Toll manufacturing: selling conversion capacity without owning the material

The invoice is for capacity and yield, never for the goods

Pricing a toll job means pricing time on a specific asset plus an agreed loss allowance. Margin comes from throughput and from keeping conversion loss inside that allowance, so an hour lost to a slow changeover cannot be recovered by selling the output at a better price — there is no output to sell. The materials that suit the arrangement are the ones customers refuse to let go of: volatile-priced feedstocks, proprietary formulations, actives whose regulatory registration sits with the owner. The commitment is real capacity on named equipment, often with a minimum booking, which is why converters treat the asset calendar as the contract.

Customer-owned material rewrites the stockroom

Segregation starts at goods-in. Each owner's material needs its own bins, its own labels and a ledger that survives an audit, because the converter holds stock it cannot value on its own balance sheet. Work in progress carries the same tag through every stage, and finished output waits for a release instruction rather than a sales order. Reconciliation becomes a monthly discipline: quantity received, quantity converted, quantity returned, quantity written off as process loss, with residues and heels accounted for line by line. Plants that treat this as paperwork discover the gap only when a customer counts.

Equipment specificity decides who gets asked to quote

Toll work is won on the asset list. A customer looking for a spray dryer with a particular chamber geometry, a reactor in a given material of construction, or a mill able to hold a fine cut is choosing between a handful of plants that already own that kit. The capital argument is therefore about breadth and cleanability rather than raw scale: shared assets need validated cleaning between campaigns, changeover parts, and utilities sized for the worst case. Dedicating a line to one customer raises the fee but hands them a claim on your capacity, and the negotiation over who funds any modification is where deals usually stall.

Yield arguments become commercial arguments

Two quantities decide whether a campaign was profitable and both are contested: how much material went in and how much saleable product came back. Analytical methods, sampling points and acceptance criteria have to be agreed before the first drum is opened, because a dispute afterwards is a dispute about money. Cross-contamination is the other exposure — a trace of the previous campaign in a customer's product can write off a batch worth far more than the conversion fee, which is why cleaning verification and campaign sequencing sit near the top of a converter's risk register. Insurance rarely covers the full replacement value of goods held for others.

Booking a plant against demand you cannot forecast

Growth comes from adding customers to the same assets, and it stops when the calendar is full or when campaigns collide. Since the converter sees no end-market signal, scheduling depends on customer call-offs, take-or-pay minimums and rolling booking windows. Execution systems have to hold ownership on every lot, price by activity rather than by item, and produce a material balance per campaign; ordinary order-to-cash logic does not fit. Purchasing narrows to consumables, filter media, packaging, spares and energy, which removes most buying leverage but also removes commodity price risk from the operation entirely.

Frequently asked questions

Who carries the loss if customer-supplied material is damaged in our plant?

Liability follows the contract, not the ownership. Material belongs to the customer throughout, but the converter is usually a bailee and answers for negligence, so the agreement should state an agreed process-loss allowance, the point at which losses become chargeable, and whether liability is capped at the conversion fee or extends to replacement value. Goods-in-trust cover is a separate insurance line and is often overlooked until a spill makes it urgent.

How should a conversion fee be built up?

Start from the asset rather than the product: occupied hours on the constraint equipment, energy and utilities at the load the process actually draws, direct labour, cleaning and changeover time, consumables, quality-control testing, and a contribution to fixed overhead. Add an explicit loss allowance and a separate line for waste disposal, which can dominate on solvent-heavy work. Fees expressed per unit of output shift yield risk onto the converter; fees per hour or per batch leave it with the owner.

Does tolling make our quality system easier because we do not own the product?

It makes it wider. The converter still holds the manufacturing record, the deviation history and the release testing for the operation performed, and customers audit that documentation because their own regulatory file depends on it. Add segregation controls, cleaning validation between campaigns and a material balance for every job, and the documentation load per unit of revenue is usually heavier than it would be for own-brand production.