Lean, JIT, Pull, and Hybrid Planning Systems

Introduction and Context

Inventory control is not simply the problem of deciding how much stock to hold. It is the design of a material-flow system that coordinates demand, production, replenishment, information, capacity, and risk. Traditional materials requirements planning (MRP) begins with a schedule and explodes it into time-phased requirements. Lean and just-in-time (JIT) systems begin with actual consumption and seek to replenish only what has been used, when it is needed, in the required quantity and quality.

Neither approach is universally superior. Lean control is highly effective where demand is relatively stable, processes are repeatable, replenishment lead times are short, and problems can be exposed and rapidly corrected. MRP is more suitable where demand is dependent, product structures are complex, lead times are long, order quantities are large, or production and procurement must be coordinated across multiple levels. Contemporary organizations therefore commonly use hybrid systems: MRP plans, while pull mechanisms execute and regulate flow.

This chapter evaluates the principal mechanisms of lean material control—kanban, supermarkets, two-bin systems, heijunka, takt, small-lot replenishment, vendor-managed inventory (VMI), and consignment—and establishes criteria for selecting lean, MRP, or hybrid planning.

Learning Objectives

By the end of this chapter, the reader should be able to:

1. Explain the logic of push, pull, JIT, and lean control.
2. Design and evaluate kanban, supermarket, and two-bin systems.
3. Relate takt time, cycle time, capacity, heijunka, and lot size.
4. Distinguish VMI from consignment and assess their governance implications.
5. Compare lean control with MRP under different demand and supply conditions.
6. Design a hybrid planning architecture and identify its risks.
7. Apply quantitative and managerial criteria to an inventory-control decision.

1. Push, Pull, and the Logic of Lean Control

A push system releases material based primarily on a plan, forecast, or calculated requirement. MRP is the classic example: the master production schedule, bill of materials, inventory records, and lead times generate planned orders. A pull system authorizes replenishment in response to downstream consumption or an explicit signal from the using process.

The distinction is not identical to the distinction between forecast and actual demand. Every system uses forecasts to some extent. A pull system may use a forecast to size capacity and inventory, but the immediate replenishment signal is consumption. Conversely, a push system may use actual orders while still releasing work centrally in batches.

Lean management seeks to create continuous flow and eliminate activities that do not add value. The commonly identified forms of waste include overproduction, waiting, transportation, overprocessing, inventory, motion, defects, and underused human capability. Inventory is not automatically waste: it may protect customers from uncertainty or decouple processes. However, excessive inventory can conceal defects, unstable schedules, long setups, unreliable suppliers, and capacity constraints.

JIT is a timing and coordination philosophy: the right item should arrive at the right process, in the right quantity, at the right time. It does not mean zero inventory under all conditions. Rather, it requires deliberate reduction of unnecessary buffers while improving process capability and response speed.

2. Takt, Cycle Time, and Flow Design

Takt time expresses the pace of customer demand. For a planning interval, it is calculated as available productive time divided by required customer output:

If a shift provides 420 minutes of scheduled time, of which 30 minutes are planned breaks, and customer demand is 390 units, then:

The assumptions are that the demand and time interval are expressed in compatible units and that the available time excludes planned non-production periods. Takt is a demand pace, not a guarantee of actual performance.

Cycle time is the elapsed time between completed units from a process. To sustain demand without accumulating a queue, the effective cycle time of the relevant process should be no greater than takt time, after considering losses and variability. If cycle time exceeds takt, output capacity is insufficient. If it is substantially less than takt, the process may have excess capacity, although this can be useful for recovery, flexibility, or mixed-model production.

For a line with total manual work content W and takt time T, the theoretical minimum number of workstations is:

Because stations must normally be whole resources, the practical minimum is the ceiling of this value. The equation assumes that work can be divided and assigned without precedence, ergonomic, skill, or equipment constraints. In practice, line balancing may require more stations.

Heijunka, or production leveling, smooths the volume and mix of production over time. Without leveling, a customer order for 500 units today and none tomorrow can create alternating overload and idleness. Heijunka does not eliminate demand variability; it converts large fluctuations into a more manageable sequence, often through mixed-model production and a repeating schedule.

For example, if weekly demand is 600 units—300 of Model A and 300 of Model B—and there are five production days, a leveled schedule may produce 120 units per day, with approximately equal quantities of A and B each day. The exact sequence depends on setup times, product constraints, and customer priorities.

3. Kanban as a Control Mechanism

A kanban is an authorization signal. It may be a card, container, electronic message, barcode event, or machine-readable signal. The signal authorizes a supplying process to replenish a specified quantity for a specified point of use. Kanban therefore controls both material movement and production authorization.

A basic two-card system uses:

• a withdrawal kanban, authorizing movement from a supermarket to the customer process; and
• a production kanban, authorizing the supplying process to replace what has been withdrawn.

A one-card system may combine these functions where the material route and replenishment relationship are simple. The fundamental rule is that no kanban should be detached from its material without a corresponding transaction or physical movement.

Kanban Quantity and Number of Cards

A commonly used sizing relationship is:

where:

• n = required number of kanban containers or cards;
• D = average demand rate in units per time;
• L = replenishment lead time in the same time unit;
• S = safety factor expressed as a decimal;
• C = standard quantity per container.

The result is normally rounded upward to a whole number. The formula assumes relatively stable demand, a defined replenishment lead time, consistent container quantity, and a safety factor that represents uncertainty. It is a policy-sizing approximation, not a substitute for capacity analysis.

Suppose demand is 240 units per day, replenishment lead time is 0.5 day, the safety factor is 20%, and each container holds 30 units:

Therefore, at least five containers or cards are required. The corresponding nominal circulating quantity is 150 units. If the system cannot replenish within the assumed lead time, five cards will be inadequate regardless of their administrative accuracy.

Kanban quantity should be reviewed when demand, lead time, container size, quality yield, product mix, or process reliability changes. Reducing cards is not itself lean if it merely causes stockouts and expediting.

4. Supermarkets and Two-Bin Systems

A supermarket is a controlled inventory location positioned between processes or between a supplier and a using operation. It holds defined quantities and locations for specified items. The downstream process withdraws material; the upstream process replenishes what was withdrawn. A supermarket is therefore a physical decoupling point governed by pull signals.

Supermarkets are particularly useful when:

• adjacent processes cannot be directly connected by continuous flow;
• replenishment lead time is longer than the customer process can tolerate;
• products have different processing routes; or
• a supplier delivers on a milk-run schedule.

They are less appropriate when items are extremely expensive, perishable, highly customized, or so unpredictable that fixed locations and quantities become obsolete.

The two-bin system is a simple point-of-use replenishment method. The first bin supplies current use. When it becomes empty, it is placed in a collection location or its signal is scanned, and the second bin supplies consumption during replenishment. The empty-bin signal is the reorder trigger.

If average demand during replenishment is 80 units and the desired protection quantity is 20 units, a two-bin arrangement might use two bins of 50 units each, providing 100 units total. The example is not a universal formula: bin size must reflect demand variability, lead time, packaging constraints, and acceptable stockout risk. Excessively large bins create hidden inventory; undersized bins create frequent shortages and transaction cost.

5. Small-Lot Replenishment and Setup Reduction

Small-lot replenishment reduces average cycle stock, shortens response time, and exposes quality problems sooner. For a constant demand rate, average cycle inventory under a replenishment lot of size Q is approximately:

This special-case relationship assumes steady demand, replenishment that is effectively instantaneous relative to consumption, and no safety stock. If replenishment is gradual, lead-time stock and safety stock must be considered separately.

Small lots are economically attractive only when ordering, setup, transportation, receiving, and transaction costs are controlled. A traditional economic order quantity model is:

where D is annual demand in units, S is ordering or setup cost per order, and H is annual holding cost per unit. The classical model assumes constant demand, constant lead time, no quantity discounts, instantaneous replenishment, and no stockouts. Lean environments often challenge these assumptions by reducing setup cost and using frequent deliveries, making smaller lots practical even when the classical model would recommend a larger quantity.

Setup reduction methods include externalizing preparation, using quick-release fixtures, standardizing tooling, separating internal and external activities, and applying single-minute exchange of die (SMED) principles. However, small lots can increase transport emissions, handling, schedule instability, and administrative workload. The correct objective is total system performance, not the smallest possible lot.

6. Vendor-Managed Inventory and Consignment

Under vendor-managed inventory, the supplier is responsible for monitoring the customer’s inventory and deciding when and how much to replenish, subject to agreed service, ownership, and operating rules. The customer supplies accurate consumption and inventory information. VMI can reduce the customer’s planning workload and improve supplier visibility, but it transfers decision authority rather than eliminating the need for control.

Consignment inventory is an ownership arrangement. The supplier retains ownership until the customer consumes, sells, or otherwise uses the material. VMI and consignment may be combined, but they are conceptually different:

A sound VMI agreement specifies data ownership, forecast visibility, reorder rules, service levels, minimum and maximum quantities, lead times, quality responsibility, obsolete stock, emergency supply, audit rights, cybersecurity, and performance measures. Consignment agreements additionally specify title transfer, shrinkage, insurance, valuation, tax treatment, obsolescence, and physical reconciliation.

Ethically and commercially, VMI should not be used to conceal excess production or shift inventory risk unfairly to a financially weaker partner. Shared data must be limited to legitimate purposes and protected against unauthorized access.

7. MRP, Lean Control, and the Hybrid Architecture

MRP is a planning logic for dependent demand. It uses the master production schedule, bill of materials, inventory records, scheduled receipts, lead times, and lot-sizing rules to calculate net requirements and planned order releases. For an item in period t, a simplified net-requirement calculation is:

The exact available balance depends on beginning inventory, prior planned receipts, allocations, safety stock, and the system’s netting conventions. MRP is powerful because it coordinates components across multiple product levels and time periods. Its weaknesses include nervousness from schedule changes, dependence on data accuracy, sensitivity to lead-time assumptions, and a tendency to create planned orders that are not operationally feasible unless capacity is checked.

Lean pull and MRP address different control problems. Pull is strong at authorizing near-term execution in repetitive flow. MRP is strong at translating end-item plans into requirements for dependent components, purchased materials, and capacity-sensitive activities.

A hybrid architecture commonly follows this pattern:

1. Sales and operations planning establishes aggregate volume and capacity decisions.
2. The master production schedule translates the plan into end-item timing and mix.
3. MRP plans long-lead, dependent, engineered, or externally purchased materials.
4. A frozen or time-fenced horizon protects near-term execution.
5. Kanban, supermarkets, or electronic pull signals control repetitive replenishment inside the execution horizon.
6. Exception management addresses shortages, quality problems, capacity failures, and demand changes.

The boundary should be explicit. If MRP releases orders into a supermarket without regard to actual consumption, the system becomes a disguised push system. If kanban is used for unique, long-lead components without adequate planning, the system may experience severe shortages.

Selection Criteria

8. Performance Measurement and Control Discipline

Lean systems require more than cards and labels. Relevant measures include:

• service level and stockout frequency;
• inventory turnover and days of supply;
• replenishment lead-time adherence;
• schedule attainment;
• first-pass yield and defect escape rate;
• changeover time;
• inventory record accuracy;
• kanban signal discipline;
• supplier delivery reliability;
• emergency freight and expediting;
• obsolescence and write-offs.

Inventory turnover is commonly calculated as:

The numerator and denominator must use compatible valuation bases. High turnover is not automatically good if it produces lost sales, overtime, or quality failures. Similarly, a high service level achieved through excessive inventory may be economically unacceptable.

A useful governance principle is to treat shortages as signals for root-cause analysis rather than immediately adding stock. Causes may include inaccurate bills of material, incorrect container quantities, supplier unreliability, unrecorded scrap, unauthorized withdrawals, poor maintenance, or an unrealistic takt assumption.

9. Illustrative Case: Hybrid Control at Meridian Appliances

Background

Meridian Appliances manufactures refrigerators in three models. Final assembly operates five days per week. Motors and electronic controllers are imported with six-week lead times. Fasteners, brackets, and standard packaging are sourced locally. A subassembly cell produces wiring harnesses and supplies final assembly.

Problem

Meridian used MRP for every item. Planners frequently changed orders in response to minor schedule variations. The result was excess fastener inventory, shortages of wiring harnesses, frequent expediting of imported controllers, and poor trust in system dates. Inventory records were accurate for high-value items but unreliable for point-of-use consumables.

Analysis

The imported controllers were dependent-demand items with long and variable replenishment lead times. A pure pull system could not respond quickly enough. Conversely, fasteners were repetitive, low-value items consumed at a stable rate and delivered by a local supplier. They were appropriate for a two-bin system. Wiring harnesses had a stable route and short internal replenishment time but were produced in batches that did not match final assembly consumption.

Approach

Meridian adopted four control segments:

1. MRP and supplier capacity reviews for imported controllers, with a time-fenced horizon and safety stock based on lead-time variability.
2. A supermarket between harness production and final assembly, replenished by electronic kanban.
3. Two-bin replenishment for standard fasteners and packaging.
4. Heijunka sequencing at final assembly, producing a repeating model pattern to reduce disruptive changeovers.

The company also reduced harness changeover time, standardized container quantities, and introduced daily reconciliation of kanban transactions. VMI was negotiated with the local fastener supplier, while ownership remained with Meridian; therefore, the arrangement was VMI but not consignment.

Outcome

After stabilization, shortages of harnesses declined because replenishment was tied to actual consumption. Fastener planner workload fell, and point-of-use stock became visible. Imported-controller risk remained and was managed through MRP rather than incorrectly forced into a pull loop. The main improvement was not a single tool but a clear division of control responsibility.

Learning Points

The case demonstrates that item characteristics, not ideological preference, should determine the control method. A hybrid system can be lean when it reduces unnecessary release, limits inventory, exposes problems, and maintains reliable flow. It becomes non-lean when multiple overlapping signals create contradictory authorizations or when buffers are increased without addressing root causes.

10. Implementation Risks and Contemporary Developments

Implementation commonly fails for five reasons. First, organizations deploy kanban before stabilizing processes. Second, they reduce inventory without improving quality, maintenance, or supplier reliability. Third, they maintain inaccurate master data and expect electronic signals to correct it. Fourth, they optimize one department’s inventory while increasing total supply-chain cost. Fifth, they treat pull as a software feature rather than a governed operating system.

Digital technologies can strengthen lean control. Barcode and RFID events improve transaction accuracy; internet-of-things sensors can report consumption and machine status; analytics can identify abnormal replenishment patterns; and AI can support demand sensing, anomaly detection, supplier-risk assessment, and dynamic parameter review. However, AI-generated recommendations require traceable data, human accountability, access controls, model monitoring, and protection against biased or strategically manipulated data. Automation should not silently change kanban quantities or safety stocks without governance.

Contemporary supply chains also face resilience, sustainability, geopolitical, and regulatory pressures. A very lean global network may be vulnerable to disruption. Firms should distinguish between wasteful inventory and strategically justified resilience stock. Sustainability analysis should include transport frequency, packaging, energy, emissions, reverse logistics, and the environmental cost of obsolescence. The best design may use local pull for common items, planned strategic stock for constrained items, and postponement for high-variety products.

Best-Practice Design Principles

1. Define the material-flow boundary and the purpose of every signal.
2. Segment items by demand pattern, value, lead time, product structure, and risk.
3. Stabilize quality, equipment, work standards, and data before reducing buffers.
4. Size kanban using demand, replenishment time, container quantity, and explicit uncertainty assumptions.
5. Use heijunka where demand mix and setup economics permit leveling.
6. Keep long-lead and complex dependent-demand items under appropriate MRP control.
7. Establish one authoritative signal for each replenishment decision.
8. Review parameters through actual performance, not arbitrary reduction targets.
9. Measure total cost and service, not inventory alone.
10. Govern VMI and consignment contracts with clear ownership, data, liability, and exit provisions.
11. Treat shortages and excesses as diagnostic evidence.
12. Use technology to improve visibility and decisions, while preserving human accountability.

Chapter Summary

Lean, JIT, and pull systems regulate replenishment through consumption and seek reliable flow with minimal unnecessary inventory. Kanban authorizes movement or production; supermarkets provide controlled decoupling points; two-bin systems offer simple point-of-use control. Takt expresses the pace required by customer demand, while cycle time describes process output pace. Heijunka levels volume and mix, and small-lot replenishment becomes practical when setup, handling, and transaction costs are controlled.

VMI concerns replenishment decision responsibility, whereas consignment concerns inventory ownership. MRP remains essential for complex dependent demand, long lead times, engineering-controlled products, and coordinated multi-level requirements. Hybrid systems combine MRP planning with pull execution, provided their boundaries and signals are clearly governed. The appropriate choice depends on demand stability, product structure, lead time, process reliability, capacity, supplier risk, and the cost of failure. Lean control is not the unconditional removal of inventory; it is the disciplined design of flow, information, buffers, and problem-solving.

Key Terms

Andon: A visual or electronic signal indicating process status or an abnormal condition.

Consignment inventory: Inventory owned by the supplier until an agreed consumption or transfer event.

Cycle time: Time between completed units from a process.

Heijunka: Production leveling by volume and product mix.

JIT: A philosophy of supplying materials when needed, in the needed quantity and quality.

Kanban: A signal authorizing material withdrawal, movement, or production.

Lead time: Elapsed time between initiating and completing a replenishment or process.

MRP: Materials requirements planning for time-phased dependent demand.

Pull system: A system in which downstream consumption authorizes replenishment.

Push system: A system in which planned or forecast requirements authorize release.

Safety factor: An allowance for demand, lead-time, quality, or process uncertainty.

SMED: Single-minute exchange of die; methods for rapidly reducing changeover time.

Supermarket: A controlled inventory location that decouples processes and supports pull replenishment.

Takt time: Available productive time divided by required customer output.

Two-bin system: A replenishment method in which an empty bin signals replenishment while the second bin supplies use.

VMI: Vendor-managed inventory, in which the supplier manages replenishment using customer inventory information.

Work content: Total labor or processing time required for a unit or task set.

Review Questions

Short Answer Questions

1. Distinguish a push system from a pull system.
2. What is the operational purpose of a supermarket?
3. Define takt time and explain how it differs from cycle time.
4. Distinguish VMI from consignment inventory.
5. State two conditions under which MRP is preferable to a simple kanban loop.

Long Answer Questions

1. Explain how kanban, supermarkets, and two-bin systems work together in a lean material-flow design.
2. Evaluate the benefits and limitations of small-lot replenishment.
3. Discuss how heijunka supports flow and what organizational conditions are needed for its success.
4. Compare MRP and pull control in terms of information requirements, demand characteristics, and risk.
5. Design a governance framework for a VMI and consignment agreement.

Analytical / Application Questions

1. A process consumes 180 units per day. Replenishment takes 0.75 day, the safety factor is 25%, and each container holds 30 units. Calculate the required number of kanban containers using the chapter formula, and explain why the result must be rounded upward.
2. A manufacturer has stable demand for standard fasteners, six-week lead times for imported electronics, and highly variable demand for customized assemblies. Recommend a control method for each category and justify the decisions.
3. An organization reduced kanban cards by 30% and then experienced stockouts, overtime, and emergency freight. Diagnose at least four possible causes and propose a corrective-action plan.

Multiple Choice Questions with Answers

Question 1

Which statement best describes a kanban?

A. A long-range aggregate sales forecast

B. An authorization signal for movement or production

C. A financial statement for inventory valuation

D. A quality inspection standard

Correct Answer: B. An authorization signal for movement or production

Question 2

A process has 360 available minutes and demand for 180 units. What is its takt time?

A. 0.5 minute per unit

B. 1 minute per unit

C. 2 minutes per unit

D. 180 minutes per unit

Correct Answer: C. 2 minutes per unit

Question 3

Which condition most strongly supports a pull system?

A. Stable demand and reliable short replenishment lead time

B. Unique products with six-month engineering lead times

C. Highly intermittent demand and no standard containers

D. A complex multilevel bill of materials with uncertain revisions

Correct Answer: A. Stable demand and reliable short replenishment lead time

Question 4

Using n = D × L × (1 + S) ÷ C, demand is 100 units per day, lead time is 1 day, safety factor is 20%, and container size is 25 units. How many containers are required after rounding upward?

A. 4

B. 4.8

C. 5

D. 6

Correct Answer: C. 5

Question 5

What is the primary purpose of heijunka?

A. To maximize batch size

B. To level production volume and mix

C. To eliminate all safety stock

D. To replace bills of materials

Correct Answer: B. To level production volume and mix

Question 6

Which statement correctly distinguishes VMI from consignment?

A. VMI concerns ownership, while consignment concerns forecasting

B. VMI concerns replenishment responsibility, while consignment concerns ownership

C. Both terms have exactly the same meaning

D. Consignment always requires supplier-managed replenishment

Correct Answer: B. VMI concerns replenishment responsibility, while consignment concerns ownership

Question 7

Under the simplified constant-demand assumption, average cycle inventory for a lot size of 200 units is:

A. 50 units

B. 100 units

C. 200 units

D. 400 units

Correct Answer: B. 100 units

Question 8

Which item is generally most appropriate for MRP rather than a simple kanban loop?

A. A repeatedly consumed standard washer with daily local delivery

B. A high-volume carton used at a stable rate

C. A long-lead imported component in a multilevel product structure

D. A standard label replenished from a point-of-use bin

Correct Answer: C. A long-lead imported component in a multilevel product structure

Question 9

In a hybrid planning system, which role is commonly assigned to MRP?

A. Authorizing every point-of-use withdrawal in real time

B. Planning complex dependent requirements and long-lead materials

C. Replacing all physical inventory locations

D. Eliminating the need for supplier agreements

Correct Answer: B. Planning complex dependent requirements and long-lead materials

Question 10

What is a major risk of reducing inventory without stabilizing processes?

A. Improved visibility of defects

B. Lower replenishment frequency in every case

C. Stockouts and expediting caused by unresolved variability

D. Automatic improvement in supplier reliability

Correct Answer: C. Stockouts and expediting caused by unresolved variability