MARGIN

Contribution margin for e-commerce

Your dashboard says revenue. It doesn't say whether you made money.

MARGIN computes daily contribution margin from your real costs — and tells you which of those costs you can actually trust. Then it hands the numbers to your AI, so you can just ask.

Read-only. Nothing we do can write to your shop.

A real shop, three days

EUR

Revenue

€15,105

Contribution margin

€675

Profit

−€1,308

2026-08-21 €3,961 −€252
2026-08-20 €5,339 −€300
2026-08-19 €5,805 −€756

Profitable-looking revenue, three losing days. The gap is cost of goods, fulfilment and ad spend — the columns a revenue dashboard does not have.

01

Costs you can verify

Every margin figure carries how much of it we could actually check. A margin computed on 40% cost coverage is not a margin, and we say so instead of printing a clean number.

02

Ratios that survive a sum

A period total sums the amounts and re-derives every ratio from them. Averaging ROAS across days is the classic way to be confidently wrong; here it is impossible by construction.

03

Two margins deeper

Contribution margin excludes fixed cost. Profit includes it. Both ship, with every component that went into them, per storefront and per day.

The daily P&L

A ledger, not a chart

One row per day, twenty-eight columns in three groups, and the figures your accountant would recognise. The period total is rendered by the same code as a daily row, so the two cannot drift apart.

Date Revenue Orders COGS Ad spend PNO Contrib. margin Profit
Three days €15,105 508 €4,123 €8,522 56% €675 −€1,308
2026-08-21 €3,961 125 €1,127 €1,989 50% €409 −€252
2026-08-20 €5,339 178 €1,460 €2,883 54% €361 −€300
2026-08-19 €5,805 205 €1,536 €3,650 63% −€95 −€756

Real days, anonymised. PNO on the total row is 8,522 / 15,105 — derived from the sums, not the average of the three daily percentages, which would read 55.7%.

Ask your AI

Theirs will answer. Ours will be right.

Point an AI client at a shop database and it writes confident, wrong SQL — cancelled orders, exchange rates, discounts as negative lines, zero-priced bundle rows. None of it visible from the column names.

Our tools have those decisions already made, once, in tested code. Every answer carries what not to trust about it, including how much of your cost data we could verify.

your AI client

$ claude mcp add --transport http margin https://…/mcp/margin

> what was my contribution margin last week?

€4,182 across 1,140 orders, after €9,610 of ad spend. Cost data covers 99.7% of goods sold. Fixed costs are excluded — profit was €248.

Pricing

One number, once you're connected

Profit

Paid

Once your orders and costs are connected.

  • Daily contribution margin and profit, per storefront
  • AOV, CAC, PNO, acquisition rate, new versus returning
  • ROAS and POAS with spend by channel
  • Cost coverage on every margin figure
  • Per-SKU profit wherever your platform can support it
  • The full tool set for your AI client, with an audit log

Find out whether yesterday made money

Connect a shop and the daily P&L fills itself in.