Free Lean Tool
Batching all of product A before touching product B feels efficient — and quietly inflates lead time, inventory, and upstream chaos. A heijunka box levels production by volume and mix: every pitch interval, one container of something, in the proportions the customer actually buys. Enter your demand below and watch the box load itself.
Production leveling by volume and mix
Heijunka is leveling production by both volume and mix over a fixed period. Instead of running all of product A in the morning and all of product B in the afternoon, you alternate small runs of each — so every part of the shift produces the same mix the customer actually buys.
| Product | Demand | Pack-out | Pitch = takt × pack | Cards = demand ÷ pack ↑ |
|---|---|---|---|---|
| A-100 Housingrunner · sets box pitch | 240 | 20 | 20 min | 12 |
| B-200 Housing | 120 | 20 | 20 min | 6 |
| C-300 Cover | 80 | 20 | 20 min | 4 |
| D-400 Bracket | 40 | 20 | 20 min | 2 |
Same shift, same demand, same containers — only the loading of the box differs. The charts show units started per quarter of the shift, by product: batching starves upstream processes of a steady mix, leveling pulls the same small amounts of everything all shift long.
| Metric | Batched | Leveled |
|---|---|---|
Longest run of one product The biggest unbroken block of a single product. Long runs are what batching looks like on the floor. | 12 cards (4 h) | 2 cards (40 min) |
Longest wait for a product Worst case across products: the longest stretch of the shift with no container of some product being made. | 7 h 20 m (D-400 Bracket) | 3 h 40 m (D-400 Bracket) |
Full mix available by Clock time when at least one container of every product has been completed. | 13:40 | 08:20 |
Want an editable copy?
Download the free Heijunka Box template — enter available time, demand, and pack-out quantities and it calculates takt, pitch, and kanban cards per shift, with a worked leveling example. Opens in Excel and Google Sheets.
Heijunka is the practice of leveling production over a fixed period in two dimensions at once: volume (the same amount of work each interval) and mix (the same proportions of each product each interval). Instead of assembling all of the month’s A-items in one long run and then switching to B-items, a leveled schedule alternates small runs of every product all shift long. The payoff shows up everywhere downstream of the decision: shorter lead times, smaller inventories of both finished goods and parts, defects caught after one container instead of after a whole batch, and a workload that is steady instead of alternating between overburden and idleness.
A heijunka box is the physical (or, here, digital) scheduling device that makes the leveled plan visible and operable. It is a grid of slots: rows are products, columns are pitch intervals labeled with clock times. Each slot holds at most one kanban card, and each card authorizes exactly one container of one product. A material handler walks a fixed route on a fixed cycle — paced withdrawal — pulling the next card from the box at every pitch, delivering it to the pacemaker process, and carrying the finished container away. The box turns the schedule into a heartbeat: if a card is still sitting in its slot after its time has passed, the line is behind — visible within one pitch, not at the end of the shift.
Pitch links takt time to a practical management timeframe. Takt time — available time divided by demand — is often under a minute, too fine-grained to schedule or check against. But customers do not buy single units; they buy containers. So the box paces itself on containers:
pitch = takt time × pack-out quantity
If takt is 60 seconds and a container holds 20 pieces, the pitch is 1,200 seconds — every 20 minutes, one full container should be completed and withdrawn. Each product’s kanban card count per shift is its demand divided by its pack-out quantity, rounded up to whole containers. This tool derives all three numbers transparently and warns when the arithmetic is telling you something: demand that is not a whole number of containers, a pitch that does not divide the shift cleanly, or more cards than the shift has intervals — which is overburden, not a scheduling detail.
The strongest argument for heijunka is invisible at the line itself: it is what the schedule does to everyone who feeds the line. A batched schedule pulls parts in batches — the supplying processes and suppliers see feast-or-famine demand, and they protect themselves the only way they can, with inventory and expediting. A leveled schedule pulls a little of everything at a steady rhythm, which is exactly the demand signal a kanban loop needs to run with few cards. Leveling the pacemaker is what makes small supermarkets, short replenishment loops, and calm upstream processes possible. The batched-vs-leveled comparison in this tool makes that visible: same demand, same containers, radically different longest runs, waits, and mix availability.
Click Load Example to see a four-product mixed-model line, or enter your own available time, products, demands, and pack-out quantities. The tool computes takt, per-product pitch, and kanban cards per shift, then loads the box with an automatically leveled sequence — each product spread as evenly as possible across the shift. Toggle Batched to see the same demand loaded the traditional way, and read the comparison panel below the box. On desktop, drag cards along their rows to experiment with your own sequence; the load row and metrics update live. When you are done, use Download Printout for a print-ready sheet, or grab the Excel template above.
One boundary worth knowing: this tool levels the schedule assuming your changeovers are already quick enough to alternate products every pitch. If they are not, the EPEI Calculator is the honest first step — it tells you how often your process can currently afford to cycle through the whole mix, and how much setup reduction it would take to level further.