D2 Commerce Knowledge · Profitability

Commerce profitability starts after GMV — and stays separate from payout.

A decision-ready commerce P&L connects valid revenue to SKU-level COGS and the variable costs required to sell it, while settlement remains a separate cash view. The result should be traceable to source evidence, and every unresolved mapping should remain visible instead of being estimated away.

See the profit waterfall Automation & Reporting

Direct answer

What makes commerce profitability decision-ready?

Revenue, variable selling costs and product cost need to reconcile at the lowest reliable level. Maintain COGS by SKU and effective date, map it to transactions, deduct the platform, Ads, creator and other variable costs in scope, and keep settlement as the cash-realization layer. If a cost or mapping is unresolved, mark the economics as incomplete instead of pretending the missing value is zero.

Profit waterfall

Move from commerce activity to contribution before interpreting cash.

Commerce activity

Start with the order events and SKU quantities that explain what was sold in the period. GMV is useful context, but it is not yet retained revenue.

Valid revenue

Apply the declared order-status, cancellation, refund and adjustment rules used by the reporting model to establish the sales basis.

Product cost

Map quantity sold to SKU or bundle COGS using a controlled cost source and the effective date that belongs to the transaction period.

Variable selling costs

Deduct the platform, offer, Ads, creator, fulfillment or other variable costs included in the operating scope.

Contribution

The amount left after the variable economics in scope have been deducted from valid revenue. This is the operating profit layer used for scale, hold and SKU-mix decisions.

Settlement bridge

Reconcile contribution evidence to platform settlement and payout without replacing the sales-period view with the cash timeline.

Operating formula

Contribution = valid revenue − COGS − platform/offer costs − Ads − creator/Affiliate costs − other variable selling costs in scope

This is an operating contribution model, not a universal accounting statement. The included cost categories, revenue basis and period rules must be declared explicitly.

Cost stack

Profitability fails when one cost layer disappears from the model.

COGS

Product cost should be governed by SKU, effective date and bundle rule so current costs do not overwrite historical order economics.

SKU · unit cost · effective date · quantity

Platform deductions

Use recorded fee and adjustment evidence where available rather than one assumed marketplace percentage across every period and product.

fee line · commission · adjustment · settlement

Paid-media spend belongs in the same period and commercial scope as the revenue it is used to evaluate. Media efficiency alone does not establish contribution.

spend · campaign · SKU scope · period

Creator / Affiliate

Commission, paid creator fees and sample costs should be visible according to the allocation logic actually supported by the evidence.

commission · sample · fee · attribution scope

Offer & variable fulfillment

Seller-funded vouchers, packaging and other variable costs matter when they change with the transaction and materially affect the decision being made.

voucher · packaging · per-order variable cost

Exception cost state

Missing or unresolved cost is not zero cost. Mark the contribution as incomplete until the evidence is restored or an explicit business-approved estimate rule exists.

missing cost · unknown SKU · unresolved fee

Evidence map

Use each source for the question it was designed to answer.

SourceQuestionEvidenceDecision use

Orders

What sold in the sales period?

Order ID, SKU, quantity, status, commercial timestamps

Valid revenue, SKU mix, order-period performance

Settlement / finance

What did the marketplace deduct or pay?

Fee lines, commission, refunds, adjustments, payable or paid amount

Realized deductions, payout timing, reconciliation

Ads

What did paid acquisition cost?

Spend and campaign/product identifiers in the chosen scope

Ads ratio, contribution after Ads, scale/hold review

SKU cost master

What did the product sold cost at that time?

SKU key, unit cost, effective date, bundle mapping

Order-level COGS and SKU contribution

Creator / other variable-cost evidence

Which additional costs moved with commerce activity?

Commission, creator fee, samples, packaging or other documented variable cost

More complete contribution economics

Exception register

Which parts of the result are not yet trustworthy?

Unmatched IDs, unknown SKUs, duplicate mappings, missing costs or unresolved deductions

Recovery ownership and decision boundaries

P&L vs cash

Reconcile sales economics to settlement. Do not merge the timelines.

Sales-period P&L

Anchor revenue and transaction economics to the period in which the commercial activity occurred. Use this view to compare SKU mix, campaigns, Ads and contribution.

Settlement-period cash

Anchor payout and realized deductions to the settlement timeline. Use this view to understand cash realization, pending balances and later marketplace adjustments.

Reconciliation bridge

Connect the two through stable order or business identifiers. Preserve pending and unmatched items rather than forcing both timelines to total to the same number on the same date.

Exception layer

Incomplete evidence should stay visible.

A complete-looking dashboard can be less trustworthy than a report that openly says which rows are unresolved. Exceptions are operational states with owners and recovery actions.

Missing COGS

Contribution cannot be finalized for the affected SKU or order.

Restore SKU mapping, effective cost date or bundle rule before treating the result as decision-ready.

Unmatched settlement

Cash evidence cannot yet be tied back to the underlying commerce event.

Investigate stable identifiers, timing and platform adjustment records instead of assigning it arbitrarily.

Unknown fee / adjustment

A deduction exists but its business meaning is unresolved.

Classify the line and document whether it belongs in contribution, cash reconciliation or another accounting layer.

Duplicate mapping

One source row may be counted more than once or one event may link to multiple business objects.

Resolve the mapping before roll-up; do not rely on aggregate totals to hide duplicate logic.

Period mismatch

Revenue, Ads or cost is being compared across incompatible time windows or time zones.

Rebase the evidence to a declared period before calculating ratios or contribution.

Attribution uncertainty

A cost or revenue signal cannot be assigned reliably at SKU or order level.

Keep the allocation level coarser or qualitative until the evidence supports a more precise claim.

Decision rules

Profitability should change the next operating decision.

GMV increases, contribution decreases

Identify whether SKU mix, discounting, Ads, commission or other variable costs changed before scaling further.

ROAS improves, contribution stays weak

Media may have improved while the underlying SKU economics remain below the required margin. Review the full cost stack.

A SKU has high gross margin but low contribution

Inspect the variable selling costs attached to acquiring and fulfilling that SKU rather than stopping the analysis at COGS.

Payout is lower than sales-period revenue

Check settlement timing, pending balances, recorded deductions, refunds and adjustments before classifying the difference as lost revenue.

Total shop is profitable but one hero SKU is negative

Do not let shop-level aggregation hide the SKU that is consuming contribution. Review its role and whether the loss is deliberate and bounded.

Exception volume becomes material

Pause fine-grained scale decisions until the unresolved data is small enough or explicitly bounded for the decision being made.

Profitability checklist

Before a profitability number becomes a decision, verify the evidence underneath it.

Define GMV, valid revenue, gross margin, contribution and payout before presenting the report.

Declare whether each view is sales-period or settlement-period based.

Keep order and SKU identifiers stable across sales, settlement, Ads and COGS sources.

Maintain SKU and bundle COGS with effective dates so historical economics remain reproducible.

Use recorded platform deductions where evidence exists instead of one assumed fee percentage.

Keep Ads spend in the same commercial scope and period as the revenue used for efficiency analysis.

Include creator, voucher, packaging and other variable costs when they are material to the decision scope.

Separate unresolved data into an exception layer rather than converting missing values into zero.

Reconcile P&L evidence to settlement and payout without collapsing the two timelines into one.

Review profitability at the lowest reliable level, then roll up to SKU, campaign, channel or shop.

Attach an owner and recovery action to every material reconciliation exception.

End the review with an operating decision: scale, hold, reprice, change mix, reduce acquisition cost, fix data or investigate cash timing.

Related paths

Go deeper into cost, marketplace P&L and the reporting system.

COGS Explained

Go deeper into SKU-level product cost, effective dates, bundles and transaction-level COGS mapping.

Open

TikTok Shop GMV vs Profit

See the TikTok Shop-specific contribution waterfall from GMV and valid revenue through fees, Ads, commission and COGS.

Open

Commerce Control Dashboard

See how D2 turns orders, fees, Ads, COGS, settlement and exceptions into a decision-support operating layer.

Open

Commerce reporting

Need one profitability logic across orders, fees, Ads, COGS and settlement?

D2 can structure the source mapping, reconciliation, P&L rules and exception layer so the team can spend less time compiling files and more time deciding what to scale, hold or fix.

Discuss commerce reporting

FAQ

Commerce profitability questions

What is the difference between gross margin and contribution margin in commerce?

For this operating model, gross margin focuses on revenue after product cost, while contribution goes further by deducting the variable selling costs included in the decision scope, such as platform deductions, Ads, creator or Affiliate commission, vouchers and other transaction-linked costs. The exact accounting definition should remain explicit rather than assumed.

Should commerce profitability be calculated by SKU or by order?

Maintain controlled product cost by SKU and effective date, then calculate transaction-level COGS from the quantity actually sold. This preserves a governed cost master while allowing contribution to roll up by order, SKU, campaign or period when the identifiers are reliable.

Why should P&L and marketplace settlement be kept separate?

P&L explains the economics of the sales period. Settlement and payout explain when cash is realized and which deductions the platform records. Reconcile the two through stable identifiers, but do not force them into one date basis because an order can be sold in one period and settled or adjusted in another.

How should missing COGS, unmatched orders or unresolved fees be handled?

Keep them visible as exceptions. A profitability report should distinguish resolved contribution from unresolved economics instead of silently estimating missing values merely to make every row complete. The exception should have an owner and a path to reconciliation.

Can a marketplace shop grow GMV while becoming less profitable?

Yes. GMV can rise while product mix shifts toward lower-margin SKUs, discounting increases, Ads consumes more revenue, creator commission rises or platform and fulfillment costs change. Growth should therefore be reviewed through contribution and cost mix, not topline commerce activity alone.

What data is needed for a decision-ready commerce profitability report?

At minimum, use order-level sales evidence, settlement or finance deductions, Ads spend, a controlled SKU cost master and the other variable selling costs material to the business model. Preserve stable identifiers, period definitions and an exception layer so the result can be traced back to source evidence.