Operations and supply chain strategy · Module 1

Operations strategy, process optimization and lean systems

How operations choices create or destroy competitive advantage, and what the laws of flow and the logic of lean say about improving any process.

Beyond the IB course UndergraduateMBA About 25 minutes · 2,997 words · checked 16 November 2026

From the IB course to this module

The IB course covers operations objectives (5.1), production methods from job to mass customization (5.2), lean production and just-in-time at HL (5.3), and capacity utilization and productivity at HL (5.6), each as a separate topic and lean mainly as a list of wastes and methods.

This module goes further in three ways. It treats operations as a pattern of strategic choices that must fit each other and the way the business competes. It adds the laws of flow (Little's law, bottlenecks and the non-linear link between utilization and waiting), which explain why some improvements work and others only move a queue. And it treats lean as a learning system, with evidence on when imitation fails. The capacity and productivity calculator applies the IB versions of these ideas.

Section 01: Conceptual foundation

Operations is where a strategy becomes something a customer experiences: the meal, the mortgage offer, the car. An operation can compete on cost, quality, speed, dependability (keeping promises) or flexibility (handling variety and change), but the layouts, systems and skills that make it excellent at one usually differ from those that make it excellent at another. An operations strategy is the pattern of decisions about capacity, facilities, technology, vertical integration, the workforce, quality and planning that determines which of these the operation is good at.

Two ideas run through this module. The first is fit: operations decisions are good or bad only relative to how the business competes, and they must be consistent with each other. A plant designed for long runs of a standard product will struggle when sales promise fast delivery of variants, however hard its people work. The second is flow: every process, whether it builds cars or approves loans, obeys simple laws linking the work waiting inside it, its rate of output and the time each item spends in it. Lean production is a disciplined way of exploiting those laws, held together by a system of learning that is far harder to copy than its tools.

Section 02: Advanced explanation

Operations as strategy: trade-offs, focus and fit

Wickham Skinner (1969) called manufacturing the missing link in corporate strategy: senior managers left operations to specialists, who optimized for efficiency whatever the business was trying to do, and plants that tried to excel at everything excelled at nothing. His remedy, the focused factory (Skinner, 1974), concentrates a plant, or a plant within a plant, on a narrow set of products, volumes and markets, so that every decision serves one task. Focus is a deliberate choice of trade-offs.

Terry Hill (1985) made the idea market-facing. For each segment, managers identify order qualifiers, which a product must meet to be considered at all, and order winners, which win the sale among those that qualify, and then test whether processes support them. A process serving one segment that wins on price and another that wins on speed is pulled two ways.

Not everyone accepts that trade-offs are inevitable. Ferdows and De Meyer (1990) proposed a cumulative "sand cone": build quality first, then dependability, speed and cost, and all four improve together. The idea of a performance frontier (Schmenner and Swink, 1998) reconciles the views. An operation far inside the frontier its technology allows can improve on several dimensions at once by removing waste; one at the frontier faces real trade-offs until a new technology or operating model moves the frontier out. The managerial question is empirical: is the operation trading off because it must, or because it is badly run?

The laws of flow: Little's law, bottlenecks and utilization

Little's law (Little, 1961) states that in any stable process the average work in process (WIP) equals throughput multiplied by average flow time. It holds whatever the pattern of arrivals and processing times. A mortgage office with 600 applications in progress that completes 50 a day has an average flow time of 12 working days, whatever its staff believe. At a given throughput, the only ways to shorten flow time are less WIP or more capacity.

Capacity is set by the bottleneck, the resource with the highest utilization. An hour lost there is lost to the whole system; an hour saved elsewhere adds nothing to throughput. Goldratt's theory of constraints makes this a routine: find the constraint, never let it idle, subordinate the rest to it, then expand it.

Waiting does not rise in proportion to utilization: it explodes near 100%. Suppose one server finishes 10 jobs an hour and jobs arrive at random. At 8 arrivals an hour (80% utilization) the average wait is 24 minutes; at 9 (90%), 54 minutes; at 9.5 (95%), 114 minutes. Kingman's (1961) approximation explains why: waiting is roughly a utilization term, ρ/(1 − ρ), times a variability term, (ca² + cs²)/2, where ca and cs are the coefficients of variation of arrival and service times, times the average service time. At 90% utilization, halving the variability of service times cuts the wait from 54 minutes to about 34. Less variability substitutes for capacity, and spare capacity is not waste where waiting is expensive.

Process choice and the product-process matrix

Hayes and Wheelwright (1979) matched the structure of a process to that of its product. Products run from unique, low-volume items to standard commodities; processes run from a jumbled job shop through batch and assembly line to continuous flow. Businesses on the diagonal match the two: a job shop for bespoke furniture, a line for cars, a continuous process for chemicals. Off-diagonal positions are a deliberate niche or an expensive mismatch: a flow line cannot recover its fixed costs on one-off items, and a job shop cannot match a line's unit cost on commodities.

The matrix moves. As a product standardizes, the efficient process moves down the diagonal; a firm that lags loses on cost, and one that moves too early loses flexibility customers still value. The IB's job, batch and mass methods are positions on this map. Mass customization (Pine, 1993) tries to escape the diagonal with modular designs, flexible automation and postponement, keeping products generic until late so that variety is cheap at the end of the process rather than throughout it.

Lean as a system, not a toolkit

The Toyota Production System, as its architect Taiichi Ohno (1988) described it, stands on two pillars: just-in-time, making only what is needed when it is needed, and jidoka, stopping the process at the first abnormality so that the cause is fixed rather than inspected out later. Womack, Jones and Roos (1990), reporting an international study of car assembly, popularised the term "lean"; Womack and Jones (1996) distilled five principles: specify value as the customer sees it, map the value stream, make work flow, let the customer pull, and pursue perfection.

In the language of flow, small batches, levelled schedules and pull signals cap WIP, and by Little's law lower WIP means shorter flow time. Quick changeovers make small batches economic, and low inventory exposes problems that buffers used to hide, which is the point rather than a side effect.

Spear and Bowen (1999) argued that the tools rest on four rules: work is specified in content, sequence, timing and outcome; connections between people are direct; flow paths are simple; and improvement uses the scientific method, close to the work. Firms that copy kanban cards without these rules, or without the job security that makes people willing to expose problems, usually get modest results. Lean also assumes fairly stable demand and reliable suppliers; without them thin buffers become fragility, a theme of Supply chain resilience, global sourcing and risk.

Section 03: Models and frameworks

Where each comes from, how to use it, and where it stops being useful.

Skinner (1969, 1974); Hill (1985)

The focused factory and operations strategy

Structural decisions (capacity, facilities, technology, vertical integration) and infrastructural ones (workforce, quality, planning) should form a consistent set serving one defined task, because each task implies different trade-offs.

How to use it

For each segment, agree order qualifiers and winners with marketing, audit every decision area against them and record contradictions. Where segments conflict, separate them physically or organizationally rather than averaging.

Where it fails

Focus can sacrifice economies of scope and strand a plant when its market changes. It assumes trade-offs bind, which is false well inside the frontier.

Wickham Skinner (1969). Manufacturing: missing link in corporate strategy. Harvard Business Review. Wickham Skinner (1974). The focused factory. Harvard Business Review. Terry Hill (1985). Manufacturing Strategy: The Strategic Management of the Manufacturing Function. Macmillan.

Hayes and Wheelwright (1979)

The product-process matrix

A map of product structure (unique to standard) against process structure (job shop to continuous flow). Firms on the diagonal match process economics to product needs, and the right position moves as products mature.

How to use it

Place each product line on the matrix, compare competitors and ask how the life cycle will move it. Use it to time automation and to decide whether to split a plant serving very different products.

Where it fails

It was built for manufacturing; services and platforms fit awkwardly. Flexible automation and modular design let mass customizers sit off the diagonal profitably.

Robert H. Hayes and Steven C. Wheelwright (1979). Link manufacturing process and product life cycles. Harvard Business Review. B. Joseph Pine II (1993). Mass Customization: The New Frontier in Business Competition. Harvard Business School Press.

Little (1961); Kingman (1961)

Little's law and the VUT equation

Work in process equals throughput times flow time in any stable system. Kingman's approximation, a product of variability, utilization and time (hence "VUT"), shows why queues explode as utilization nears 100% and why variability does as much damage as a lack of capacity.

How to use it

Count work in process and measure throughput to estimate flow time. Find the bottleneck by utilization. Before raising utilization targets, estimate the effect on waiting and compare added capacity with less variability.

Where it fails

Little's law describes long-run averages, not the spread of flow times or a backlog after a disruption. Kingman's formula approximates one stage and is least reliable with rework.

John D. C. Little (1961). A proof for the queuing formula: L = λW. Operations Research. J. F. C. Kingman (1961). The single server queue in heavy traffic. Proceedings of the Cambridge Philosophical Society. Wallace J. Hopp and Mark L. Spearman (1996). Factory Physics: Foundations of Manufacturing Management. Irwin. Eliyahu M. Goldratt and Jeff Cox (1984). The Goal: A Process of Ongoing Improvement. North River Press.

Ohno (1988); Womack, Jones and Roos (1990)

Lean thinking and the Toyota Production System

A system built on just-in-time flow and built-in quality, extended through five principles: value, value stream, flow, pull and perfection. Its advantage lies less in its tools than in specified work and scientific problem-solving by the people who do the work.

How to use it

Map the value stream and measure value-adding time as a share of flow time. Attack WIP and batch sizes first, because they set flow time, and make exposing problems part of daily work.

Where it fails

It assumes stable demand, reliable suppliers and a workforce confident that improvement will not cost jobs. In volatile settings, removing buffers lengthens queues and raises fragility.

Taiichi Ohno (1988). Toyota Production System: Beyond Large-Scale Production. Productivity Press. James P. Womack, Daniel T. Jones and Daniel Roos (1990). The Machine That Changed the World. Rawson Associates. James P. Womack and Daniel T. Jones (1996). Lean Thinking: Banish Waste and Create Wealth in Your Corporation. Simon & Schuster. Steven Spear and H. Kent Bowen (1999). Decoding the DNA of the Toyota Production System. Harvard Business Review.

Section 04: Real business applications

Real organizations appear only with dated, cited facts; illustrative examples are labelled.

Real organization, 2023-03

Toyota's production system outside the factory

In 2011 Toyota made its Production System Support Center (TSSC), set up in 1992 to teach the system to suppliers, a not-for-profit open to non-profit organizations. It cited pilots: waits at a Harlem community kitchen run by the Food Bank for New York City fell from more than an hour to 18 minutes after its meal service was redesigned, and just-in-time supply of pharmacy items at Allegheny General Hospital was projected to save $391,000 a year. In 2023 Toyota reported two more: the St. Bernard Project, which rebuilds homes after disasters, cut its average rebuild from 116 days to 60 and raised homes completed a month from 8.6 to 12.8, and the Harbor-UCLA eye clinic halved the time from check-in to discharge.

Interpretation. If the St. Bernard figures are averages from a stable process, Little's law implies that homes in progress fell from about 33 (8.6 a month × 3.8 months) to about 25 (12.8 × 2.0) while output rose by half: shorter flow, less WIP and more throughput at once, the mark of waste removal inside the frontier. The sponsor reports the results without a comparison group: they show what is possible, not what is typical.

Real organization, 2026-04

BYD: vertical integration as an operations choice

BYD describes itself as a company founded in 1994 that has grown across electronics, cars, renewable energy and rail transit, with 2025 revenue above RMB 804 billion and new energy vehicles in more than 120 countries and regions by April 2026. Its account of its technology stresses components it developed itself: the Blade Battery, hybrid power systems, an electric powertrain combining eight functions in one unit, a four-motor drive and its own vehicle operating system. It lists electronics and batteries as businesses alongside cars.

Interpretation. Making rather than buying the battery and much of the drivetrain is a structural operations decision in Skinner's sense. It offers control of the costliest, most differentiating component, faster coordination of product and process design, and the supplier's margin; it costs capital, commitment to one technology path and competition among suppliers. It rests on capabilities built over decades, and can ruin an entrant that tries to master every stage at once (see Northvolt in Scaling, innovation management, failure and recovery).

Illustrative

An emergency department that ran too hot

A hospital (illustrative) removes one doctor from an evening rota, raising utilization from 80% to 92%. Arrivals are unpredictable and consultation times vary widely (coefficients of variation near 1). By Kingman's approximation, the average wait for a 20-minute consultation rises from about 80 minutes (4 × 1 × 20) to about 230 (11.5 × 1 × 20), with knock-on delays to ambulance handovers. Separating minor injuries from complex cases would cut variability within each stream and recover part of the loss without the extra doctor.

Section 05: Case studies

Fictional case, simulated figures

Linden Lane Mortgages: a slow application process

Linden Lane Mortgages (fictional) is a regional lender whose brokers complain that offers take "about a month". Applications pass through a document check, credit underwriting, a review of the external valuation and the issue of the offer. They arrive at an average of 22 a working day; files in progress averaged 440 last quarter; hands-on work per application totals 135 minutes.

Marketing plans a campaign to raise applications to 25 a day. The head of underwriting wants a fifth underwriter; the finance director prefers a fast track in which straightforward applications (70% of the total) take 40 minutes to underwrite and complex ones 120.

Linden Lane Mortgages, capacity by stage at current staffing
Stage Staff Minutes per application Capacity per day Utilization at 22 a day
Document check3304252%
Credit underwriting4702492%
Valuation review2204252%
Offer issue1152879%

Each member of staff has 420 productive minutes a working day; capacity is staff × 420 ÷ minutes per application. Simulated data.

Questions

  1. Identify the bottleneck. Use Little's law to estimate the average time from application to offer, and the share of that time spent working on the file.
  2. If applications rise to 25 a day with nothing else changed, what happens at the bottleneck? Use the utilization term in Kingman's approximation to explain why lead time worsens faster than demand rises.
  3. Calculate underwriting capacity with a fifth underwriter and with the fast track. Which adds more capacity, and which also reduces variability?
  4. The director proposes capping files in progress at 220 by giving new applicants a booking slot. What does Little's law predict for flow time, what must hold for throughput not to fall, and what is the commercial risk?

Section 06: Critical evaluation

What the evidence supports, where the debate is, and what the models leave out.

Three claims here are established beyond serious dispute. Little's law is a theorem, not a hypothesis. The non-linear rise of waiting with utilization under variability is a robust result of queueing theory, borne out from call centres to hospitals. And the view that operations decisions embody trade-offs, to be chosen to fit the strategy, rests on decades of case research, even if its strong form is contested.

The contested questions concern how far the ideas generalise. Trade-offs or cumulative capabilities? The sand cone rests largely on surveys in which better plants report better performance on everything, as the frontier view also predicts for plants far inside the frontier; it does not show that plants at the frontier escape trade-offs. Does lean cause performance? Shah and Ward (2003) found bundles of lean practices associated with better operational performance, but adopters are not a random sample: better-managed plants adopt more practices and might perform better anyway. Frequent implementation failures suggest that practices travel more easily than the management system behind them. Lean in services. The TSSC projects show striking gains, but published cases are chosen by the people who ran them; reviews of lean in healthcare tend to find improvements in waiting and throughput with weaker, less consistent evidence on patient outcomes, partly because demand variability and professional autonomy limit standardization.

Two critiques deserve weight. Labour process researchers argue that lean can intensify work by removing the slack people used for recovery; whether it does depends on whether improvement is done with workers or to them. And thin buffers raise exposure to supply shocks, though defenders reply that the danger lies in hidden single points of failure rather than low inventory itself. Both concern conditions; the laws of flow hold regardless.

Section 07: Decision exercise

A fictional situation. Decide before you read the guidance.

Focus or flexibility at Harrowgate Foods

Harrowgate Foods (fictional) runs one sauce plant for two kinds of customer. Supermarket own-label contracts (70% of volume, 45% of contribution) need the lowest unit cost and near-perfect delivery, with heavy penalties for misses. Premium chilled sauces for restaurants (30% of volume, 55% of contribution) need many recipes, small batches and delivery within 48 hours. Changeovers take up to four hours, so premium orders wait for weekly runs, and a competitor now offers restaurants 24-hour delivery.

  1. A. Plant within a plant

    Split the site into two focused units with their own lines, teams, planning rules and measures. Costs £3.5 million for duplicate equipment; the premium unit would run at about 55% utilization.

  2. B. One flexible plant

    Keep one shared plant, spend £1.2 million on quick-changeover equipment and training to cut changeovers to 40 minutes, and schedule both products daily on the same lines.

  3. C. Focus and outsource

    Dedicate the plant to supermarket volume and move premium production to a specialist co-packer on a three-year contract, keeping recipes and customers. No capital, but about 8 points of margin go to the co-packer, and control of quality and speed weakens.

Judge the options on

  • Fit with the order winners of each segment
  • Capital and transition cost
  • Risk to delivery on supermarket contracts
  • Control of quality and of capabilities that may matter later
  • Reversibility if the restaurant market changes
Read the guidance

Start with the order winners. The supermarket business wins on cost and dependability, which favour long runs and high utilization; the restaurant business wins on variety and speed, which need spare capacity and fast changeovers. A shared plant serves both only if changeovers become nearly costless, so option B stands or falls on whether 40 minutes is credible. Even then the two schedules would compete for the same lines daily, the conflict Skinner described, and the supermarket penalties make the downside asymmetric.

Option A is the textbook answer and the most expensive. A premium unit at 55% utilization looks wasteful, but the VUT logic says that spare capacity buys a 24-hour response to unpredictable orders, and it keeps the more profitable segment's capabilities in-house. Option C buys focus cheaply but hands that segment's order winners, speed and quality, to a supplier who may serve competitors.

A defensible answer is A if the premium segment is the future of the business, B as a staged experiment (prove the changeover time on one line first), and C only if premium is a sideline. The weakest answer is to keep the current plant and ask planners to try harder.

Section 08: Further reading

  • Wickham Skinner (1974). The focused factory. Harvard Business Review.

    The original case for focus, short and still provocative; read it with the 1969 article on manufacturing as the missing link in strategy.

  • Robert H. Hayes and Steven C. Wheelwright (1979). Link manufacturing process and product life cycles. Harvard Business Review.

    The product-process matrix in its first form, with the argument for moving along the diagonal as products mature.

  • Steven Spear and H. Kent Bowen (1999). Decoding the DNA of the Toyota Production System. Harvard Business Review.

    Why copying Toyota's tools rarely copies its results, and the four rules behind the system.

  • Wallace J. Hopp and Mark L. Spearman (1996). Factory Physics: Foundations of Manufacturing Management. Irwin.

    The standard text on the physics of flow, including Little's law, variability and the VUT equation, at a level suited to MBA and engineering students.

  • James P. Womack, Daniel T. Jones and Daniel Roos (1990). The Machine That Changed the World. Rawson Associates.

    The international study of car assembly that popularised the term lean production.

Every work cited here is listed in the platform's sources register. Real organizations appear only with facts from the sources shown; everything else is labelled illustrative or fictional.