01 / The order
What is the right starting number?
Start with cash that the business actually expects to keep from an order, excluding tax collected for the government and accounting for discounts. Model expected refunds and return costs separately, or net them from revenue, but never both. A displayed MRP, platform gross merchandise value and money settled to the seller are different numbers. Reconcile the customer receipt to the settlement statement before making a margin claim. If the channel pays weeks later, note the delay as a cash-cycle issue even when the order is profitable on paper.
Keep each channel separate. A sale on the brand's own site can incur payment, shipping and support costs; a marketplace sale adds its fee schedule and may change fulfilment costs; a retail sale has a different trade price, promotions and possible returns. Amazon's India seller page makes the point plainly: fees depend on category, closing fee and shipping or fulfilment arrangement. The relevant figure is the current fee for your actual SKU, not a generic marketplace percentage.
| Line | Ask the finance owner | Treatment |
|---|---|---|
| Net realised revenue | What will the business keep after tax and discount, before expected refunds if they are modelled separately? | Start here |
| Variable product and channel costs | What cost exists because this order exists? | Subtract per order |
| Acquisition cost | Which spend brought this customer, and over what cohort? | Show separately before and after |
| Fixed operating costs | What persists if orders stop? | Model below contribution |
02 / The subtraction
Which costs belong in contribution margin?
Subtract landed product cost, any packaging added for the order, pick and pack, delivery, payment processing, marketplace commissions and expected reverse logistics. Include damaged goods and failed deliveries as an expected cost across the cohort, not only when a complaint arrives. Contribution before acquisition shows whether a sale can support marketing. Contribution after acquisition shows what the first order leaves to pay overhead and fund the next order.
Do not bury customer acquisition in a blended monthly ratio when some orders are organic, some are repeat and some are paid. Track new-customer cohorts by source and first purchase date. A strong repeat rate can support a costly first order, but only if the repeat behaviour is observed and the cash required to reach it is available. Put the acquisition payback horizon beside the reorder interval and inventory payment terms. The rupee example below is hypothetical, excludes tax, and is not an Ospyr or brand result.
| Illustrative line | Rupees per order | Why it matters |
|---|---|---|
| Net realised revenue | 1,000 | Cash expected after discounts, excluding tax; returns are modelled below |
| Landed product and pack | -400 | The unit and its packaging |
| Fulfilment, payment and channel fees | -150 | Costs caused by this order |
| Expected returns and damage | -50 | Average the cost over the order cohort |
| Contribution before acquisition | 400 | Amount available for acquisition and overhead |
| Customer acquisition | -250 | Attribute spend to this customer cohort |
| Contribution after acquisition | 150 | Amount left for fixed costs and future investment |
- Product and inbound freight
- Order pack, pick, delivery and payment
- Platform and promotional charges
- Expected returns, damage and customer support
- Acquisition by channel and cohort
03 / The cash cycle
Why can a profitable order still starve the business?
A D2C founder often pays for production and packaging before a customer pays. The next batch may be due while the first batch is still moving through warehouses, platforms and returns. Write down when cash leaves for material, production, freight and advertising, and when it comes back from each channel. The difference is working capital. It is not solved by a positive gross margin line.
Forecast stock in units as well as rupees. A slow variant ties up cash and shelf life; a fast variant can be unavailable just when marketing has created demand. Test the reorder point against realistic production and transit time, not the best delivery date in a supplier's pitch. If the plan depends on every unit selling at full price, run a second scenario for a slower month, more returns and promotional pressure.
04 / The decision
What should the founder review every week?
The weekly page should show orders, net realised revenue, gross margin, contribution before and after acquisition, return rate, repeat orders and cash tied up in stock. Split the page by SKU, channel and new versus returning customers. Show the number of orders behind each percentage so a small sample does not masquerade as a stable trend. Add a note for the operational cause of a change: damaged packaging, a new fee, stockout or discount, rather than treating the chart as an explanation.
A founder should make one decision per review. Stop a loss-making campaign, improve a pack that drives damage, renegotiate a shipping tier, or reduce an unproven variant. Large consumer companies treat packaging and product design as operating variables; Unilever's packaging research is one reminder that a cheaper material can create different distribution or performance costs. The purpose of unit economics is to make those tradeoffs visible before volume magnifies them.