Repetitive manufacturing: running a line to a rate instead of a work order

Rate replaces the work order as the unit of control

Scheduling a repetitive line means agreeing a daily or weekly quantity and letting the line run, with consumption posted automatically as units are declared complete. The plant commits to holding that pace: staffing, maintenance windows and material delivery all key off it. Systems have to support rate-based schedules, backflush at a defined point, and scrap reporting that corrects the automatic consumption, because an uncorrected backflush quietly destroys stock accuracy. Cost collection changes shape too — variances appear per period and per line rather than per job, so trouble shows up as drift rather than as one bad order.

The products that earn a dedicated line

A line is worth dedicating where the design is stable, volume is steady enough to fill it, and variants differ in ways the same stations can absorb. Capital goes into fixed conveyance, purpose-built fixtures, station tooling and test equipment matched to one family, which is cheaper per unit than flexible equipment and worthless for anything else. That trade is the whole decision: the plant buys low unit cost by surrendering the option to make something different next season. Families with long lives and predictable engineering changes justify it; volatile ranges do not.

Where the stock goes when work in progress disappears

A balanced line holds very little between stations by design, so material that would otherwise sit part-finished moves to the two ends. Incoming components need a feed buffer sized to delivery frequency and supplier reliability, while completed units accumulate against a shipping pattern that rarely matches the production pace exactly. The consequence is that accuracy depends on flow rather than on counting: a stoppage drains the feed buffer within a shift, and an unnoticed scrap rate leaves the record holding material that physically left the building hours earlier. Cycle counting at the buffers, rather than across the whole store, catches most of that error early.

Quality built at the station, not in the inspection bay

Because a defect made at pace is repeated at pace, detection has to sit close to the operation that creates it. Error-proofing at the station, first-off checks after any adjustment and simple in-process measurement are the practical controls; a final inspection bay merely tells you how many units to sort. Process drift is the characteristic failure — a tool wearing, a fixture loosening, a supplier's material shifting within tolerance — producing a run of conforming-looking parts that fail later. Containment planning follows: knowing which units were made between two checks defines the size of any recovery.

Where a fixed pace stops paying

Rate-based lines lose their advantage once demand turns irregular or the mix broadens. Every additional variant adds changeover, and when changeovers consume a meaningful share of available time the line behaves like a batch operation with worse flexibility. Upstream, suppliers must deliver in step with the pace rather than in convenient lots, which pushes purchasing toward long-term agreements, frequent small deliveries and quality measured per shipment. Sourcing turns into a long-lived commitment: qualifying a replacement for a component consumed continuously is a project, not a purchase. The signal to abandon the arrangement is usually visible in the changeover log long before it shows up in the cost report.

Frequently asked questions

How do we keep stock accurate when materials are backflushed?

Backflush only works when the bill of materials matches what the line actually consumes, so audit usage against theoretical consumption on a fixed cycle and correct scrap factors rather than leaving them at their original values. Report scrap and rework where they occur, not at period end. Set the backflush point where physical confirmation exists — a completion scanner, a test pass — and count key components physically often enough to catch drift the transactions cannot see.

Can a rate-based line handle more than one product?

Yes, provided the variants share stations, cycle time and tooling closely enough that switching costs little. Mixed-model running with sequenced variants works when work content per station stays roughly even; where one variant takes materially longer at a single station, that station sets the pace for everything behind it. The honest test is whether a change of variant needs a stoppage. If it does, the schedule is really a series of batches and deserves to be planned as such.

What breaks first when demand for a dedicated line falls?

Overhead recovery, then morale, then discipline. A line designed for a pace absorbs fixed cost through volume, so a sustained drop spreads the same staffing and depreciation thinner. The tempting responses each carry a cost: intermittent running loses the rhythm the model depends on, slowing the line lengthens cycle without cutting labour proportionally, and building ahead converts the problem into finished goods and eventual write-downs.