Choose a unit definition and keep it consistent
MarginTrial starts from gross promotion units before separately entered full-refund units. It subtracts the refunded promotion revenue but conservatively leaves the variable cost of every gross unit unrecovered. If you import units already net of those returns and also enter returned units, the same loss is counted twice.
A payment refund without a physical return, a partial goodwill adjustment and an order cancellation can have different quantities and costs. Reconcile those transactions before converting them to the workbench's modeled units. A platform column containing a broad reversal count should not be mapped merely because its name sounds similar.
Worked example: cost recovery changes the answer
A campaign ships 1,000 gross units at €40 with €25 variable cost per unit. One hundred units receive full refunds. Before refunds the product contribution is €15,000; refunding €4,000 reduces the workbench result to €11,000.
| Component | Calculation | Amount |
|---|---|---|
| Gross product revenue | 1000 × €40 | €40,000 |
| Cost incurred on gross units | 1000 × €25 | −€25,000 |
| Full-refund revenue | 100 × €40 | −€4,000 |
| Contribution before other campaign effects | €11,000 |
Do not disguise recovered inventory
If all returned units actually recover their full €25 variable cost, that alternative would add €2,500 relative to the conservative result. The workbench does not assume that recovery. Return freight, inspection, damage, resale value and payment-fee treatment may prevent full recovery.
Do not reduce returns to make the result match another accounting report. Keep the operational return count and document the different cost-recovery model. Use a separate reconciliation when recovered inventory is material.
Map the returns-contribution worksheet
The returns-contribution template uses 1,000 gross units and 100 full-refund units. A baseline of 800 units at €50 and €25 cost contributes €20,000; the campaign's contribution difference is −€9,000 before other effects.
| Worksheet item | CSV field | Value |
|---|---|---|
| Units before these refunds | promo_units | 1000 |
| Full-refund units | returns | 100 |
| Realized price before these refunds | promo_price | 40 |
| Variable cost per gross unit | unit_cost | 25 |
Choose a review date that captures the cohort
A campaign can end before its customers complete the return window. State the review date and whether later returns are still expected. Comparing a mature baseline with a fresh promotion cohort can understate the promotion's eventual refund burden.
The workbench also has no separately modeled baseline refund process. If baseline returns are material, its simple relative comparison can penalize promotion refunds asymmetrically. That decision needs a fuller benchmark rather than silently relabeling the baseline quantity.
Frequently asked questions
Why might my sales report differ? Transaction timing, reversals, tax, shipping and cost definitions can differ. Reconcile those definitions before looking for a rounding error.
Can returned units exceed promotion units? Not for one consistently defined promotion cohort in this model. An excess usually signals a mixed cohort, timing mismatch or an inappropriate quantity mapping.
Put it to work
Source context
Guide text and examples were prepared independently for MarginTrial. Sources provide the stated methodological or product context. They do not validate the hypothetical commercial outcomes.