How is mixed-model production different from batch production?

Batch production makes a long run of one variant, changes over, then runs the next. Mixed-model production interleaves the variants into one repeating sequence, changing over between almost every unit. That keeps finished inventory of every variant low and lets you serve any customer quickly, but it only works when changeovers are nearly free.

How do you calculate a mixed-model sequence?

Total each variant's demand over a level period, reduce it to its simplest ratio (600:400:200 becomes 3:2:1), then lay out a short repeating cycle that holds that ratio, such as A-A-B-A-C-B for a six-unit cycle. Spread each variant out so the same model does not clump, then release the pattern through a heijunka box.

What do you need for mixed-model production to work?

Four things: quick changeover so switching variants is nearly free, balanced work content so heavy and light variants both fit inside takt, reliable parts presentation so operators never hunt for components, and a reasonably level demand signal. A weakness in any one pulls the line back toward big batches.

What is the difference between mixed-model production and heijunka?

Heijunka is the broader idea of leveling production by both volume and product mix. Mixed-model production is heijunka applied to the mix on a single line: it is the actual repeating variant sequence the line builds to hold demand steady across the day.

Does mixed-model production increase changeovers?

Yes, dramatically. A changeover happens between almost every unit rather than once per long run. That is why mixed-model lines depend on SMED quick-changeover to drive setup time down to seconds or a fast fixture swap. Without cheap changeovers the interleaved sequence would waste too much capacity.

How does mixed-model production relate to one-piece flow?

Mixed-model production is one-piece flow extended across variants: one unit of whatever product is next flows through the line, rather than one unit of a single product. It is also the operational face of every-part-every-interval, since running the full variant cycle every day means you can make every part number daily.

Mixed-Model Production Explained

Q: What is mixed-model production?

Mixed-model production makes several product variants on one line, one unit at a time, in a repeating sequence built to match the ratio of customer demand. Instead of a long run of model A followed by a long run of model B, the line runs a short repeating pattern like A-B-A-C that holds the demand mix, so every variant ships within one cycle.

Build several variants on one line in a short repeating sequence, and you match the demand mix without the waste of big batches.

Mixed-model production builds several product variants on one line in a repeating, blended sequence, such as A-B-A-C-A-B-A-C, instead of long single-model batches. The sequence matches output to the demand mix and spreads the workload of every variant evenly across the day.

Run one model all morning and another all afternoon, and the line looks orderly, but it lurches: the first customer waiting for the afternoon model waits half a day, and every shared resource gets slammed then starved as the batch changes over. Mixed-model production is the fix that keeps the variety but throws out the big batches. It is heijunka applied to product mix, and it is one of the harder, higher-value moves in lean manufacturing because it forces the whole line to get good at change.

What Is Mixed-Model Production?

Mixed-model production is the practice of making more than one product variant on the same line, one unit at a time, in a repeating sequence built to match the ratio of customer demand. If customers buy twice as many of model A as model B, the line runs a pattern like A-B-A-A-B-A that holds that two-to-one ratio in a short, repeating cycle rather than a long run of A followed by a long run of B.

The goal is to produce a little of everything, often, instead of a lot of one thing, rarely. That keeps finished inventory of every variant low, shortens the wait for any given model to the length of one cycle, and smooths the load on shared upstream cells and suppliers, because each part is consumed at a steady trickle rather than in a monthly gulp.

How Is Mixed-Model Different From Batch Production?

In batch production you make a long run of one variant, change over, then make a long run of the next. Mixed-model production interleaves the variants into one repeating sequence, so a changeover happens between almost every unit. That only works if the changeover is nearly free, which is why mixed-model lines live or die on quick changeover.

The difference shows up hardest in responsiveness. With batches, the answer to "when can I have a model C?" depends on where model C sits in this month's run. With a mixed-model sequence, a model C rolls off within every cycle, so the honest answer is "within a cycle or two." You trade the comfort of long uninterrupted runs for the ability to serve any customer, any variant, almost any time, with very little finished stock on the shelf.

Both schedules build the same daily count of A, B, and C. The mixed-model sequence holds the demand ratio in a short repeating cycle so no variant's customer waits half a day.

How Do You Build the Mixed-Model Sequence?

Building the sequence is arithmetic first, then engineering. You start from the demand mix, translate it into a repeating pattern, then make the line capable of running that pattern one unit at a time:

Total the demand by variant over a level period. Take a representative week or month and count how many of each variant customers actually pull. Suppose A is 600, B is 400, C is 200 per week.
Reduce the mix to its simplest ratio. 600:400:200 becomes 3:2:1. That ratio is the DNA of your repeating cycle, and tools like the Glenday sieve help decide which runners get leveled first.
Set the cycle and lay out the pattern. A six-unit cycle of A-A-B-A-C-B holds the 3:2:1 ratio. Spread the variants so the same model does not clump, which keeps the workload even.
Load it into a leveling device. A heijunka box or a leveled schedule turns the pattern into physical pull signals released at a steady pace, so the floor builds to the sequence, not to a monthly quota.
Balance the work content of each variant to takt. Because variants differ in labor, use line balancing and clear standard work so a heavy model and a light model still fit inside the same takt time without overburdening the operator.
Present parts for every variant at the point of use. Kit or line-side each variant's components so the operator never leaves the station to hunt parts. Feeding the mix reliably is usually the make-or-break detail.

What Do You Need to Make It Work?

Mixed-model production sits on four capabilities, and a weakness in any one drags you back toward batches. Fast changeover is first: if switching variants costs more than a few seconds of operator motion or a quick fixture swap, the interleaved sequence bleeds capacity. Balanced work content is second, because a variant that takes fifty percent more labor will overburden the line every time it appears unless the tasks are re-divided. Reliable parts presentation is third; the operator has to reach the right components for whatever variant arrives without walking or searching, which ties mixed-model directly to visual management and good line-side design. And a leveled demand signal is fourth: mixed-model smooths a mix that is roughly stable, so a wildly spiky order pattern has to be leveled with heijunka first, or buffered, before the sequence can hold.

Four capabilities carry a mixed-model line, all resting on standard work sized to takt. Weaken any pillar and the line drifts back toward big batches.

What Does Mixed-Model Production Return, and What Does It Cost?

The return is shorter lead time, lower finished-goods inventory, and steadier demand on everything upstream, because you build a little of every variant every cycle instead of a mountain of one. That steadiness is the whole point of leveling: the Lean Enterprise Institute defines heijunka as leveling production "by volume and by product mix" so that a plant can serve varied demand without the waste that lumpy batches create (Lean Enterprise Institute, Heijunka). The cost is discipline: mixed-model production only pays off after you have driven changeover down and balanced the work, which is why it usually follows, rather than precedes, solid SMED and one-piece flow. The two wastes it targets hardest are inventory and overproduction, the two that batch scheduling generates most (Lean Enterprise Institute, The Seven Wastes).

Holding the sequence day after day is where many lines slip. A team that can see, in real time, whether it is building to the leveled pattern or quietly drifting back into a comfortable batch of A can correct within the shift instead of discovering the drift in next month's inventory. That is the practical case for live visibility over the existing line: mixed-model production is a plan the floor has to execute every hour, and it holds far better when the plan and the actual build are visible side by side. Harmony's line is simple here: no rip-and-replace, just a live signal on top of the line you already run, the same idea behind our connected-worker rollout at CLS.

What Are the Common Mistakes in Mixed-Model Production?

Most failures are not exotic. They are the same handful of mistakes, and each one traces back to attacking the sequence before the line was ready for it:

Sequencing before fixing changeover. Interleaving variants when a swap still costs ten minutes turns the schedule into a changeover-eating machine. Drive setup down first, then sequence.
Ignoring uneven work content. If model C carries fifty percent more labor and you slot two Cs back to back, you have manufactured muri overburden, at that station. Spread the heavy variants and rebalance the tasks.
Sequencing to a forecast instead of real pull. Build the pattern from what customers actually consume over a level period, not from an optimistic plan, or the finished inventory the method was meant to shrink piles right back up.
Letting the line quietly re-batch. Under pressure, crews clump the easy variant to "get ahead." Without a visible signal, that drift is invisible until month-end inventory exposes it.

None of these are reasons to avoid mixed-model production. They are the checklist for readiness. A line that has genuinely conquered changeover, balanced its work, and levelled its demand finds the sequence almost anticlimactic to run, which is the sign the hard work was done in the right order.

How Does Mixed-Model Relate to One-Piece Flow and Every-Part-Every-Interval?

Mixed-model production is one-piece flow extended across variants: instead of one unit of one product flowing, one unit of whatever variant is next flows. It is also the operational face of every-part-every-interval (EPEI), the metric that asks how often you can make every part number. A mixed-model line that runs its full variant cycle every day has an EPEI of one day, which is the ambition, and shrinking that interval is what mixed-model production is for. Where the mix is huge and demand is spiky, most plants level the high-volume runners tightly and buffer the rare specials, which is a pragmatic blend rather than a failure of the method.

Think of the progression as three rungs on the same ladder. First you make flow work for a single product, one unit moving cleanly through the steps. Then you drive changeover low enough that switching products costs almost nothing. Then you interleave the products into a leveled sequence, and the line quietly earns the ability to build anything, in the right proportion, every single day. Each rung stands on the one below it, which is why plants that skip straight to mixed-model scheduling usually slide back down to batches within a quarter. Built in order, though, mixed-model production is what lets a plant carry real variety without carrying the inventory and lead time that variety normally drags along.