gains in delivery efficiency will have only marginal impact.

A company may spend three quarters developing a sophisticated feature that almost nobody uses. Another may invest two years building an orchestration platform that never enables a single production use case. A payroll SaaS provider may devote seven years to development without establishing a genuinely reliable production pipeline.

In each case, local indicators improve while the global outcome remains largely unchanged.

Goodhart's Law in Practice

Charles Goodhart captured the danger succinctly: when a measure becomes a target, it ceases to be a good measure.

Teams begin serving the metric rather than the mission.

At that point, organisations often descend into a familiar form of corporate theatre. People run from meeting to meeting, chasing indicators, updating dashboards, and reporting apparent progress, much like headless chickens reacting to numbers that have become disconnected from reality.

Activity intensifies, yet understanding declines.

Over time, the organisation confuses numerical performance with genuine customer impact. Dashboards turn green while the business weakens underneath.

The Biological Parallel

Living organisms do not optimise organs independently.

The heart adjusts to oxygen demand. The immune system responds selectively. Cells divide within strict regulatory mechanisms.

Each component functions in service of the whole organism.

Cancer represents the most destructive form of local optimisation.

A subset of cells maximises its own growth without regard for the larger system that sustains it. What appears as local success ultimately destroys both the host and the cells themselves.

Many corporations display an eerily similar pathology.

Why Organisations Fall into This Trap

Local optimisation persists because it offers political convenience and, perhaps more importantly, flatters executive ego.

Departmental targets remain easy to define. Managers gain visible ownership. Executives receive polished dashboards that suggest control and progress. Responsibility stays neatly compartmentalised, allowing each leader to point to a set of improving numbers and claim success.

This arrangement creates a powerful illusion.

Leaders can present growth in velocity, utilisation, feature output, or cost reduction as evidence of effective stewardship, even when the overall system delivers little additional value to customers.

The price, however, proves substantial.

Every organisation functions as an input-output system.

Capital, talent, time, and market information enter the system. Products, services, customer satisfaction, and financial returns emerge at the other end.

When one function optimises itself at the expense of the whole, the organisation breaks the conversion mechanism that transforms inputs into value.

Local efficiency may improve while overall return on investment stagnates or declines.

In its simplest form, business performance can be viewed as:

Customer value created ÷ Capital, talent, and time invested

Compartmentalised responsibility erodes systemic accountability. No one owns the whole, and each function can declare success while the enterprise quietly deteriorates.

In that sense, local optimisation does not merely distort performance. It destroys economic efficiency, protects status, reinforces hierarchy, and sustains the comforting narrative that leadership remains in control while the system itself drifts further from reality.

The Executive Responsibility

The purpose of management is to optimise the organisation's ability to convert customer pull into profitable outcomes.

Every budget, hire, and initiative represents a bet on the system's ability to transform scarce resources into customer value and financial return.

Senior leadership exists to optimise the system rather than the silos. That responsibility demands more difficult questions.

• Where does work actually stall?
• What constrains throughput?
• Which activities create measurable customer value?
• Which metrics reward behaviour that harms the whole?
• Where do local improvements conceal global decline?

The answers often expose uncomfortable truths.

The most celebrated initiatives may contribute little. Highly praised departments may worsen system performance. Attractive dashboards may conceal structural decay.

Real Performance Emerges from Coherence

The strongest organisations do not chase isolated efficiency.

They align every component around a shared purpose and manage the interactions between them with discipline.

Purpose provides direction. Feedback enables learning. Constraints focus improvement. Metrics reinforce behaviour. Coherence converts local effort into meaningful outcomes.

Only under these conditions does local excellence strengthen global performance.

In a fluid organisation, teams do not optimise their own activity in isolation. They optimise the organisation's collective ability to sense, decide, deliver, and learn as one coherent system.

The Law in One Sentence

When every part tries to win independently, the system loses collectively.

The Economic Reality

Local optimisation is not merely an operational mistake. It is an economic failure disguised as managerial success.

The purpose of an organisation does not involve maximising the efficiency of each department. Its purpose involves converting customer pull into profitable outcomes more effectively than competitors.

When that conversion mechanism functions smoothly, value compounds. When departments optimise in isolation, the organisation consumes more capital, more time, and more talent for progressively weaker returns.

Final Reflection

One of the enduring illusions of modern management lies in the belief that aggregate departmental success guarantees organisational success.

Systems do not function in that manner.

Performance emerges from relationships, interactions, and alignment rather than isolated achievements.

The objective does not involve building the best departments. The objective involves building the best system.

The companies that endure do not optimise departments. They optimise the conversion of reality into value.

That distinction separates organisations that compound value over decades from those that slowly optimise themselves into irrelevance.