Consumer Goods: Sequencing Lines to Cut Changeover Waste

Consumer goods changeover sequencing has to model a cost that depends on what ran before, then order the run so that cost is as small as possible. A filling line that goes from a light fragrance to a heavy one pays a small change; going the other way needs a full clean. EDGEBIC by User Solutions prices each transition per line in a setup matrix, then lets its optimizer reorder the queue to minimize total changeover while protecting your due dates.

For the matrix mechanism on its own, see the setup matrix explained. For the optimizer that does the reordering, see mathematical schedule optimization. For the sector view, see consumer goods production scheduling, and pair this with seasonal capacity for consumer goods. The full map is at how different industries use EDGEBIC.

The changeover asymmetry consumer goods live with

Packaged goods lines run families of similar products that differ in one dimension: scent, flavor, color, formula. Switching between them costs time, and the cost is rarely symmetric. Unscented to a light fragrance is a rinse. A heavy fragrance back to unscented is a full clean-in-place cycle to avoid carryover a customer would notice. Clear to a deep color, then back, is the same story. Light to dark is cheap; dark to light is expensive.

A schedule that stores one average setup number gets both directions wrong. It underestimates the expensive reverse change, so the plan promises a finish the line cannot hit, and it overestimates the cheap same-family change, so the plan hides capacity the line actually has. Neither error is visible until the floor runs behind or idle, and by then the shift is spent.

The fix is two steps. First, price the real cost of each transition. Second, order the run so the expensive transitions happen as rarely as possible. A personal care plant lives at the sharp end of that asymmetry, which is why batch scheduling and changeovers for cosmetics treats fragrance and color order as the main lever it has.

Step one: price every transition in a matrix

EDGEBIC stores changeover as a lookup keyed by the line and by the pair of products involved. You keep it maintainable with families.

Setup families group products that change over identically on a line. For consumer goods the useful grouping is almost always the changeover dimension itself: for a fragrance line, something like Unscented, Light, Medium, Heavy, and a Specials group for anything with a dye or an allergen that forces an extra clean. Eight families covering a hundred SKUs turn a ten-thousand-cell product grid into 64 cells.

The family matrix is the per-line grid of from-family to to-family minutes, where roughly 95 percent of behavior lives. Product-level overrides cover the handful of SKUs that break their family rule, such as one pigment that fights every light color after it, and an override always wins over the family cell.

One rule matters more than any other on your first load: a matched cell replaces the routing's flat setup rather than adding to it. Your cells hold the total changeover, teardown and clean included. If the mechanical swap is 20 minutes and the clean is 40, the cell holds 60. Same product back to back returns zero automatically, so a repeat run on a line already set carries no changeover.

Each line gets its own grid, because a clean-in-place filler and an older manual line do not change over in the same time. The same scent change can be 15 minutes on one and 60 on the other, and the scheduler charges whichever line the job lands on. The concept on its own is covered in what a setup family is.

Step two: let the optimizer order the run

Once the matrix is loaded, the scheduler knows the real cost of every ordering. Now it can do something a whiteboard cannot: search for the order that minimizes total changeover.

EDGEBIC's optimizer works on goal presets, each a ranked list of what to protect. The least-setup preset ranks total setup hours first, then due dates, then overall finish. When you run it on a line's queue, it reorders the jobs to group compatible products and step through the changeover dimension in one direction, so a run that bounced between light and heavy fragrances becomes a run that sweeps light to heavy once.

Two properties make this safe to use. The optimizer proposes; nothing is committed until you accept it, so you see the reordered plan and its numbers before it touches the schedule. And the result protects the priorities you set: it is guaranteed not to be worse than the plan you started with on the goals you ranked. The least-setup preset becomes available once your line has a populated setup matrix, because there is nothing to optimize against until the transitions are priced. The mechanism at the sector level is covered in campaign sequencing to minimize changeovers.

A worked line grid

A personal-care filling line, clean-in-place, mechanical swap 20 minutes. Five families. Values are total changeover minutes including the swap and any clean.

Read the asymmetry. Unscented to Heavy is 35 minutes. Heavy to Unscented is 90. The diagonal is the mechanical swap alone, because a same-family run needs no clean. A queue that runs Heavy, Unscented, Medium, Heavy pays 90 plus 30 plus 25 across three changes. Reordered by the optimizer to Unscented, Medium, Heavy, Heavy, it pays 30 plus 25 plus 20, and two of those SKUs never triggered a clean at all.

Twenty-five cells cover every combination of every SKU that runs on this line, however many there are. A second line with manual teardown gets its own grid with a longer diagonal and heavier clean times, same products, different numbers.

What honest data plus sequencing does

The published worked case uses a paint booth, and a filling line is the same shape. Three jobs, one machine, one 8-hour shift. With a flat 30-minute setup the plan showed an early finish while the floor needed far longer, because the expensive reverse changeovers were never charged. With the matrix loaded and a due-date order, the plan honestly charged 330 minutes of changeover and showed the work overflowing the shift. With the matrix loaded and the jobs campaigned like to like, changeover fell to 90 minutes and all three finished with hours to spare: a 73 percent reduction from sequencing alone.

That is the two-step benefit. The matrix buys honesty. Sequencing buys back capacity. A scheduler with no matrix has no basis for preferring one order over another, so the optimizer step is unavailable until the matrix is done.

Common mistakes in the first week

Only one product in a pair has a family. The family lookup needs both sides. Assign one and leave the other unassigned and the engine falls through to the flat routing default with no error. This is the most common cause of "the matrix is not working."

Cells hold only the clean. A matched cell replaces the flat setup, so a cell holding 40 minutes of clean quietly deletes the 20-minute mechanical swap from the plan. Cells hold the total.

Running the least-setup goal with an empty matrix. The preset needs priced transitions to have anything to optimize; it stays out of the picture until the line's matrix exists. Build the matrix first.

Trusting the reorder without checking due dates. The optimizer protects due dates, but you should still read the proposal before accepting it, because it is showing you a real trade you may want to adjust. Accept nothing you have not looked at.

Rolling it out

1. Pick your two most contended lines. The lines that decide your ship dates are where the payback is.
2. Define four or five families by the changeover dimension: scent, flavor, or color band, not customer.
3. Assign every SKU that runs on those lines to a family; anything genuinely unique gets a product-level override.
4. Fill each line's grid with total changeover including teardown and clean, from the people who do the changeovers.
5. Run the schedule and confirm the plan charges the real changeover, using the setup-source detail to check which rule fired.
6. Run the least-setup goal on a busy queue, read the reordered proposal, and accept it once the trade looks right.

Where the clean is validated rather than merely expensive, the same matrix carries more weight: scheduling pharmaceutical cleaning changeovers honestly covers that case.

Bring one line, its real changeover times and a week of orders to a demo of consumer goods production planning, and we will build the grid and run the optimizer with you.