Exotic variants cost far more than their share of line time. The changeovers, the error risk and the material handling all argue for splitting them out — and the numbers usually agree.
Every high-volume line accumulates them: the export variant, the legacy customer’s special, the version with the different connector. Individually each looks harmless — twenty units a day on a line that runs a thousand. Collectively they are often the most expensive products in the plant, and standard costing hides it, because changeover time and error cost get smeared across the whole line.
Do the arithmetic honestly. A 30-minute changeover before and after a 20-piece run means the line spends more time converting than producing that variant. At a thousand-a-day takt, an hour of changeover is a thousand units of forgone standard product. The variant’s true cost per piece can be several times its calculated cost before a single quality incident is counted.
Quality incidents will be counted. Mixed-model lines fail in a characteristic way: the wrong component from the adjacent bin, the label from the previous run, the test program not switched. These are not operator failures, they are system failures — a line tuned for rhythm being asked to handle exceptions. The 2% variant routinely generates 20% of the containment actions.
The structural answer is segregation. Give low-volume variants their own cell — slower, more flexible, staffed by people whose job is variation rather than rhythm, with its own kitted material supply so nothing shares a bin with the main line. The cell’s cost per piece looks high on paper. It is honest, which the smeared number never was.
This is where an external partner earns a place. Building a variant cell in-house means dedicating floor space and people to your least attractive volumes. Placing those variants with a manufacturer built around changeovers — kitted materials, flexible test rigs, batch traceability — often costs less than the internal cell and frees the main line completely.
The test we suggest to plant managers: list every product that runs less than one hour per day on your main line, and assign each one its real changeover and containment history. The list is usually short and the numbers are usually decisive. The thousand-a-day line got fast by doing one thing. Let it.
— GANI Manufacturing engineering team