How do you calculate the pitch for a heijunka box?

Pitch equals takt time multiplied by the pack-out quantity, the time it takes to build one container of parts. If takt is one minute per unit and a container holds 20 units, the pitch is 20 minutes, which becomes the width of each column and how often the handler withdraws kanban.

How do you load a heijunka box?

Count how many containers of each product are needed per day by dividing demand by the pack-out quantity, reduce those counts to their smallest whole-number ratio, and lay out the shortest repeating cycle that honors the ratio. Then place one kanban per container into the slots following that pattern, repeating the cycle across the shift.

What is the difference between heijunka and a heijunka box?

Heijunka is the concept of leveling production by volume and mix. The heijunka box is the physical tool that carries out a leveled plan, the slotted rack of kanban that paces the line. You can understand heijunka without a box, but the box is how most plants actually execute leveling on the floor.

Who moves the cards in a heijunka box?

A material handler, sometimes called a water spider or runner, withdraws one column of kanban at each pitch interval and delivers the cards to the pacemaker process. This is often part of a timed material-delivery route that also brings components and removes finished containers.

What does it mean when cards pile up in the box?

It means the line is falling behind schedule. Because the handler should have pulled those cards by now, stale cards sitting in a past column are an immediate, visible signal of a problem, a long changeover, a late component, or an unstable cycle, and the number of stale cards shows how far behind the line is.

Do you need short changeovers to use a heijunka box?

Reasonably short changeovers help a lot, because every pitch boundary can require a setup. Slow changeovers force a coarser pitch and less leveling. Running a SMED program to cut changeover time lets you shrink the pitch and level the mix more finely, which is why the two practices are usually done together.

The Heijunka Box (Leveling Box)

Q: What is a heijunka box?

A heijunka box, also called a leveling box, is a physical grid that holds a leveled production plan. Rows represent products, columns represent fixed time intervals called pitches, and kanban cards in the slots dictate what the line builds and in what order. A material handler works the box one column per pitch.

How to size the pitch, load the kanban, and read the leveling box that turns a smoothed plan into a build sequence anyone can see.

A heijunka box (leveling box) is a physical grid that holds a leveled production plan. Each row is a product, each column is a fixed time interval called a pitch, and kanban cards placed in the slots tell the line exactly what to build and in what order. A material handler works the box one column at a time.

The heijunka box is the tool that makes a leveled schedule real. Leveling itself, spreading volume and mix evenly instead of building in big batches, is the concept, and our guide to heijunka and production leveling covers why it matters and what it does to a plant. This article is about the box: how to size it, how to load the cards slot by slot, and how to run it on the floor. It is a piece of visual management for the schedule and a practical anchor of lean manufacturing on the floor: once it is loaded correctly it lets anyone standing in front of it see what should be building right now, what comes next, and whether the line is keeping pace.

What Is a Heijunka Box?

A heijunka box is a slotted rack, usually mounted on a wall near the pacemaker process. Horizontal rows each hold one product or part number. Vertical columns each represent an identical, short time interval for the paced withdrawal of kanban (Lean Enterprise Institute, Heijunka Box). The kanban cards sitting in the slots each authorize one increment of production for one product. Read down a column and you see everything due in that pitch; read across a row and you see how often that product is scheduled through the day. The box does two things a spreadsheet cannot: it makes the leveled sequence physical, so the plan cannot silently drift, and it makes falling behind visible, because cards that should have been pulled are still sitting in the box.

How Do You Calculate the Pitch?

The pitch is the width of each column in time, and it is the first number you need. Pitch is takt time multiplied by the pack-out quantity, in other words, the time it takes to produce one full container of parts. Takt time is available production time divided by customer demand; the pack-out quantity is how many units go in the container or tote you move as one unit. Multiply them and you get the rhythm at which the material handler will work the box.

Pitch is takt time multiplied by the pack-out quantity: the time to build one container. It sets how wide each column of the heijunka box is and how often the handler works it.

Pitch is a design choice as much as a calculation. A shorter pitch levels the mix more finely and keeps the box responsive, but every pitch boundary is a potential changeover, so a fine pitch is only affordable when setups are quick. That is why SMED and the heijunka box go together: cheaper changeovers let you shrink the pitch and level harder. If setups are slow, start with a coarser pitch and tighten it as changeover times come down.

How Do You Load the Box, Slot by Slot?

This is where the box gets real. Take a worked example. A line builds a family of three products, A, B, and C, and total demand is 480 units a day against 480 minutes of available time, so takt is one minute per unit. The container holds 20 units, so pitch is 20 minutes and there are 24 pitches, hence 24 containers, in the day. Demand splits 50/33/17: product A is 240 units a day (12 containers), B is 160 (8 containers), and C is 80 (4 containers).

Now level the mix. This assumes three products stable enough to level; before you get here, a Glenday sieve is a good way to find the handful of high-runners worth putting on a fixed cycle in the first place. Twelve, eight, and four containers reduce to a ratio of 3 to 2 to 1, so a repeating six-pitch cycle of A-B-A-C-A-B delivers exactly that ratio, and the cycle repeats four times across the 24-pitch day. You load the box to that pattern: a kanban for A in pitches 1, 3, and 5; a kanban for B in pitches 2 and 6; a kanban for C in pitch 4. Repeat the block four times and every container of the day has a slot.

The loaded box for the worked example: three A cards, two B cards, and one C card per six-pitch cycle hold the 50/33/17 mix. Repeat the block four times to cover the day.

How Do You Build a Heijunka Box? Step by Step

With the pitch and pattern understood, here is the build sequence for a first box:

Confirm the pacemaker and its takt. The box paces one process, the pacemaker. Set takt from averaged demand over your planning period, not from a single spiky day.
Pick the pack-out quantity and compute pitch. Choose the container size you already move as one unit, multiply by takt, and you have the column interval.
Count containers per product per day. Divide each product's daily demand by the pack-out quantity to get the number of kanban cards that product needs in the box.
Reduce the mix to a repeating pattern. Simplify the container counts to their smallest whole-number ratio and lay out the shortest cycle that honors it, keep it boring and repeatable.
Check the pattern against changeover reality. If the cycle demands more setups than current changeover times allow, coarsen the pattern or run SMED first. Do not load a box the line cannot physically run.
Build the physical rack and cut the cards. One row per product, one column per pitch, enough slots for a full shift. Make one kanban per container and load them to the pattern.
Assign the handler route and start. Define who pulls a column each pitch, where they deliver the cards, and how empties return. Then run it and watch what breaks first.

How Does the Material Handler Run the Box?

The box does nothing on its own; a material handler makes it move. At each pitch interval the handler withdraws the next column of kanban and delivers those cards to the pacemaker, which builds exactly what the cards specify, in sequence. The withdrawal interval is the pitch, so on a 20-minute pitch the handler works the box every 20 minutes, all shift. This is often folded into a timed material-delivery route so the same trip drops kanban, brings components, and takes away finished containers. The handler's steady loop is what converts the static grid into a live pacemaker: the plan does not advance because a computer says so, it advances because a person pulls the next card on time.

How Do You Read the Box at a Glance?

The best feature of a heijunka box is that it fails loudly. If the current column is empty and the pitch is not yet over, the line is on or ahead of schedule. If cards are still sitting in a column whose time has passed, the line is behind, and you can see exactly how far by how many stale cards remain. No status meeting, no report, the box shows adherence continuously to anyone walking by. That visibility is the point. When cards start piling up, you have found your next problem to solve: a changeover that ran long, a component that arrived late, an unstable cycle. Measure schedule adherence by the pitch, not by the day, and the box tells you where to look. Keeping the pacemaker steady this way is also what lets downstream standard work hold, since the sequence stops lurching.

By the Numbers: What Leveling With a Box Buys You

The payoff of loading a box instead of running campaigns shows up in inventory. When production runs in large batches, finished goods pile up ahead of real demand, and U.S. Census Bureau data on manufacturing shows inventories-to-shipments ratios running around 1.4 to 1.5 in recent years (U.S. Census Bureau, M3 Survey), roughly a month and a half of shipments sitting as stock across the sector. A heijunka box attacks the batch-driven share of that directly: building small quantities of every product every day, paced by the box, means less cash parked in the warehouse and fresher stock when demand moves. The trade is explicit, a small finished-goods buffer absorbs daily order noise, but the buffer is far smaller than the batch inventory it replaces.

Where Harmony fits: the classic box is cardboard and kanban, and it works, but it cannot see whether the line actually hit each pitch or which changeover ran long on nights. Plants that put live floor visibility and AI scheduling over their existing systems keep the box's leveling logic and add a feedback loop: pitch-by-pitch adherence tracked automatically, and the sequence recalculated when demand or downtime moves, instead of waiting for the next planning cycle. See it in a plant like yours in the CLS case study. The principle stays Toyota's; the clipboard just gets a live signal, with no rip-and-replace to get there.