Heijunka Practical Guide: From Batch Fluctuations to Smooth Flow Production

Abstract: Many factories face production bottlenecks not due to equipment but due to "scheduling fluctuations" — idle at the beginning of the month, rushed at the end, with concentrated large batches and frequent changeovers leading to quality issues. Heijunka (levelled production) is the key to solving this dilemma: it evenly distributes the variety and quantity of production over a period, ensuring a smooth flow on the production line every day and every hour. This article systematically explains the essence of Heijunka, two leveling methods, the Heijunka Box, and EPE scheduling rules, and provides a five-step implementation path along with methods for integrating it with quality management.

1. The Essence of Heijunka: Leveling Demand Before Production

1.1 What is Heijunka Production

Heijunka production originates from the Toyota Production System. It involves evenly distributing the variety and quantity of production over a period to meet total customer demand, rather than following the original order arrival rhythm of "do whatever comes in." The core question it addresses is: how to ensure consistent delivery while maintaining a nearly constant production pace every day, rather than fluctuating between "idle" and "overloaded."

Understanding Heijunka requires recognizing its distinction from traditional scheduling: traditional scheduling passively responds to demand fluctuations, leading to overtime during peak periods and idle personnel and equipment during troughs. In contrast, Heijunka actively "shapes" demand, using a balanced plan to level out fluctuations outside the production system, ensuring that internal processes face a steady pace.

1.2 Two Types of Leveling: Quantity Leveling and Variety Leveling

• Quantity Leveling: Distributes the total volume over a period evenly to each day and shift. For example, if a customer places a week's worth of orders on Monday, Heijunka does not mean working hard on Monday and finishing up from Tuesday to Friday. Instead, it divides the week's total volume by five working days, producing about one-fifth each day.
• Variety Leveling: Interleaves different varieties into the production sequence to avoid "only producing A this month and only producing B next month." For instance, if the demand ratio for products A and B is 3:1, the production schedule follows an "A, A, B" cycle rather than completing all of A before moving to B.

These two types of leveling are often used together: first, the total volume is leveled by quantity, then the sequence is mixed by variety, ultimately forming a production rhythm that is "similar every day and similar every hour."

1.3 Heijunka is Not "Deviation from Orders"

Some worry that Heijunka might deviate from customer needs. In reality, Heijunka changes the "production start rhythm" rather than the "delivery commitment." As long as the finished goods supermarket (buffer inventory) is set reasonably and the delivery cycle meets customer requirements, Heijunka can bring more stable delivery performance by eliminating the quality and delivery risks caused by internal rush work. Heijunka and order-based production are not mutually exclusive but rather a combination of "front-end leveling and back-end pull": the fixed-rate process follows a balanced plan, while downstream processes are pulled by kanban, absorbing demand changes through supermarket levels and rolling plans.

2. Why Heijunka is the Common Foundation for Quality and Efficiency

2.1 Leveling Exposes "Hidden Rocks"

Taiichi Ohno used the metaphor of "water level and hidden rocks" to describe inventory: when the water level is high, quality issues, equipment failures, and slow changeovers are hidden beneath the surface. As the water level drops, these hidden rocks become exposed. Heijunka is a proactive method to "lower the water level" — with demand leveled, the buffer inventory between processes no longer needs to cover peaks and troughs, and the issues hidden by inventory will surface one by one, forcing the team to address the root causes. In this sense, Heijunka is not just a scheduling technique but a management tool for exposing problems and driving improvements.

2.2 Stable Pace Ensures Stable Quality

Concentrated large-batch production may seem to reduce the number of changeovers, but each changeover involves a high-risk window of "parameter adjustment — trial production — first article inspection," and any issues that arise are often batch-related. Heijunka breaks down large changeovers into multiple small ones, making each switch simple, standardized, and predictable. Combined with standardized changeover procedures and first article inspections, the risk of each changeover is significantly reduced. More importantly, a steady pace ensures that each workstation operates at a fixed rhythm, stabilizing the work rhythm of operators and significantly reducing human errors such as missed inspections and incorrect installations.

2.3 Prerequisite for Pull Systems

Kanban, supermarkets, and other pull mechanisms require relatively stable demand over time. If downstream demand fluctuates significantly, the number of kanban cards will either need frequent adjustments or become ineffective, degrading the pull system into a "disguised push." Heijunka first "shapes" the demand entering the system into a stable sequence, allowing the pull system to operate reliably under a steady pace — this is why Heijunka is called the "prerequisite for pull production."

2.4 Suppressing the Bullwhip Effect

Demand information amplifies as it moves through the supply chain, known as the bullwhip effect: a small fluctuation at the customer end can become a severe oscillation upstream. Heijunka reduces the amplitude of fluctuations at the source, combined with a balanced ordering rhythm, significantly mitigating the bullwhip effect and benefiting the scheduling and quality control of suppliers — stable suppliers naturally provide more stable incoming quality.

3. Core Methods of Heijunka: Heijunka Box and Scheduling Rules

3.1 Heijunka Box

The Heijunka Box is the most classic tool for implementing Heijunka: a panel divided into small grids along a time axis, each grid representing a production cycle (e.g., every 30 minutes or per shift). Production kanban cards representing different varieties are inserted into the corresponding grids according to the plan. Production scheduling follows the sequence of grids, taking cards and organizing production, advancing one grid at a time. The Heijunka Box transforms a "balanced plan" into a visible production rhythm, making it clear to anyone what should be done today or in the current hour, and making it easy to identify plan deviations.

3.2 Calculation of Quantity Leveling

For example, with a monthly demand of 8000 units and 20 working days in a month: 400 units should be produced each day. If a single shift has 400 effective working minutes, the takt time = 400 minutes ÷ 400 units = 1 minute/unit. Quantity leveling involves breaking down the "monthly total" into "daily, shift, and hourly quantities," making the production target for each time unit clear and executable.

3.3 Variety Leveling: EPE and E/I Rules

The core concept of variety leveling is EPE (Every Part Every Interval): each variety is produced at least once within a fixed time interval. The shorter the interval, the more frequent the changeovers but the lower the inventory; the longer the interval, the fewer changeovers but the higher the inventory. The choice of EPE interval must balance changeover costs and inventory costs.

The E/I rule (Every Interval) further specifies the frequency of variety interleaving: if the demand ratio for A and B is 3:1 and EPE is set to 1 day, the daily cycle is "A, A, B"; if the demand ratio is 5:1, it can be set to a two-day cycle of "A, A, A, A, B, A, A, A, A, B." Scheduling example: A has a weekly demand of 300 units, B has a weekly demand of 100 units, with a single shift of 5 days and a takt time of 1 minute, then 60 units of A and 20 units of B are produced each day, interleaved in an "A, A, B" cycle, rather than "only A for the first two days, only B on the third day."

3.4 Connection Between Fixed-Rate Processes and Supermarkets

Heijunka plans are typically only issued to "fixed-rate processes" (usually assembly or bottleneck processes). The fixed-rate process follows a balanced sequence, while other processes are pulled by supermarkets and kanban. The upstream of the fixed-rate process flows smoothly, and the downstream is replenished by pull, ensuring the entire system's balance without imposing complex plans on every process.

4. Five-Step Implementation Method for Heijunka

Step One: Clarify Real Demand. Collect customer orders, sales forecasts, and historical shipment data, distinguishing between "confirmed orders" and "forecasted demand," and determine the base demand for the leveling cycle. Accurate demand data is the foundation of Heijunka; without it, everything else is futile.

Step Two: Determine Takt Time and Cycle Time. Calculate the takt time (available time ÷ customer demand), measure the actual cycle time of each process, identify bottleneck processes, and confirm whether the Heijunka plan can be implemented at the bottleneck.

Step Three: Design EPE and Scheduling Sequence. Determine the EPE interval based on variety demand and changeover time, generate a cyclic scheduling sequence, and simulate and validate: whether the total changeover time fits within the available time, whether the supermarket level is sufficient, and whether the delivery commitment is met.

Step Four: Set Up the Heijunka Box and Fixed-Rate Processes. Create a Heijunka Box (or digital scheduling board), define the card and grid rules, specify the fixed-rate processes and pull connection points, and simultaneously establish finished goods supermarket and safety stock rules.

Step Five: Pilot Run and Visual Management. Select representative varieties for a pilot run of one to two weeks, observe the plan achievement rate, changeover frequency, and inventory levels, expose anomalies through visual kanban, and iteratively adjust EPE and supermarket parameters until a stable rhythm is formed.

5. Synergy Between Heijunka and Quality Management

5.1 Controlling Quality Risks During Changeovers

Heijunka transforms "few large-batch changeovers" into "many small-batch changeovers," making changeovers a daily routine. The quality risks associated with changeovers must be systematically managed: use SMED (Single-Minute Exchange of Die) to reduce changeover time to the minute level; standardize changeover steps and document them in work instructions; perform first article inspections after each changeover, and do not proceed to batch production if the first article is nonconforming. The more frequent the changeovers, the more important it is to create a standard closed loop for "changeover — first article — release."

5.2 Nonconforming Products are the Top Disruptors of Heijunka

Nonconforming products in the production process can lead to material shortages and line stoppages, forcing emergency orders to recover lost production, which immediately disrupts the balanced sequence. Therefore, the stability of Heijunka is highly dependent on quality stability: on one hand, process control and poka-yoke intercept nonconformities at the source; on the other hand, maintaining reasonable safety stock for key bottleneck varieties serves as a "buffer." Stable quality is the prerequisite for the continuous operation of Heijunka, and Heijunka, in turn, promotes quality improvement by exposing issues, creating a cause-and-effect relationship.

5.3 Using Quality Data to Inform Scheduling

Mature factories use data-driven Heijunka decisions: incorporating first-time yield, nonconforming distribution, overall equipment effectiveness (OEE), and first-time changeover success rate into scheduling inputs. For example, if a variety has a high recent nonconforming rate and long rework time, its daily production volume is appropriately reduced and more buffer is added; if a process's OEE declines, maintenance windows are scheduled in advance to avoid "bottlenecks" in the balanced sequence. Heijunka is not a static plan but a dynamic rhythm continuously calibrated based on quality and equipment data.

5.4 Heijunka and Digital Scheduling

Traditional Heijunka Boxes are made of paper and cards, but in the digital age, APS (Advanced Planning and Scheduling) systems can be used: automatic re-scheduling of balanced sequences in response to demand changes, automatic verification of changeover times and capacity constraints, and automatic alerts for anomalies. However, digital tools do not change the logic of Heijunka — clear balanced rules must come first, followed by systematic automatic execution. Do not blindly implement systems without clear rules.

6. Common Misconceptions and Applicability Boundaries

6.1 Five Common Misconceptions

• Misconception One: Heijunka is Just Average Allocation. Without takt time calculation, EPE design, and supermarket support, "averaging" is only superficial balance and will inevitably lead to chaos.
• Misconception Two: Ignoring Changeover Time. Long changeover times without SMED will result in significant capacity loss, forcing a return to large-batch production.
• Misconception Three: No Buffer Support. Heijunka requires finished goods supermarkets and safety stock to absorb demand fluctuations; a Heijunka without any buffer cannot withstand any disturbances.
• Misconception Four: Treating Forecasts as Orders. Large forecast deviations but rigid Heijunka plans can lead to stockouts or overstock; it is essential to distinguish between confirmed orders and forecasts and handle them in layers.
• Misconception Five: Leveling Only the Main Plan, Not Execution. A balanced plan that is not followed by consistent execution on the shop floor, with random order insertions and sequence jumps, will render Heijunka ineffective.

6.2 Applicability Boundaries and Flexibility

Heijunka is most suitable for scenarios with relatively stable demand, clear variety structure, and repetitive production. For businesses with highly volatile demand or project-based or custom production, flexible approaches can be adopted: segment leveling for common components, delaying differentiation (postponing differentiation) for differentiated parts, or using modular design to convert custom demands into combinations of standard modules, thereby achieving Heijunka benefits on a larger scale. Understanding boundaries is not about abandoning Heijunka but knowing where and at what granularity to implement it.

7. Conclusion

The entire value of Heijunka production can be summarized in one sentence: using a balanced plan to ensure stable production and, in turn, reliable quality. It levels out demand fluctuations outside the system, ensuring a steady flow on the production line, exposing issues, making pull systems truly operational, and making changeover risks manageable.

When demand data is solid, takt time is accurately calculated, EPE is reasonably designed, supermarkets and kanban are well-connected, and quality data continuously informs scheduling, Heijunka is no longer a theoretical concept on a scheduler's desk but a daily rhythm in the workshop — every minute is used evenly, every variety appears at the right time, and every product is manufactured in a stable process. This is the most harmonious cooperation between lean manufacturing and quality management.

The essence of Heijunka production is to use a balanced plan to ensure stable production and, in turn, reliable quality — leveling demand fluctuations before production and exposing issues at a controllable time.

Knowledge code: 7.1.2

Version: v20260805

Author: Quality Think Tank The Quality Think Tank is dedicated to providing systematic professional knowledge, methodologies, and practical tools for quality management practitioners, helping companies continuously improve their quality capabilities.