Heijunka: What it is and how to level Lean production

What is Heijunka and why is it key in Lean Manufacturing?

Heijunka is a Japanese term, which translates as “leveling” and is one of the fundamental principles of the Toyota Production System.

To better understand what Heijunka is, its main objective is to keep the production of a product or its variants at regular levels as sustained as possible, and produced in as small batches as possible with frequent deliveries to the following processes.

What are the benefits of implementing Heijunka?

• Waste Reduction: By avoiding overproduction, reducing excess inventory, minimizing waiting times, lowering the number of necessary changeovers, and decreasing the number of defects produced.
• Improved Production Flow, making it more predictable.
• Greater Flexibility to Demand.
• Optimization of Existing Resources.
• Increased Stability for Workers.

How to implement Heijunka step by step

To explain the implementation of Heijunka more clearly, let's use an example.

In a home appliance factory, three types of washing machines are manufactured. The monthly demand for these washing machines is:

• Washing Machine A: 800
• Washing Machine B: 900
• Washing Machine C: 100

Considering this, the first step is to determine the basic leveled sequence. To do this, we must divide the production volumes of each item by the greatest common factor, which is the Greatest Common Divisor (GCD).

We start by breaking down the different volumes into their prime factors:

• Washing Machine A: 2 × 5 × 2 × 5 × 2 × 2 × 2 = 800
• Washing Machine B: 2 × 5 × 2 × 5 × 3 × 3 = 900
• Washing Machine C: 2 × 5 × 2 × 5 = 100

The GCD, common to all three products, is highlighted:

GCD = 2 × 5 × 2 × 5 = 100

To obtain our leveled sequence, we divide the total volumes by 100, resulting in:

• Washing Machine A: 800 / 100 = 8
• Washing Machine B: 900 / 100 = 9
• Washing Machine C: 100 / 100 = 1

A leveled production sequence would be:
A-A-A-A-A-A-A-A-B-B-B-B-B-B-B-B-B-C
This sequence should be repeated 100 times to reach the total demand.

Assuming that the available monthly working time for this production line is 300 hours, we proceed to calculate the cycle time of the basic sequence:

Cycle Time (Basic Series) = (300 × 60) / 100 = 180 minutes.

When scheduling production, it could be done in two ways:

• Mixed production: It is useful when there is a great variety of products and the preparation times between varieties are short. In the example case it would be: B-A-C-B-B-A-B-A-B-A-B-A-B-A-B-A-B-A-B-A-B-A-B-A-B-A-B.
• Production of monomodel series: In the case that the variety of products is not so wide, and there are considerable preparation times between varieties. In the case example it would be to make the leveled series: A-A-A-A-A-A-A-A-A-A-A-A-A-A-B-B-B-B-B-B-B-B-B-B-B-B-B-B-B-C

Another practical example to understand Heijunka compared to traditional planning

We have a toy factory in which 4 models are manufactured: A, B, C and D. The average monthly demand for each of them is as follows:

• Toy A: 32
• Toy B: 28
• Toy C: 10
• Toy D: 10

The following assumptions will be taken into account:

• The factory works one shift per day of 8 hours and 20 days per month
• The changeover time between products is 0
• Each unit costs 2 hours to produce.

If we were to do a traditional type of planning, we would produce first the A, then the B, then the C and then the D in large batches, and the planning would look like this:

As can be seen, in this type of traditional planning we have the following characteristics:

• Long production batches are chosen, as long as possible.
• The high runner products, the most ordered, are the first to be manufactured (A and B), leaving the least ordered products (C and D) for the last days of the month.

Let us imagine that a customer places the following order on the first day of the month:

• 1 product A.
• 1 product B.
• 1 product C.
• 1 product D.

The waiting time until the order is manufactured, i.e. the lead time, will be 19 days, since:

• We manufacture unit A on day 1.
• We manufacture unit B on day 9.
• We manufacture unit C on day 16.
• We manufacture unit D on day 19.

This is a very long waiting time, which can cause us to lose potential customers to competitors because of the wait.

However, if we apply Heijunka we can achieve drastic improvements.

The first thing we will do is to take the DCM (Greatest Common Divisor) to obtain the level series. For this case we will make the following approximation:

• Toy A: 32 --> Approximate to 30
• Toy B: 28 --> Approximate to 30
• Toy C: 10
• Toy D: 10

Thus we get the MCD to be 10, and it will be in the last series where we adjust the quantities, therefore the planning would be like this (we have opted for the type of production of monomodel series):

As can be seen in this case, the lead time delivery to the customer who has placed the previous order, would be in this case of 2 days, so the gain in delivery time is very substantial.

In addition, the fact of producing with these short production batches, makes the payback, that is, our investment and costs in materials to manufacture, is returned in a shorter time than traditional planning, thus improving the cash flow of the factory.

What is Heijunka Box?

The Heijunka Box is basically a box, which contains the following:

• A horizontal rows representing each product.
• Vertical columns for each day of the work week.

The Heijunka Box can be visualized in different ways, from a real box, to a flat drawn surface including the work items with notes with sticky notes.

Following the example above, let's imagine for example cards for the first given week, which would be placed as follows, and would be picked up as Kanban cards.

Common mistakes when applying Heijunka

The implementation of Heijunka is often faced with several common mistakes:

• Not having a stable or predictable demand, as this is the basis for being able to apply Heijunka.
• Lack of flexibility in the line: if quick reference changes cannot be implemented, it can be counterproductive.
• Resistance to change by personnel: this change in the way of scheduling can cause rejection in the team if not properly communicated
• Insufficient capacity in downstream processes, which would lead to bottlenecks and unnecessary inventories.
• Not having an efficient Kanban system: the Kanban system must be properly designed.
• Not considering variability in suppliers: if supplier deliveries are not made on time they will affect heijunka leveling.
• Not measuring the impact of the implementation with KPIs: kpis must be implemented to see if there is a real improvement in terms of accumulated inventory and lead time.