---
name: orphex-product-profitability-review
description: "Review product advertising economics using aligned net revenue, returns, and explicit variable costs; distinguish contribution from gross ROAS and company profit."
license: MIT
metadata:
  version: "2.0.0"
---

# Orphex Product Profitability Review

Assess product-level contribution under the user's supplied cost and revenue definitions. Do not equate attributed gross revenue with profit or assume a universal margin.

## Reconcile economic scope

Align order/acquisition cohort, product/variant, currency, tax/shipping convention, returns maturity, discounts, refunded revenue, recognized/recovered COGS, fulfilment, payment fees and advertising allocation. Identify whether revenue is already net of discounts/refunds and whether COGS already includes other expenses. Subtract each economic cost once. Keep unidentified shared advertising/overhead separate unless a documented allocation is supplied; do not distribute it arbitrarily by attractive ROAS.

Use totals for the same scope and horizon. Under the explicit component contract: net revenue = gross revenue − discounts − refunds; contribution before ads = net revenue − distinct COGS − fulfilment − payment fees; contribution after ads = contribution before ads − allocated ad spend. Negative net revenue or contribution can be real; report it with its definition. A missing cost is unknown, not zero, and prevents a complete contribution claim. A gross-revenue field already net of refunds must be normalized before using the helper, with the transformation documented.

Calculate gross and net revenue ROAS with their named numerators. Contribution break-even net-revenue ROAS = net revenue / pre-ad contribution only for positive net revenue and positive pre-ad contribution under stable observed cost mix. Zero/negative contribution does not have a meaningful positive advertising break-even ratio. Missing/zero ad spend makes ROAS unavailable. These are descriptive cost identities, not marginal response forecasts or target bids.

## Recommend economic checks

Rank material adverse contribution, uncertain allocations, immature returns and cost gaps. Identify whether a recommendation concerns ad allocation, merchandising, stock, pricing, returns or feed health; do not prescribe an unrequested price change. Repeat-customer value needs a separate supported cohort analysis. Report the table, explicit cost inclusions, missing overhead/future value, and a bounded next check. No company net-profit or causal incrementality claim follows from allocated contribution alone.

## Portable inputs and examples

- Read [the input contract](references/input-contract.md) when mapping a new export or checking the example's scope and definitions. Copy [the header-only CSV template](assets/input-template.csv) when preparing data; equivalent supplied exports remain acceptable.
- Read [the reusable business context](references/business-context.md) only for business facts or constraints this task needs. Reuse user-supplied facts with their source/date; the template contains no default targets.
- Inspect [the complete fictional input](assets/example-input.csv) with [its example output](references/example-output.md) when learning the output and calculation boundaries. Never use fictional values for a real account.

State whether the result is complete, partial, or blocked for the requested decision. Link material findings to actual supplied rows/sources and separate observed metrics, hypotheses, and estimates. Lead with a short business conclusion, then evidence, uncertainty, and the next measurable check. A data export or installed skill does not authorize account changes.

For the supported arithmetic only, optionally run [the bundled calculator](scripts/marketing_math.py) with Python 3: `python3 scripts/marketing_math.py profitability < calculation.json`. Read its input mapping in the input contract before preparing JSON. It reads JSON, not CSV directly. If Python or the requested method is unavailable, show a reproducible alternative calculation or mark it unsupported; do not report an uncomputed result as verified.