1. Define period, status and unit of analysis
FSV provides seven complete records: four sales, two returns and one cancellation. P1 and P2 are consecutive equal-length laboratory periods; P3 is missing. S1 and S2 are fictional items. MU is not a real fee.
A confirmed sale adds quantity times price and quantity times reference cost. A confirmed return subtracts both at its recorded price and cost; cost reversal is an exercise assumption, not a universal accounting rule. Cancellation contributes nothing. Counts are not unique buyers.
Activity: supply a field/status dictionary. Identify missing effective expenses, return dates, condition of returned inventory and net profitability data. Never fill missing P3 with zeros.
Worked feedback
P3 = missing; cancelled C01 is excluded, not an invented zero sale. Seven rows are not seven sales.
2. Reconcile sales and returns
P1: gross sales 1400 MU, returns 100, net sales 1300; gross cost 960, reversal 60, net cost 900; reference difference 400. P2: gross 1500, returns 200, net 1300; gross cost 1200, reversal 180, net cost 1020; difference 280.
Observed P1+P2 total: gross 2900, returns 300, net 2600; net cost 1920; difference 680. Cancelled C01 does not add its 300 MU. Difference is not net profit: expenses and actual conditions are unknown.
Activity: show signed contributions, SKU/period subtotals and consolidated totals. A return is not another sale; retain its fictional identifier and source.
Worked feedback
P1 1300−900=400; P2 1300−1020=280; 2600−1920=680. Reconcile rows before comparisons.
3. Compare using the right base and weights
Reference margin = difference / net sales. P1: 400/1300 = 30.77%; P2: 280/1300 = 21.54%; total: 680/2600 = 26.15%. Net sales change: (1300−1300)/1300 = 0%; difference falls 120 MU or 30% against base 400.
Never average margins without weights. In P1, S1 net sales 900, difference 360, margin 40%; S2 net sales 400, difference 40, margin 10%. Simple average 25% is wrong for P1: (360+40)/(900+400) = 30.77%.
Activity: explain two comparisons with explicit denominators and three limitations. Missing P3 prevents P2→P3 change; it does not establish a fall to zero, annual trend, forecast, ROI or causality.
Worked feedback
A simple SKU average gives 25%, but weighted P1 margin is 30.77%. The base and weights explain why.
4. Test exceptions and present limits
Independent cases: exact duplicate V01 is counted once; V01 reused with another quantity conflicts and blocks calculation. Negative quantity, unknown SKU, unrecognized type or nonfinite amount is an error, never silently corrected.
A return exceeding sold quantities for the same period/SKU invalidates this bounded model. An explicit period with matching sale and return has zero net sales: margin is not calculable. This differs from a missing period.
Submit dictionary, reconciliation, weighted analysis and exceptions log. Editorial self-inspection: traceability, arithmetic, denominators and limits; no grade, approved PRO rubric, benefit, enrollment or completed human review.
Worked feedback
Exact repetition counts once; conflict or error stops calculation. Zero net means not calculable; missing means unobserved.
Complete source FSV
| id | period | sku | type | qty | price | cost | status |
|---|---|---|---|---|---|---|---|
| V01 | P1 | S1 | sale | 10 | 100 | 60 | confirmed |
| V02 | P1 | S2 | sale | 2 | 200 | 180 | confirmed |
| D01 | P1 | S1 | return | 1 | 100 | 60 | confirmed |
| V03 | P2 | S1 | sale | 5 | 100 | 60 | confirmed |
| V04 | P2 | S2 | sale | 5 | 200 | 180 | confirmed |
| D02 | P2 | S2 | return | 1 | 200 | 180 | confirmed |
| C01 | P2 | S1 | sale | 3 | 100 | 60 | cancelled |