Co-Manufacturer vs. Co-Packer: The Difference Is the Recipe

Two quotes for the same jar of hot sauce. Under the toll quote, the one Cinderhaven Provisions runs on today, it buys its own peppers, vinegar, and glass at $44.60 a case and pays the plant $17.80 for filling: $62.40 all in. Under the turnkey quote, the plant buys everything and delivers a finished case at $57.60. Across the 45,000-case line, turnkey is $216,000 a year cheaper. It is also the only one of the two quotes that comes with a lock on the door. Cinderhaven Provisions is a fictional company and both quotes are a synthetic dataset; the trade they illustrate, price against custody, is the real difference between a co-packer and a co-manufacturer.

The vendor glossaries will not settle which name applies. The industry uses co-packer, co-manufacturer, and co-man loosely, and half the plants in the category answer to all three. What settles it is the contract, and the contract turns on one operational fact with three consequences hanging off it.

Ask who buys the peppers

Strip the terminology and one question separates the two models: who issues the purchase orders for ingredients. In a toll arrangement, the brand does. It owns the formula, sources the inputs, holds the supplier relationships, and pays the plant a per-case fee to run them through the line. In a turnkey arrangement, the plant procures against the brand's specification, batches at its own scale, and sells back finished cases. Same kettles, same filler, different org chart.

The $4.80-a-case spread in Cinderhaven's quotes is what turnkey procurement is worth: the plant buys pepper mash and glass across every client it runs, at volumes a $25M brand cannot match alone. The direction of the spread varies by category and by year. What does not vary is what rides along with the purchasing. When the plant buys the inputs, the plant holds the supplier list, the ingredient specifications, and the certificates of analysis. The brand holds an invoice for finished goods.

Ask who buys the peppers. The answer is the org chart.

The recipe that runs is not the recipe you wrote

Formulas do not survive contact with a production line unchanged. The brand's recipe says 12% mash; the line runs at 12.4% because that is where the fill viscosity holds. Cook times shift a batch at a time until the scorching stops. Two years in, the document that describes what is in the jar is the plant's batch sheet, versioned by their QA team, and the brand's founding recipe is a historical artifact.

Under a toll structure, those revisions are the brand's property being maintained by a contractor, and a quality agreement that says so makes it enforceable. Under turnkey, the default drifts the other way: process knowledge accumulates as the plant's know-how, and a turnkey agreement can assign process parameters developed in production to the manufacturer outright. Every downstream fact the brand publishes, the nutritionals on the label, the master data fields the retailers key on, derives from documents it may not hold.

The recipe you own is version one. The jar ships version seven, and version seven lives at the plant.

The FDA files plants, not recipes

For Cinderhaven's category the custody question has a regulatory floor. Cinderhaven's hot-fill salsas, low-acid vegetables brought under pH 4.6 with added acid, are acidified foods, and FDA requires the commercial processor to register each establishment and file a scheduled process for each acidified food in each container size, on Form FDA 2541e, listing the foods processed in each establishment. The registration numbers the industry calls FCE and SID attach to the processing plant. They are not the brand's to take.

That reshapes what switching costs. Even when Cinderhaven owns its formula outright, moving the salsa line to a new plant means the new establishment's registration, new scheduled process filings for every jar size, and a process authority validating the thermal process on the new line's equipment. If the incumbent's undocumented adjustments are what made the process work, the brand is not transferring a recipe. It is reconstructing one, while the traceability records FSMA 204 wants restart under a new roof.

The FDA does not file recipes. It files plants.

Price the exit before you sign the entry

The two quotes are two custody structures, and the differences line up:

Read the last row against the first paragraph. The $216,000 discount is real money in year one. It is also the co-manufacturer's renewal-day arithmetic: a buyer who must reconstruct a formula, refile with FDA, and revalidate a kill step is a buyer who accepts the price increase. The discount and the moat are the same clause viewed from opposite ends of the contract.

None of this argues for tolling. It argues for pricing the exit while both quotes are still on the table, because that is the last day the price of leaving is zero.

Put the custody question in writing first

Forward me the draft agreement before anyone signs it, along with one current batch sheet if you are already in production. I will send back a custody map, who holds the formula, the filings, the supplier list, and the process knowledge as written, plus the clauses to change while you still hold the pen. The draft costs nothing to check. The renewal letter is more expensive.