Define contribution profit before comparing channels

High GMV does not guarantee retained economics once fees, media, freight, refunds, and subsidies settle.

Build a bridge from transaction value to net revenue, gross profit, and contribution profit without arbitrary fixed-cost allocation.

Attach each cost at its valid grain

Join line-level fees directly, allocate order freight once, and attribute media by campaign, item, or channel. Keep unassignable cost visible.

Platform settlement, payment, and internal order data need stable reconciliation keys.

Handle maturity and accounting time

Sales, media charges, returns, and settlement can cross periods. Offer operating-event and financial-settlement views and flag immature orders.

Do not compare new unsettled business to mature history as if both were final.

Control for mix before attributing a channel effect

Product, customer, campaign, and inventory mix may explain differences. Compare aligned items or cohorts.

Observed difference is a lead, not proof that moving volume will reproduce the result.

Reconcile from invoice to order

Sample platform and advertising statements, orders, logistics, and returns, checking allocation conservation and unmatched cost.

BI0 can be evaluated for multichannel analysis; statement connectors, allocation, and finance reconciliation require confirmation.

Build a reconciled bridge to contribution profit

Move from transaction value through merchant discounts, cancellations, and refunds to net revenue; subtract product cost for gross profit, then attributable platform, payment, fulfilment, media, and service cost for contribution profit. Keep corporate overhead above this decision layer unless a governed driver exists.

Document tax, subsidy, freight, and returned-commission treatment. The metric is a decision contract agreed by finance, channel, and supply owners, not a convenient sum of available columns.

Model order lines, cost events, and settlement batches

Keep merchandise at line grain, charges as dated cost events, and actual transfers and adjustments in settlement facts. Reconcile platform order, suborder, line, and settlement identifiers; route unmatched records to an explicit queue.

Allocate order-level discount or freight using a declared driver such as paid value, weight, or units. Preserve original amount, allocation version, and unmatched balance, and enforce conservation before and after allocation.

Account for return tails and maturity

Orders, fulfilment, returns, advertising, and settlement cross accounting periods. Offer operating-event and financial-settlement views and mark after-sales and charge maturity.

Compare equal-age cohorts or show an estimate range. A current month with open return windows is not directly comparable with fully matured history.

Control mix before attributing a channel effect

Product margin, region, fulfilment, campaign, and customer return behavior can create apparent platform differences. Drill to comparable SKU and operating cohorts before interpreting residuals.

Historical association is not the guaranteed gain from moving volume because price, traffic, competition, and cannibalization change. Use a limited experiment for consequential moves.

Accept with conservation and state the BI0 boundary

Reconcile payment to transactions, charge statements to allocated totals, and receivables less deductions to cash. Sample partial return, cross-period settlement, subsidy, bundles, and unmatched cost.

BI0 may be evaluated for multichannel operating analysis. Connector coverage, allocation policy, accounting treatment, and reconciliation scope require project verification; material unmatched cost should remain visible rather than being converted into false precision.

Public references

Take the next step with BI0.AI

Talk through a real business scenario and see how governed AI BI can fit your team.

Explore BI0.AI