# Returns-adjusted cohort review

By George Kelly · Published September 5, 2026
Guide: https://www.iamgeorgekelly.com/field-guide/returns-adjusted-contribution

An original management-analysis template. The example is fictional. It is not an accounting statement or a bid recommendation. All money uses one currency; sales tax and fixed overhead are excluded.

## Define the cohort

- Order-date start/end and timezone: [fill]
- As-of date: [fill]
- Currency and channels: [fill]
- Included order IDs / secure source location: [fill; keep customer data private]
- Status: [provisional / mature under the stated return-window rule]
- Return-window rule and unresolved cases: [fill]
- Spend allocation method and scope: [fill; allocate each cost once]
- Source exports, versions, and reviewer: [fill]

## Reconcile the figures

| Line | Your cohort | Illustrative USD cohort | Evidence / exclusions |
| --- | --- | --- | --- |
| A. Revenue after discounts, before refunds | [fill] | 10,000 | 100 orders; no shipping revenue or tax |
| B. Actual refunds assigned to these orders | [fill] | 2,000 | 20 fully refunded orders |
| C. Net revenue = A − B | [calculate] | 8,000 | Do not subtract refunds again if input was already net |
| D. Original product cost | [fill] | 5,500 | Includes all originally shipped units |
| E. Inventory cost actually recovered | [fill] | 1,280 | 16 resalable returns × 80; four unsalable returns recover zero |
| F. Net product cost = D − E | [calculate] | 4,220 | Retained cost already includes unsalable units |
| G. Original variable costs, net of actual credits | [fill] | 1,300 | Shipping, payment, fulfillment; none credited in this example |
| H. Additional return costs | [fill] | 200 | Transport and handling; exclude costs already in G |
| I. Advertising assigned once | [fill] | 2,000 | Same cohort/scope |
| J. Contribution after ads = C − F − G − H − I | [calculate] | 280 | Before fixed overhead |
| K. Additional expected net return cost | [fill separately] | 275 | Sensitivity assumption, not another actual refund |
| L. Provisional contribution after reserve = J − K | [calculate] | 5 | Replace estimates as actual costs arrive |

The sensitivity assumes five additional refunds of 100, with 55 recovered product cost and 10 return cost each: 5 × (100 − 55 + 10) = 275. These are different assumed orders from the twenty already recorded.

Control: pre-return contribution 1,200 − (refunds 2,000 − recovered cost 1,280 + return costs 200) = 280.

## Reproduce the observed result

Use https://www.iamgeorgekelly.com/field-guide/roas-poas-profit#calculator with:

- Net revenue: 8,000
- Cost of goods: 4,220
- Other variable costs: 1,500
- Advertising: 2,000

Expected: contribution before ads 2,280; after ads 280; net-revenue ROAS 4; contribution return 1.14; break-even ROAS approximately 3.51. Ratios are multiples. Do not compare the net-revenue break-even ratio with pre-refund dashboard revenue.

## Decision record

- Missing costs, credits, or duplicate adjustments: [fill]
- Which values are actual and which are estimates: [fill]
- Platform attribution versus cohort allocation differences: [fill]
- Decision supported by this evidence: [fill]
- Owner and next review date: [fill]

## Sources

- Shopify sales/reversal definitions and dates: https://help.shopify.com/en/manual/reports-and-analytics/shopify-reports/report-types/default-reports/sales-report
- Google Ads conversion-adjustment prerequisites: https://support.google.com/google-ads/answer/7686280?hl=en

Checked September 5, 2026. This document performs no uploads, conversion changes, or campaign actions.