The Yamazumi chart: seeing imbalance before it costs money

The Yamazumi (山積み, "to stack") is manufacturing's most honest chart: one bar per station, the work content stacked inside it, and a horizontal line marking TAKT. Nothing else. And yet, in a single glance it answers the three questions that define a line's health:

Where does it break? Any bar peeking above the TAKT line is a bottleneck: that station can't reach the pace demand requires, and it drags the whole line with it.
Where is waste hiding? Any bar far below TAKT is paid capacity that isn't producing: structural idle time, shift after shift.
How fragile is the balance? A line where several bars graze the TAKT "delivers" today, but any demand increase or absence breaks it tomorrow.

Imbalance has a price — and it can be calculated

The gap between the tallest bar and the rest is not a cosmetic detail: it's money. If one station runs at 96% and another at 69%, that second station spends roughly a third of its shift waiting. Multiply those minutes by the cost of a labor-minute and by the month's shifts, and the "detail" becomes a figure that justifies — or doesn't — reassigning operations, adding a station, or changing staffing.

That calculation is the argument engineering teams usually need in front of management: not "the line is unbalanced," but "this line's imbalance costs X per shift."

The problem isn't the chart — it's the snapshot

Almost every plant has Yamazumis. The problem is that they tend to be snapshots: someone builds them in a spreadsheet from a one-off time study, they're shown in a meeting, and they're obsolete by the next change in demand or mix. The chart was correct; the line is already a different one.

For a Yamazumi to drive decisions — not just documentation — it has to recalculate at the pace the operation changes:

When demand changes, TAKT moves and stations recolor: yesterday's slack can be today's bottleneck.
When the mix changes, work content per station shifts even if total demand stays the same.
When new personnel come in, their stations' real capacity temporarily drops — the learning curve stacks into the bar too.

From snapshot to living model

With Atlas we turned the Yamazumi into exactly that: a living model. Demand moves with a slider and every station recalculates instantly — bottlenecks in red, risk in amber, slack in green — along with optimal staffing and the cost of unproductive time. The same chart as always, but answering for today's line, not last month's.