Why Co-Manufacturers Lose Money on Changeovers They Quoted Correctly

The quote was right and the job still lost money

You ran the numbers on a new customer's line time. Your operators hit the changeover you promised, give or take ten minutes. Quality signed off. The customer paid on time. And when the monthly close came through, that account was underwater, and the person who quoted it cannot explain why.

Every result you will find on this problem sells you a faster changeover. Quick changeover programs, tool-less fittings, SMED workshops, a scheduling module that sequences light to dark. Some of those are worth buying. None of them address the reason the job lost money, because the money did not leak in the changeover. It leaked in the definition.

Two numbers that share a name

A co-manufacturer quotes changeover as clock time on the line: last case of product A off, line reconfigured, first case of product B on. That number is real, your operators can see it, and it is the one your customers ask about.

The number that decides whether the job is profitable is different. It is time to first saleable case. Equipment vendor INSITE by Douglas puts it plainly: changeover is "the total elapsed time from the moment you stop running one product until the line is producing good product again, at full rate," and it includes "shutdown, cleanout, adjustments, material loading, testing, and anything else that inhibits your machine from running full production again." The same source says the single most important question to ask about any quoted changeover speed is whether it includes the time to first good product.

That question is aimed at machine buyers. Turn it around and aim it at your own quoting process. Between the clock-time changeover and the first saleable case, a co-man on a food or beverage line typically absorbs:

• Cleanout and sanitation, including any hold time the cleaning procedure requires before the line can be released.
• Allergen verification, where the line waits on a swab result before anyone is allowed to start.
• First-article approval, where the first cases are pulled, checked against spec, and either released or scrapped while the line idles or runs at reduced rate.
• Startup scrap, the product made while fills, seals, code dates and weights settle in. The customer does not pay for this, and on a short run it is not a rounding error.

Quote the first number, incur the second, and the job is unprofitable at exactly the volumes your salesperson was told to chase. Short runs, frequent switches, new customers with unfamiliar formats. The accounts that look like growth on the sales board are the ones where the gap between the two numbers matters most.

Why SMED cannot compress the half that costs you

A quick changeover program is a good thing. We are not arguing against it. We are arguing about where it reaches.

SMED attacks the mechanical portion of a changeover: fittings, adjustments, staging, operator motion. On a food line, a meaningful portion of the time between last good and first good is not mechanical. An allergen changeover is governed by a validated cleaning procedure, one that has been proven through worst-case testing to remove residue consistently. That procedure defines the steps, the chemistry, the contact times, and the verification swabs. It does not care how fast your operator is. A faster operator does not shorten a mandated contact time or make a lateral flow test return in fewer minutes.

The same is true of first-article approval. It exists because the customer's spec says it does, and the line waits on quality, not on an operator.

So the portion of the changeover that a lean project can compress is often the smaller half of the real cost, and the larger half is fixed by food safety and by the customer's own requirements. A co-man that buys a SMED program to fix an unprofitable account is speeding up the part it already had under control and leaving the part that actually bled untouched.

How the gap gets into the quote

None of this happens because someone is bad at their job. It happens because the quote is built from the number the plant measures, and the plant measures what it can see.

The changeover log records line stop to line start, because that is when the operators are standing at the machine. Sanitation is logged by a different team on a different sheet. Startup scrap goes into a waste figure that is averaged across the whole run rather than attributed to the first twenty minutes. First-article delays show up nowhere at all, or as "quality hold" in a downtime code that is never cross-referenced to a customer.

By the time these reach the person quoting the next job, they have been averaged, reclassified, or dropped. The quoter is not being optimistic. They are quoting the only changeover number that exists, and it is the wrong one.

We wrote about a version of this pattern in exception handling as manufacturing's blind spot. The cost that never gets a line item is the cost that never gets managed. Changeover-to-saleable is that line item for a co-manufacturer.

What the brand side sees, and why it makes this worse

Your customers have been coached to negotiate exactly the fees that would have covered this gap. Brand-side contract guidance now circulates with template numbers: a small flat fee for a simple pack change, a larger one for a moderate retool, a five-figure fee plus yield guarantees for a major process change. The same guidance tells brands to push a co-man's scrap allowance down from the 5 to 8 percent commonly requested toward 2 to 4 percent, and to have scrap reimbursed at fully landed ingredient cost.

None of that is unreasonable from the brand's chair. It does mean your customers arrive with a clear picture of which fees to resist, and no picture at all of what your changeover actually costs you. If your quote does not carry the sanitation hold, the first-article wait and the startup scrap as explicit items with a defensible basis, they get negotiated away by default, because nobody on either side can see them.

A brand that has just landed its first national account is under its own pressure, and we have described that from their side in what a DTC brand has to automate before its first big retail PO. They will squeeze on run size and on switch frequency. That squeeze lands on you at the point where the two definitions of changeover diverge the most.

The fix is a scheduling and quoting change, not a purchase

The obvious response is a system. A scheduling module that sequences by allergen and pack format, a quoting tool with a changeover matrix, an MES that timestamps every transition. Vendors will happily sell you all three, and there are plants where one of them is the right next step.

But the sequencing matters, and a tool bought before the definition is fixed will encode the wrong number faster. That is the same trap we describe in the automation ROI math vendors show you, recalculated: the software does not know which changeover it is measuring, and it will report whichever one you feed it with a great deal of confidence.

What has to change first is smaller and less comfortable:

• One definition of changeover across the plant, running from last saleable case to first saleable case, owned by one person and used by quoting, scheduling and finance alike.
• A changeover matrix that reflects the real transitions, not the mechanical ones. Allergen to non-allergen is a different row from format change on the same formula, and the difference is measured in hours, not minutes.
• A quote that prices short runs as short runs. Setup time is mostly fixed, and fixed time is cheap per unit only when the run is long. A minimum run fee or a true-up charge is a normal instrument in this market and brands increasingly expect to see one.
• Startup scrap attributed to the changeover rather than averaged across the run, so the quote for a new customer's format carries its own scrap estimate rather than the plant average.

Where those lines sit for your plant, which transitions dominate your calendar, and how far your current quotes are from your real cost is a different question for every co-man, and it is the question that decides whether this is a spreadsheet fix or a software purchase. We do not sell the software, which is why we are willing to say that most plants we look at need the spreadsheet first.

The number to go and check this week

Take the three accounts with the shortest average run length and the most frequent switches. For each, pull the quoted changeover and then reconstruct, from sanitation logs, quality holds and scrap records, the elapsed time from last saleable case to first saleable case. The gap between those two numbers, multiplied by your line cost per hour, multiplied by switches per month, is what those accounts are costing you that nobody quoted.

If that figure is small, your quoting is fine and a quick changeover program may be a good use of the year. If it is not, you have found the leak, and it was never in the changeover. It was in what you agreed to call one.

If you would like a second pair of eyes on which changeovers are actually paying their way, start a conversation. We will bring the questions, and we will tell you if the answer is a spreadsheet rather than a system.