metric 82

Synchronises pace of production to pace of sales.

Shift length minus breaks, meetings and cleanup.

Denominator of takt time.

Divisor that converts inventory quantity into days of supply.

Consistent increment of work released and withdrawn at the pacemaker.

Time between one part coming off the process and the next.

Time to switch the process from one product type to another.

Fraction of time the machine runs when it is asked to run.

How frequently a process changes over to make all variations.

Quantity produced between two changeovers.

People required to run the process.

How many part numbers this process must serve.

Number of parts one finished-goods container holds.

Fraction of output that must be scrapped or reworked.

Fraction of parts that clear the process with no rework.

The share of parts crossing the whole stream with no rework anywhere, being the product of every process's first-time-through.

Good parts as a share of what the scheduled time could have produced: availability times performance times quality.

Parts produced against the target output for the time the plant ran — what short stops, idling and reduced speed cost once availability is accounted for.

Operating pattern that scales available time.

Converts monthly demand into a daily rate.

Output achievable with no changeovers.

Sum of operator work across the cell — numerator for manning.

Continuous-flow manning implied by work content and takt.

Loading of one operator against takt; > 1 means the cell cannot hold pace.

How evenly work is distributed across operators.

Sets the smallest practical batch size at a fabrication process.

What the current batch policy actually demands of the process each day.

Typical target is ~10% of available time spent changing over.

A large positive gap signals hidden downtime being compensated for.

Observed accumulation at a triangle — where flow has stopped.

Lead time contributed by one inventory triangle.

Time one part takes from raw material arrival to customer shipment.

Sum of value-adding times — the shocking comparison against lead time.

Fraction of elapsed time that actually adds value.

Financial mirror of production lead time.

Little's Law inside one process box.

Controlled inventory sized to cover replenishment plus a hedge.

Cards in circulation cap the WIP of a pull loop.

Hard cap that stops the upstream process from overproducing.

Temporary hedge against downtime or demand spikes.

How long a pull loop takes to put parts back in the supermarket.

More frequent delivery is the cheapest inventory reduction available.

Only processes from the pacemaker downstream count toward the quote.

Column width of the load-levelling box.

Drives the repeating sequence loaded into the heijunka box.

A lean future state schedules exactly one point: the pacemaker.

Every push arrow is an isolated island — future state should drive it to zero.

Processes producing to their own schedule rather than a customer.

Transport waste made visible: metres one part travels across the stream.

Space released is a headline benefit of removing inventory.

Customer-facing outcome measure.

The single best lean measure: is the takt-based target hit every day?

Symptom of releasing too much work too early.

Direct measure of the most significant source of waste.

N — the units the shift is asked to finish. Sets both pace and manning.

n = ⌈Σtₑ/c⌉. Below this the shift schedule cannot be met at all.

c̄ — the work each operator would carry if it divided evenly. The target a balance aims at.

c* = max(cᵢ) — the pace the line actually runs at.

d — the share of paid time the line stands idle because work was assigned unevenly.

Balance delay expressed as labour seconds thrown away on every unit built.

Nₑ — how finely the work is described. Too few elements and no balance is reachable.

tₑ — the standard time of one task, the grain a line can be rebalanced at.

Places on the line. Never more than the operator count, and fewer when a station is doubly manned.

Copies of one station working alternate units, so a task longer than takt stops setting the pace.

w — centre-to-centre distance between consecutive units, which fixes conveyor speed and line length.

v = w/c* — how fast the line must move to deliver a unit every effective cycle.

Floor the line needs once the stations and unit space are fixed.

How much work content the variants share — the test of whether they are one product family.

Share of incoming work this process can start without sending it back.

Product of every process's %C&A — the share of work that crosses the stream without a single information loop-back.

Seconds the operator spends walking per unit in the cell.

How much of the operator cycle is walking rather than work — the figure a U-cell is meant to shrink.

Unattended machine time after load. Compared with operator work it says whether the operator waits on the machine or works in-cycle.

The allowed pieces in the cell. A map that cannot state this cannot claim one-piece flow.

How long a stoppage lasts, as distinct from how often one happens. Uptime gives the frequency; this gives the duration, and it is the duration that decides whether a boundary can be coupled and how deep a lane has to be.

What a process can deliver in a day once its own losses are taken out, set against what the day asks of it. A cycle time inside takt is not capacity, and a process below one here cannot be buffered into adequacy: the shortfall recurs every day the buffer is drawn from.

How often the process actually stops to change over in a shift — the load SMED has to buy down.

How many source-inspection devices sit on the process. Zero with a scrap rate is a finding, not a style choice.

The part of the changeover the machine is stopped for. This — not the total setup — is what limits batch size and EPE.

Setup work already done while the machine still runs. Stating it is how a plant proves stage 1 of SMED actually happened.

How much of the whole setup still stops the machine. SMED stage 2 is the business of driving this down.

Process time over lead time, as Martin and Osterling print it on the map header.