What is COGS in ecommerce?

For ecommerce, COGS (Cost of Goods Sold) is the total variable cost of fulfilling one order — the product itself, the freight to get it to your warehouse, the packaging consumed per order, and the labour to pick-pack-ship it. Ad spend and Shopify processing fees are not COGS; they live elsewhere on the P&L. nouz takes a COGS snapshot at the moment of sale, so an order placed today is costed against today's landed unit cost — not whatever the cost happens to be next month when the supplier raises prices.

TL;DR

The definition, in shop-owner English

COGS is the cost that exists only because the sale exists. If you sold one fewer unit this month, your product cost drops by one unit, your inbound freight drops by the per-unit share, your packaging drops by one mailer, and your fulfillment labour drops by the minutes that order would have taken. Everything else — rent, salaries, ads, software — stays the same. The four items above are COGS. Everything else is something else.

The clean test: if the order had not happened, would I have spent this money? Yes → not COGS. No → COGS. Ad spend would still have been spent (you ran the ad before the sale). Shopify's monthly subscription would still have been paid. Rent would still have been paid. None of those are COGS for that order.

The formula and what to include

COGS per order = Product unit cost + Inbound freight + Packaging + Per-order fulfillment labour

Include:

• Product unit cost — what you paid the manufacturer or supplier for one unit, plus any tariffs or duties.
• Inbound freight — shipping from supplier to your warehouse, divided per unit.
• Packaging consumed per order — mailer, box, void fill, tissue, stickers, thank-you card.
• Per-order fulfillment labour — the human (or 3PL pick-pack fee) cost of getting that order out the door.

Exclude:

• Ad spend — that is variable marketing cost, tracked separately for CAC.
• Shopify subscription fees — fixed software cost.
• Card processing fees — variable payment cost, not COGS.
• Warehouse rent — fixed cost, allocated separately.
• Salaries (non-fulfillment) — operating expense, not COGS.

The reason the split matters: gross margin is computed as (Net revenue − COGS) ÷ Net revenue. If you stuff non-COGS items into COGS, gross margin drops and EBIT looks better than it should. If you leave true COGS items out (like packaging or freight), gross margin looks higher than reality.

Worked example: €29,50 per €60 order

A small DTC accessories brand selling a €60 leather wallet:

Now compare to the version most brands use, where COGS is just the product cost:

The product-only version overstates gross margin by 9 percentage points. Plug that into a CLV calculation and customer value comes out 22% too high. Plug it into break-even AOV and you set a free-shipping threshold €8 too low. The pricing decisions, the channel-spend decisions, the bundle decisions all anchor on a number that does not match the bank.

Benchmarks and gross margin targets

These benchmarks assume fully-loaded COGS — product, freight, packaging, and fulfillment labour. A brand reporting 70% margin in apparel using product-only COGS is probably actually at 55-60% on a fully-loaded basis. The difference is exactly the gap between "looks profitable" and "is profitable."

Why getting COGS wrong silently kills margin

COGS is the input to almost every other unit-economics number: gross margin, CLV, break-even AOV, contribution margin, ROAS judgment. Get COGS wrong by 9 points and every downstream metric is wrong by a corresponding amount. The failure is silent because nothing breaks — the store still ships, the dashboard still shows a margin, the number is just quietly the wrong one. You only feel it at the bank, months later, when the profit the reports promised never arrives.

Common mistakes

The most common COGS mistakes in small ecommerce — each one pushes reported margin above real margin, which is the dangerous direction:

• Forgetting inbound freight, especially when ordering from overseas — the per-unit share of a sea or air shipment is a real landed cost, not a footnote.
• Underestimating packaging — a custom mailer + tissue + insert + sticker easily runs €1-€2 per order, and it scales with every single one.
• Ignoring fulfillment labour when the founder still picks-and-packs orders themselves. Free-feeling founder time is still a real per-order cost; the day you hand it to a 3PL, it appears as a line, so cost it from the start.
• Using old supplier prices from six months ago instead of the latest invoice — margin drifts down invoice by invoice while the spreadsheet says it is flat.
• Stuffing ad spend into COGS — that compresses gross margin and inflates EBIT, and it hides your true CAC, so two numbers go wrong at once.
• Averaging one COGS across every SKU — a €5-landed keychain and a €40-landed jacket do not share a margin. Blended COGS hides which products actually make money and which you are subsidising.
• Ignoring returns and refunds — a returned order still ate packaging, outbound shipping and labour. On a daily P&L, track it as a refund plus a returns-handling cost rather than pretending the order never cost anything.

How it shows up in your daily P&L

A daily P&L only tells the truth if COGS is right at the line level, every day. In nouz the chain runs the same way each day: gross revenue minus tax minus card fees gives net revenue; net revenue minus COGS minus the remaining variable costs (shipping-out, non-COGS packaging) minus the day's slice of fixed cost gives EBIT. If COGS is understated by 9 points at the top of that chain, the EBIT at the bottom is overstated by the same amount — every single day, compounding into a month-end number that never matches the bank.

This is why nouz snapshots COGS at the moment of sale. An order placed today is costed against today's landed unit cost, so when the next supplier shipment arrives dearer, new orders carry the new cost while old orders keep theirs — historical margin stays accurate instead of being retroactively rewritten. The result is a daily EBIT you can act on the same evening, built on a fully-loaded COGS rather than a product-only guess that flatters every number downstream of it.

For a store selling several products, that same-day accuracy is what lets you see per-order profit at the SKU level rather than a single blended average. A promotion that pushes volume toward a thin-margin line shows up in the day's EBIT immediately, not in a quarter-end review. The habit worth building is to read the day against its fully-loaded COGS every evening — the number takes seconds to check and it is the difference between running the store on what the bank will actually show and running it on a hopeful estimate.

Related concepts:

• COGS snapshot explained — why nouz freezes COGS at order time.
• True profit per order on Shopify — the full per-order calculation.
• The Shopify profitability guide — how COGS fits into the full daily P&L.

Common questions