Ask a food manufacturer what their best product is and you will usually get the highest-revenue one. Ask for contribution margin per unit of constrained capacity and the conversation goes quiet. That gap is where a lot of margin lives.
Build the cost from physical reality
A defensible unit cost has five components: raw material at actual yield, direct packing material, direct conversion cost, variable logistics, and channel-specific selling cost. Fixed overhead is deliberately excluded, because contribution — not fully absorbed cost — is what should drive product and channel decisions.
The component most often wrong is the first. Raw material cost at standard recipe is not raw material cost at actual yield. A 3% yield loss that nobody measures shows up as an unexplained gap between theoretical and actual gross margin, and at scale it is frequently the entire profit.
The yield loss nobody books
Peeling, trimming, evaporation, line spillage, rework and returns all consume raw material that produces no saleable output. In fruit and vegetable processing these routinely total 5–9% before anyone measures them; in dairy value addition, 2–4%.
The fix is a mass balance per batch, which requires weighing inputs and outputs rather than inferring them. One spice processor found 7.1% loss where the team assumed 4% — worth ₹1.4 crore a year at their volumes, and invisible in a monthly P&L.
Cost to serve by channel
Charging these costs to the channel that causes them changes decisions immediately. Institutional supply with a 12% margin frequently out-earns modern trade at 28%, because the listing fees, visibility charges and working capital cost of extended terms never get charged to modern trade in most MSME costing.
| Channel | Typical trade margin | Other costs to charge | Net realisation |
|---|---|---|---|
| General trade (distributor) | 18 – 26% | Schemes, damages, credit cost | Moderate |
| Modern trade | 22 – 32% | Listing, visibility, 60–90 day terms | Lower than it looks |
| Institutional / HoReCa | 8 – 15% | Credit cost, customisation | Often the best |
| D2C | 0% | CAC, shipping, returns, payment gateway | Highly AOV-dependent |
| Export | 10 – 18% | Certification, documentation, freight | Good at scale |
Contribution per unit of constraint
When capacity is the binding constraint, the right metric is contribution per hour of the constrained resource, not contribution per unit. A product with a lower margin per kilogram can be the better product if it runs twice as fast through the bottleneck.
This single reframing has changed the product mix on roughly half of our operations engagements, usually by revealing that the SKU the sales team pushes hardest is the one that occupies the bottleneck for the least return.
When capacity binds, the product that earns most per kilogram is not necessarily the product that earns most per hour.
Doing this without an ERP
You do not need a system to start. A single spreadsheet with your ten largest SKUs, actual yields measured over two weeks, real packing material costs from invoices, and channel costs pulled from your own credit notes will get you 80% of the value.
What the system buys you later is currency — the difference between a number that was true in March and a number that is true this week. Start with the spreadsheet; the discipline of maintaining it is what justifies the system.
Frequently asked questions
A note on the numbers in this article
Benchmarks come from engagements we have delivered and are indicative at 2025–26 prices. Scheme details, rates and eligibility change with each policy cycle — verify the current position before you commit capital. Nothing here is legal, tax or investment advice.