Short answer: contract brewing and co-packing sell capacity, not a brand — so the whole business runs on two numbers: how full the plant is, and whether each customer contract actually makes money after every loss. The analytics that matter are capacity utilisation (fermentation tank-days and line hours), changeover and CIP time, yield per recipe, and true cost-to-serve and margin per contract. AI sharpens the schedule and the costing; it does not change the fundamentals.
A contract brewery’s product is its brewhouse, fermentation tanks and packaging lines — sold by the tank-day and the line hour. That makes it a different analytical animal from a brand-owning brewery: success is not depletions, it is a full plant of profitable contracts. The same lens applies whether you are the co-packer or you use one. It builds on cost of goods per hectolitre and cost-to-serve profitability.
Capacity is the product
The binding constraint in contract brewing is rarely the brewhouse — it is fermentation and conditioning tank-days, and packaging-line hours. A tank tied up an extra two days for a slow-attenuating recipe is capacity you cannot sell. So the first analytics job is utilisation: how many of your available tank-days and line hours are actually billed, and where they leak to changeovers, CIP between brands, and idle gaps. Scheduling analytics — sequencing brands to minimise CIP and matching recipe fermentation profiles to tank turns — directly create sellable capacity, the theme of production scheduling.
True cost and margin per contract
A contract looks profitable on the quote and loses money in reality when the losses are not attributed. Real cost-to-serve per customer has to carry that brand’s actual brewhouse yield and extract efficiency, its fermentation losses and tank time, its changeover and CIP burden, its packaging-line OEE and material waste, and its share of utilities and overhead. Only then does margin per contract become honest — and it routinely reveals that a high-volume customer paying a low rate is worth less than a smaller, simpler one. Attributing yield per recipe (gravity, attenuation and brewhouse efficiency vary by brand) is the unglamorous core of the costing.
Traceability and spec control across many brands
Running several customers’ brands through shared vessels and lines makes traceability and quality non-negotiable. Each brand is a distinct recipe and specification, and the IBD fundamentals still govern: hit the customer’s target original gravity and attenuation, their colour and bitterness, their dissolved-oxygen and carbonation specs — on their recipe, not a house default. Batch genealogy (lot to tank to package to pallet, per customer) protects against mix-ups, supports each brand’s recalls independently, and underpins the billing. Shared lines also raise allergen and cross-contamination control to a first-order analytics concern, not an afterthought.
Where AI helps — and where it doesn’t
Machine learning sharpens the two hard problems: scheduling (sequencing brands and tank turns to maximise billed tank-days and minimise CIP) and forecasting (predicting a brand’s fermentation time and yield so the schedule is realistic), with packaging-line OEE analytics squeezing the line side. A generative-AI copilot can draft the per-contract margin review and the customer-facing production report from the data.
Three honest limits. First, garbage costing in, garbage margin out — if you do not attribute yield, tank-time and CIP to the right brand, the per-contract margin is fiction, and AI just computes the fiction faster. Second, capacity is lumpy and contractual — you cannot smoothly optimise around minimum-order commitments and a customer’s fixed launch date, so the schedule has hard human constraints. Third, quality is the contract — a co-packer’s reputation is hitting every brand’s spec every time; no utilisation gain is worth an off-spec batch that loses the customer.
The bottom line
Contract brewing and co-packing are a capacity business wearing a brewery’s clothes. Measure utilisation in tank-days and line hours, attribute every loss to the brand that caused it, and compute honest margin per contract — then let scheduling and forecasting analytics keep the plant full and the costing tight. Keep traceability and spec control absolute, because the product you actually sell is reliable capacity.
Frequently asked questions
What is contract brewing and co-packing? Contract brewing is making beer for another company’s brand; co-packing (co-manufacturing) is packaging it — canning, bottling or kegging — for them. In both, the producer sells capacity and execution rather than its own brand, so the business is run on tank-days, line hours and cost per contract, not on depletions.
What metrics matter most in contract brewing? Capacity utilisation (fermentation tank-days and packaging-line hours are the real constraints), changeover and CIP time between brands, yield and brewhouse efficiency per recipe, true cost-to-serve per contract, and margin per customer after every loss. Shipments tell you little; utilisation and margin per contract tell you everything.
How does analytics help a contract brewery? It answers the two questions the business turns on: is the plant full, and does each contract make money? Scheduling analytics maximise tank-days and minimise changeovers; cost-to-serve and yield analytics expose unprofitable contracts; and batch genealogy keeps every customer’s brand traceable and on-spec across shared lines.
Part of the Brewing Science & AI track.