5. Flow, Takt Time, Pull and Just-in-Time

Introduction and Context

Lean management seeks to create value with fewer resources by removing activities that do not contribute to customer value. In earlier discussions of waste and value-stream thinking, the central concern is identifying what interrupts value creation. This chapter examines how work can be reorganised so that value moves smoothly, at the rate required by customers, and is replenished only when needed.

Flow, takt time, pull and just-in-time (JIT) are closely related, but they are not interchangeable terms. Flow concerns the movement of materials, information and work through a process. Takt time expresses the rhythm of customer demand. Pull determines when production or replenishment should occur. JIT is the broader operating philosophy of making and delivering what is needed, when it is needed, and in the required quantity. Together, these concepts provide the operational foundation for cellular manufacturing, line balancing, Kanban and heijunka.

The managerial challenge is not simply to accelerate production. A fast process that produces excess inventory, defects or unstable output is not lean. The objective is controlled, visible and reliable flow, supported by appropriate buffers and the capability to respond to variation.

Learning Objectives

After studying this chapter, students should be able to:

1. Explain the relationship among flow, takt time, pull and JIT.

2. Calculate takt time and use it in line-balancing decisions.

3. Distinguish continuous flow, cellular flow, batch flow, FIFO lanes and supermarkets.

4. Design basic Kanban and replenishment systems.

5. Explain heijunka, lot-size reduction and SMED as responses to variability.

6. Evaluate the managerial risks of applying pull systems without process stability.

7. Recommend flow improvements in manufacturing and service operations.

8. Interpret digital and analytical technologies that support flow management.

From Functional Batches to Connected Flow

Traditional operations are frequently organised by function. Similar machines or skills are grouped together, and work is produced in batches before being transferred to the next department. This arrangement may provide high local utilisation, but it often creates queues, long lead times, excess work-in-process and delayed detection of defects.

Flow reorganises the process around the sequence of value creation. Ideally, one product or service moves from one step to the next without waiting. The aim is not necessarily physical movement at all times; some processes require controlled waiting or inspection. Rather, flow means that interruptions, queues and unnecessary handoffs are systematically reduced.

Three forms are particularly important:

• Continuous flow: Items move one at a time, or in very small quantities, from process to process.
• Cellular flow: Equipment and people are arranged in a product or service family sequence, often in a U-shaped or compact cell.
• Batch flow: Items are processed in groups. Batch flow may be necessary where changeovers, transport, technical constraints or demand patterns make one-piece flow impractical.

One-piece flow is an ideal rather than a universal rule. A technically sound lean design may use small batches, FIFO lanes or supermarkets where complete continuity is impossible. The important question is whether the chosen arrangement exposes problems and limits inventory rather than concealing instability.

Flow and the Seven Flows

A mature flow design considers more than physical materials. It also considers the movement of information, people, equipment, quality decisions, cash and customer requests. When information moves slowly or is distorted, material flow becomes unstable. For example, an inaccurate production schedule may cause a machine to produce items that are not currently required, while a downstream shortage remains unresolved.

Flow therefore requires visibility. Standard work, visual controls, status boards, and clearly defined replenishment signals allow abnormal conditions to be detected quickly. A process that appears smooth only because inventory hides its problems is not genuinely controlled.

Takt Time: The Customer-Determined Rhythm

Takt time is the available production time divided by customer demand during the same period:

Takt time = Net available operating time / Customer demand

If a shift provides 420 net working minutes and customers require 210 units, takt time is two minutes per unit. This does not mean that every operation must literally complete its work in exactly two minutes. It means that the complete system must be capable of delivering one finished unit every two minutes on average.

Net available time excludes planned breaks, meetings and other periods during which production is not expected. The time basis must be consistent with the demand basis. If demand varies by product, shift or season, takt should be calculated for the relevant planning interval rather than using an annual average that hides actual requirements.

Takt time is distinct from:

• Cycle time: The actual time required to complete one cycle at a process.
• Lead time: The elapsed time from order initiation to delivery.
• Throughput time: The time a unit spends being processed, including relevant waiting.
• Pitch: A practical, often larger, interval used to move a defined quantity and communicate production status.

If a workstation’s cycle time exceeds takt time, it is a bottleneck under the stated conditions. If cycle time is much shorter than takt, the station may have unused capacity, though the apparent capacity may be misleading if it depends on overtime, excessive work-in-process or unstable quality.

Takt Time and Capacity Decisions

Takt is a demand signal, not a guarantee of demand. Managers must compare required capacity with available capacity, considering uptime, changeovers, absenteeism, maintenance, product mix and quality losses. A line designed precisely to average takt may fail when demand or process conditions fluctuate. Practical design therefore includes capacity protection, cross-training and carefully chosen buffers.

For example, if a process has a 90-second cycle time and takt is 120 seconds, the process has apparent spare capacity. If its first-pass yield is poor or changeovers consume significant time, however, the effective capacity may be inadequate. Capacity analysis must therefore use effective rather than purely theoretical time.

Line Balancing and Cellular Design

Line balancing assigns tasks among workstations so that work is distributed as evenly as possible without exceeding takt. The theoretical minimum number of stations is:

Minimum stations = Total work content / Takt time

The result is rounded upward. Suppose total work content is 18 minutes and takt time is 2.5 minutes. The theoretical minimum is 7.2, so at least eight stations are required. Actual requirements may be higher because of precedence restrictions, equipment differences, ergonomic considerations and skill availability.

A useful measure of balance is line efficiency:

Line efficiency = Total work content / (Number of stations × Takt time) × 100

Balance delay is the unused capacity resulting from uneven task allocation. High efficiency is desirable, but it should not be pursued by forcing workers to operate at unsafe speeds or by eliminating all flexibility. A balanced line must also protect quality and accommodate normal variation.

Cellular design groups equipment, methods and people around a family of products or services. A cell reduces transport and waiting, encourages multi-skilled work and makes abnormal conditions visible. Product-family analysis is essential: products should not be placed together merely because they pass through similar departments. Their process routes, demand patterns, quality requirements and changeover needs must be sufficiently compatible.

Pull, Kanban and Controlled Replenishment

A push system releases work according to a schedule or forecast, regardless of whether the next process is ready. A pull system authorises upstream production in response to actual downstream consumption. Pull does not mean that every activity waits passively; it means that replenishment is triggered by use and constrained by agreed limits.

Kanban is a signalling mechanism used to operate pull. A Kanban may be a card, container, electronic signal or clearly defined digital transaction. It normally identifies the part, quantity, source, destination and replenishment rule. When a downstream process consumes a quantity, the associated signal returns upstream to authorise replenishment.

A simple Kanban quantity can be estimated as:

Number of Kanban containers = (Demand during replenishment lead time × (1 + safety factor)) / Container quantity

The result is rounded upward. The safety factor should not become a permanent substitute for problem solving. Excess Kanban increases inventory and can conceal defects, unreliable suppliers or inaccurate lead times.

Supermarkets and FIFO Lanes

A supermarket is a controlled inventory point from which a downstream customer withdraws items. The upstream process replenishes what has been consumed. It is appropriate when processes cannot be directly connected because of distance, different operating hours, long changeovers or differing production rhythms.

A FIFO lane is a first-in, first-out queue with a defined maximum quantity. It is useful when direct continuous flow is not possible but production should remain sequenced. When the lane is full, the upstream process stops or investigates the abnormal condition. This limit is crucial: an unlimited queue is simply hidden inventory.

Supermarkets provide decoupling, whereas FIFO lanes preserve sequence. Both should be deliberately sized and regularly reviewed. Their purpose is not to maximise stock availability but to protect flow while exposing variation.

Heijunka and the Management of Demand Variation

Customer demand is rarely level. If production follows every demand fluctuation directly, the operation may experience uneven workloads, overtime, idle time and unstable supplier requirements. Heijunka, or production levelling, smooths the volume and mix of production over a planning horizon.

Volume levelling distributes the quantity of work over time. Mix levelling distributes different products in a repeating pattern, such as A-B-A-C rather than producing a large batch of A followed by large batches of B and C. The appropriate pattern depends on demand, changeover time, product families and available capacity.

Heijunka does not eliminate genuine demand variation. It separates short-term production stability from the customer’s immediate order pattern by using carefully controlled finished-goods or component buffers. It is therefore a managerial compromise: some inventory may be accepted to prevent greater instability, but the buffer must be visible, bounded and governed.

Lot-Size Reduction and SMED

Large lots are often justified by the desire to reduce setup frequency. However, large lots create inventory, delay defect detection and reduce responsiveness. Reducing lot size requires reducing the time and risk associated with changeovers.

Single-Minute Exchange of Die (SMED) is a structured approach to reducing setup time. Its logic includes:

1. Observe and document the existing changeover.

2. Separate internal activities, which require the equipment to be stopped, from external activities, which can occur while it is running.

3. Convert internal activities into external activities wherever possible.

4. Simplify, standardise and improve both categories.

5. Eliminate adjustment through locating devices, presets, quick fasteners and visual standards.

6. Sustain the new method through training and measurement.

SMED is not limited to presses or dies. It applies to cleaning, software deployment, healthcare room preparation, restaurant menu changes and administrative process transitions. The strategic outcome is not merely a lower setup metric; it is the ability to produce smaller lots economically and respond more closely to demand.

A Practical Design Sequence

A flow-and-pull redesign should begin with demand and process facts rather than with a preferred tool. A practical sequence is as follows.

First, define the customer, product families, demand period and net available time. Second, observe the current process directly and record cycle times, queues, changeovers, defects, uptime and replenishment delays. Third, calculate takt and compare it with each process’s effective cycle time. Fourth, identify constraints and redesign the physical and informational sequence into a cell or connected flow where feasible.

Next, establish standard work and quality-at-the-source practices. Only then should the organisation define supermarkets, FIFO lanes or Kanban limits for points that cannot be directly connected. Level the production schedule where demand and mix variation would otherwise destabilise the system. Reduce changeover time and progressively reduce lot sizes. Finally, conduct frequent reviews using actual consumption, service performance, lead time, inventory and abnormality data.

This sequence matters. Installing Kanban before stabilising quality often causes repeated shortages and emergency expediting. Balancing a line before clarifying standard work may simply distribute inconsistency among stations.

Illustrative Case: Flow Redesign in a Medical-Device Assembly Cell

Background

The following is an illustrative case. MedSure Components assembles a family of diagnostic kits in a regulated facility. Customer demand averages 240 kits per day. The operation has 480 scheduled minutes per shift, including planned breaks and meetings, leaving 420 net minutes. The current arrangement uses separate departments for component preparation, subassembly, testing and packing.

Situation and Problem

Although individual departments report high utilisation, total lead time is three days. Finished kits frequently wait for testing, while subassemblies accumulate between preparation and assembly. A quality defect discovered during final inspection may require the review of several hundred units. Changeovers between three kit variants take 75 minutes, encouraging weekly batch production.

Takt time is:

420 minutes / 240 kits = 1.75 minutes per kit.

Observed effective cycle times are 1.2 minutes for preparation, 2.1 minutes for assembly, 1.4 minutes for testing and 1.0 minute for packing. Assembly is therefore above takt and is the principal capacity constraint. The operation also experiences frequent short stoppages and variation in operator skill.

Analysis

A cross-functional team maps the material and information flow. It finds that the departmental layout creates transport and queueing, while the weekly schedule pushes large batches without reference to actual downstream consumption. The team also discovers that testing equipment is shared with another product family and that changeover adjustments are not standardised.

The apparent problem is insufficient labour at assembly, but deeper causes include poor task allocation, long changeovers, weak quality-at-source controls and an inappropriate release system. Adding inventory would protect delivery temporarily but would increase the number of units exposed to defects.

Approach and Solution

The team designs a U-shaped cell for the product family. Work content is redistributed so that two operators can perform complementary tasks, while a trained floater supports abnormal conditions. A simple line balance is established around the 1.75-minute takt, with quality checks placed immediately after the operations most likely to generate defects.

A small supermarket is retained for a component that must be prepared in a separate controlled room. The assembly cell withdraws components using electronic Kanban signals; the preparation room replenishes only the quantities consumed. A FIFO lane is used between testing and packing because testing equipment operates in a different rhythm. The lane is limited to one hour of demand.

The team applies SMED to the three kit variants. Materials and tools are prepared externally, fixtures are colour-coded, and adjustment points are replaced with preset locations. The changeover method is documented and practised. The weekly batch is replaced by a levelled sequence that repeats the product mix several times per shift, subject to regulatory and demand constraints.

Outcome

The redesign reduces work-in-process and makes shortages visible at their point of origin. Quality issues are detected earlier because checks occur within the cell. Operators can see whether the cell is ahead or behind its pitch, and supervisors can distinguish a genuine capacity issue from a material or equipment problem. The organisation does not treat these improvements as permanent guarantees; it continues to review takt, product mix, equipment availability and regulatory requirements.

Learning Points

The case illustrates five principles. First, high departmental utilisation does not prove good system flow. Second, takt identifies the required rhythm but does not solve capacity constraints by itself. Third, pull works through defined limits and signals, not through an assumption that inventory is unnecessary. Fourth, SMED enables smaller lots and better levelling. Finally, visual control and standard work are prerequisites for sustained improvement.

Risks, Limitations and Managerial Judgement

Flow systems are vulnerable to variability. Demand changes, supplier delays, machine failures, absenteeism and quality problems can interrupt a tightly coupled process. A system with no buffers may be efficient under ideal conditions but fragile in practice. Managers must decide where buffers are economically and operationally justified, and must ensure that buffers are governed rather than allowed to grow without limit.

Common mistakes include using average demand to calculate takt when demand is highly seasonal; treating takt as a target for individual workers rather than a system rhythm; balancing work without considering precedence or ergonomics; implementing Kanban with inaccurate container quantities; using supermarkets as permanent storage; and reducing inventory before improving process capability.

There are also human risks. Excessive pressure to meet takt can encourage unsafe work, skipped inspections or concealment of problems. Lean flow should not be interpreted as a demand for continuous acceleration. Work design must comply with occupational health and safety requirements, include reasonable recovery time and respect employee knowledge. Operators should participate in designing standard work because they understand practical sources of variation.

In regulated industries, traceability, validation, segregation and documentation may restrict the use of certain flow arrangements. A faster flow is unacceptable if it weakens product safety or regulatory control. Governance should therefore define who may alter Kanban quantities, release rules, standard work or inspection points.

Digital Support for Flow Management

Digital technologies can improve visibility, but they do not substitute for process understanding. Manufacturing execution systems, barcode scanning, RFID, electronic Kanban and real-time dashboards can record consumption, inventory location, cycle time and abnormalities. Statistical process monitoring can identify drift before it becomes a major interruption. Simulation and digital-twin tools can test alternative cell layouts, staffing patterns and buffer sizes.

Artificial intelligence may support demand sensing, anomaly detection and predictive maintenance. Its value depends on data quality, transparent decision rules and appropriate human oversight. An algorithm that responds to every short-term demand fluctuation may amplify instability rather than support heijunka. Digital systems should therefore preserve clear replenishment rules, audit trails and escalation responsibilities.

Current Developments and Best Practice

Contemporary flow systems increasingly combine lean principles with flexible automation, modular equipment and postponement. Collaborative robots may assist with repetitive handling while people perform judgement-intensive tasks. Cloud-based supply-chain platforms improve coordination across suppliers, although they also create cybersecurity and data-governance obligations.

Best practice is characterised by several habits: calculate takt from clearly defined demand and available time; observe the process at the workplace; balance effective capacity rather than theoretical capacity; use small, visible and bounded buffers; stabilise quality before reducing inventory; reduce changeover time to support smaller lots; review Kanban limits as conditions change; and treat abnormalities as opportunities for learning rather than occasions for blame.

Chapter Summary

Flow connects operations so that value moves with minimal waiting, transport and interruption. Takt time translates customer demand into an operating rhythm, while cycle time, capacity and quality determine whether that rhythm is achievable. Line balancing and cellular design distribute work around takt and expose constraints.

Pull systems replenish work in response to actual consumption. Kanban provides the signal, while supermarkets and FIFO lanes provide controlled decoupling where direct flow is not feasible. Heijunka reduces the disruptive effects of volume and mix variation. Lot-size reduction becomes practical when SMED lowers changeover time. These methods must be applied in sequence and adapted to process stability, safety, regulation and demand uncertainty.

The central managerial lesson is that lean flow is a system design problem, not an isolated inventory-reduction programme. Sustainable results require standard work, quality at the source, employee involvement, visible controls, bounded buffers and continuous learning.

Key Terms

Flow: The progressive movement of work through value-creating steps with minimal interruption.

Continuous flow: Movement of items through successive processes with little or no waiting.

Cellular flow: Arrangement of resources around a product or service family and its process sequence.

Takt time: Net available operating time divided by customer demand.

Cycle time: Time required by a process to complete one work cycle.

Lead time: Total elapsed time from initiation of an order to delivery.

Line balancing: Allocation of tasks among workstations to meet takt with minimal imbalance.

Pull system: A system in which replenishment is authorised by downstream consumption.

Kanban: A signal that authorises production, movement or replenishment.

Supermarket: A controlled inventory point replenished according to downstream withdrawal.

FIFO lane: A bounded queue that preserves first-in, first-out sequence.

Heijunka: Production levelling by volume and product mix.

Just-in-time: Producing and delivering the required item, quantity and timing with minimal unnecessary inventory.

Lot size: Quantity of units processed between changeovers.

Changeover: Activities required to switch a process from one product or condition to another.

SMED: A method for reducing setup time by separating and simplifying internal and external activities.

Pitch: A defined time interval and quantity used to communicate production progress.

Work-in-process: Items currently being processed or waiting between process steps.

Effective capacity: Practical output capability after considering downtime, quality losses, changeovers and other constraints.

Review Questions

Short Answer Questions

1. Define takt time and distinguish it from cycle time.

2. What is the purpose of a FIFO lane?

3. Explain how a Kanban signal supports pull production.

4. What is meant by heijunka?

5. Identify two activities involved in SMED.

Long Answer Questions

1. Explain how cellular design, line balancing and standard work contribute to continuous flow.

2. Discuss the relationship among supermarkets, FIFO lanes and Kanban in a pull system.

3. Explain why large lot sizes arise and how SMED enables lot-size reduction.

4. Analyse the benefits and limitations of using takt time for capacity planning.

5. Discuss the organisational, safety and governance risks of implementing JIT without process stability.

Analytical / Application Questions

1. A plant has 360 net operating minutes and daily demand of 180 units. Calculate takt time. If a workstation has a cycle time of 2.4 minutes, assess its capacity against takt and recommend possible responses.

2. Design a basic pull arrangement for two processes that cannot be directly connected because one has a long changeover and the other operates continuously. Explain where a supermarket or FIFO lane would be used and why.

3. A service centre experiences highly variable arrivals, frequent overtime and long queues despite high staff utilisation. Apply flow, takt, levelling and capacity principles to propose a redesign and identify the information required before implementation.

22. Multiple Choice Questions with Answers

Question 1

A production line has 420 net operating minutes and customer demand of 210 units per shift. What is the takt time?

A. 0.5 minute per unit

B. 2 minutes per unit

C. 210 minutes per unit

D. 630 minutes per unit

Correct Answer: B. 2 minutes per unit

Question 2

Which statement best distinguishes a pull system from a push system?

A. Pull always eliminates inventory.

B. Pull releases work in response to downstream consumption.

C. Pull requires every process to operate at maximum utilisation.

D. Pull is based exclusively on long-range forecasts.

Correct Answer: B. Pull releases work in response to downstream consumption.

Question 3

A FIFO lane is most appropriate when an organisation needs to:

A. Create an unlimited queue between departments

B. Preserve sequence with a bounded amount of work-in-process

C. Replace all quality inspections

D. Produce large batches without signals

Correct Answer: B. Preserve sequence with a bounded amount of work-in-process

Question 4

If total work content is 20 minutes and takt time is 2.5 minutes, what is the theoretical minimum number of stations?

A. 5

B. 7

C. 8

D. 10

Correct Answer: C. 8

Question 5

What is the principal purpose of a supermarket in a lean flow system?

A. To maximise finished-goods inventory

B. To provide controlled decoupling and replenish what downstream consumes

C. To remove the need for demand information

D. To ensure upstream processes never stop

Correct Answer: B. To provide controlled decoupling and replenish what downstream consumes

Question 6

Which action is most consistent with SMED?

A. Increase batch size to reduce setup frequency

B. Perform all preparation after the machine is stopped

C. Convert internal setup activities into external activities where possible

D. Ignore adjustment time because it is unavoidable

Correct Answer: C. Convert internal setup activities into external activities where possible

Question 7

Heijunka primarily addresses which operational problem?

A. Unlevelled volume and product-mix demand

B. The definition of customer value

C. The elimination of all preventive maintenance

D. The calculation of theoretical minimum stations

Correct Answer: A. Unlevelled volume and product-mix demand

Question 8

A workstation has a cycle time greater than takt time. What does this generally indicate?

A. The station has excess capacity

B. The station may be a capacity constraint under the stated conditions

C. Demand is necessarily too low

D. The process should immediately receive unlimited inventory

Correct Answer: B. The station may be a capacity constraint under the stated conditions

Question 9

Why can implementing Kanban before stabilising quality create problems?

A. Kanban prevents visual management

B. Kanban increases the number of defects by definition

C. Defects and unreliable processes can cause repeated shortages and expediting

D. Kanban requires every product to be made in one piece flow

Correct Answer: C. Defects and unreliable processes can cause repeated shortages and expediting

Question 10

Which practice best reflects responsible lean management?

A. Set takt as an individual worker speed requirement regardless of safety

B. Remove all buffers immediately to expose employees to pressure

C. Use bounded buffers, employee involvement and quality-at-source controls

D. Prioritise utilisation even when it creates excess inventory

Correct Answer: C. Use bounded buffers, employee involvement and quality-at-source controls