Consolidated COGS reconciliation across a multi-brand DTC portfolio

If you run 5+ DTC brands, you've had this conversation with your CFO: "What's our actual per-brand contribution margin?" And the answer is always some variant of "let me pull three exports, do a VLOOKUP, and I'll get back to you Thursday." This article is the operator-level guide to making that answer arrive in 10 minutes instead of 4 days.

1. Why multi-brand COGS is harder than single-brand COGS

Single-brand COGS is mechanical. Revenue = Shopify report. Product cost = 3PL invoice. Shipping = Shopify + 3PL. Payment processing = Stripe export. Marketing = Meta + Google. Subtract costs from revenue. You have contribution margin.

Multi-brand breaks the mechanism in three places:

Revenue attribution. Five brands on five Shopify stores, five processor accounts, one shared email service provider that sends across all five. Revenue gets double-counted if any brand has a "purchased on brand A, fulfilled from brand B's 3PL" cross-sell.

Shared overhead allocation. The shared 3PL charges a master warehouse fee. The Meta account runs ads for 5 brands simultaneously. The customer service team covers all brands. How do you allocate these fairly? Most operators either over-allocate to the fastest-growing brand (making it look worse) or under-allocate (making it look artificially profitable).

Processor settlement timing. Stripe settles in 2 days, Authorize.net settles in 1 day, Square settles same-day. A COGS report that aligns revenue (settlement date) with costs (shipment date or invoice date) double-books everything across month-end boundaries.

2. The three-system-agreement model

Every COGS reconciliation process has to reconcile three systems:

• OMS (order management system): what orders were placed, what shipped, what was refunded. Shopify, Magento, WooCommerce, ShipStation.
• Processor: what charges cleared, what was disputed, what was refunded back. Stripe, Authorize.net, your acquiring bank.
• Accounting GL: what landed in QuickBooks / Netsuite / Xero as revenue and cost entries.

These three systems agree less often than operators expect. Common disagreements: an order placed on Dec 30 but the payment settled on Jan 3 lives in one month in the OMS and another month in the processor report. A chargeback filed in February reverses a transaction from October that was booked as revenue in October.

The reconciliation process: pick a single reporting period (calendar month), pull all three systems with the same date range, compute the deltas, explain them. The deltas are almost always timing differences. Once labeled, they can be rolled forward or backward to align.

3. The descriptor discipline prerequisite

Reconciliation requires you to tag every transaction to a brand. If your 5 brands share a merchant account but don't use per-brand billing descriptors, the processor report just says "XYZ Holdings LLC" on every transaction. You can't reconcile by brand because the brand tag doesn't exist in the processor data.

Fix this first, before trying to reconcile anything else. Every brand gets a unique descriptor (see descriptor strategy). Every processor export then carries the brand tag at the transaction level. Your OMS already has brand tags (one Shopify store per brand). Your GL can accept a class/department field per entry.

Once all three systems carry the same brand tag on every record, reconciliation becomes a filter-by-brand operation rather than a guess-and-check.

4. COGS line items in a multi-brand model

Per-brand contribution margin calculation includes:

• Revenue: processor-settled (net of refunds, net of chargebacks that landed in period)
• Product cost: 3PL invoice per brand, allocated by SKU
• Shipping cost: 3PL-charged shipping, per-brand
• Payment processing fees: processor-reported per-brand (this is where descriptor discipline pays off)
• Chargeback losses: per-brand; include representment wins and losses separately
• Fulfillment handling: 3PL pick-and-pack per brand
• Returns processing: refund logistics per brand
• Paid marketing: Meta + Google at the ad-account level, per-brand if dedicated ad accounts; allocated by revenue share if shared
• Customer service: allocated by ticket count per brand from Gorgias / Zendesk
• Warehouse storage: 3PL storage fee, per-brand SKUs allocated

Below contribution margin but above net income:

• Shared overhead: executive team, accounting, tooling — allocated by revenue share, or held in an unallocated corporate bucket
• Tax, interest, depreciation: typically kept unallocated at the parent level

5. Automating the reconciliation

Operators at 5–10 brands with $2–10M in combined revenue have three realistic architectural options:

Option A: Shopify → Google Sheets → QuickBooks manual. Breaks at 5+ brands. The team spends 1+ week per month on close.

Option B: Purpose-built accounting layer (A2X, Finaloop, Settle). These tools auto-map Shopify revenue → processor settlements → GL entries per brand. Cost is $500–$2,500/month. Cuts close from 1 week to 2–3 days for most portfolios.

Option C: Data warehouse + dbt model. Pull OMS, processor, 3PL, Meta, Google, and GL into Snowflake/BigQuery. Build dbt models that reconcile at the brand-day grain. Cost is $1–3k/month in tooling + engineering setup. Close becomes a SQL query, but you need engineering capacity.

Most multi-brand operators we work with land on Option B until their portfolio crosses $50M combined, then migrate to Option C. Option A is a bridge that breaks at scale.

6. The audit-trail requirement

Once you have 5+ brands reconciling cleanly, you'll hit the next problem: auditors, lenders, or acquirers want to see the reconciliation. They want to know your per-brand EBITDA, per-brand contribution margin, per-brand unit economics, and they want to see that the numbers tie back to primary sources.

The audit trail requirement:

• Each brand's monthly contribution margin is computable from source data in < 60 minutes
• Every line item has a documented source (processor export, 3PL invoice, ad platform API)
• The reconciliation methodology is documented (which allocation keys are used, how shared costs are split)
• Journal entries in the GL reference the source reports

This is the same reconciliation a clean close would produce anyway. The difference for a multi-brand operator is that "clean close" means across every brand, every month, with documented methodology. Operators who build this in Year 1 have it forever. Operators who don't rebuild it during the first lender due diligence, under deadline pressure, usually more expensively.

The short version: per-brand descriptors give you clean revenue tagging; clean revenue tagging lets three systems reconcile; three systems reconciling gives you trustworthy per-brand contribution margin; trustworthy contribution margin lets you actually manage the portfolio instead of guessing at it. See the multi-brand reconciliation playbook for the operational specifics, or how multiflow prices into the stack.

7. Inventory valuation across brands with shared SKUs

A subtle multi-brand complication: shared SKUs. Some portfolios sell the same underlying product under different brand names (a private-label supplement sold under three different brand labels, or a dropshipped apparel item listed in multiple stores). COGS calculation has to handle this without double-counting or under-counting.

The clean approach:

• Every SKU has a master inventory record at the parent entity level
• Each brand's OMS carries a brand-specific SKU that maps to the master record
• Inventory movement (purchase orders, receipts, sales, returns) is tracked at the master SKU level
• Per-brand COGS allocation uses the standard cost of the master SKU at the moment of sale, not a brand-specific cost

This prevents the common bug where Brand A sells 100 units of SKU X at $10 COGS/unit while Brand B is simultaneously reporting 50 units of SKU X at $12 COGS/unit, and finance can't figure out why inventory balance doesn't reconcile.

8. Variable cost per shipment complications

Shipping cost is rarely a fixed per-unit number — it depends on origin warehouse, destination zone, carrier, package weight, and current carrier rate tier. For multi-brand reconciliation, shipping cost needs to come from the 3PL's actual shipping charge per order, not from a standard rate card.

The 3PL typically produces a daily shipment ledger: order ID, actual shipping cost, carrier. That ledger joins to the OMS (order ID match) and tags brand from the OMS record. Per-brand shipping cost aggregates cleanly from there.

Where it breaks: 3PLs that invoice monthly rather than per-shipment, or 3PLs that apply bulk carrier discounts at the portfolio level rather than per-brand. In those cases, the allocation rule becomes "allocate portfolio shipping cost by per-brand shipment count" or "by per-brand weight," which is an approximation rather than an exact match. Operators should pick one rule, document it, and stick with it across reporting periods — switching allocation methods between quarters breaks trend comparison.

9. The quarterly close cadence that actually works

For multi-brand operators at $5–50M in combined revenue, the close cadence that produces trustworthy per-brand reporting without consuming half the finance team's time:

• Daily (automated): revenue and processing fees flow from processor APIs to accounting system, tagged by brand descriptor
• Weekly (30 min): OMS-to-processor reconciliation for the prior 7 days; flag any discrepancies > 2% for investigation
• Monthly (2–3 days): full close with all COGS line items allocated; per-brand contribution margin report distributed to brand leads
• Quarterly (5–7 days): audit-trail-complete close with every line item sourced back to primary documents; board-ready per-brand P&L
• Annually (2–3 weeks): tax-ready close with accruals finalized, depreciation run, inter-company allocations documented

The cadence only works if the systems are wired correctly and the brand tagging is consistent. Operators who try to run this cadence without the per-brand descriptor foundation spend the entire month reconciling instead of managing. Operators with the foundation in place run close as a background process and spend the month on the actual portfolio decisions the close numbers support.