Skip to content

Profit by SKU: find best-sellers that lose money

Build a product P&L to find best-selling SKUs that lose money: revenue × margin matrix, causes of loss, a traffic-builder check and decisions.

By The DataxLand team· · 6 min read

Best-selling products usually get the most favorable treatment: more ads, more campaigns, more restocking. But "best-selling" is about revenue, while the owner's real question is how much money that product leaves behind. Some SKUs at the top of the sales ranking make the shop poorer with every order.

This post shows how to build a product-level profit and loss report, sort SKUs into four groups, pin down why a SKU loses money, and decide what to do about it — including when a losing SKU is worth keeping.

The numbers use DataxLand sample data (a simulated scented-candle and essential-oil shop, 1 Aug – 29 Oct 2026) and clearly stated illustrative assumptions. They are not a real customer's numbers.

What a best-seller that loses money looks like

In the sample data, product MH-NEN-03 brings in 52,741,533 ₫ of net revenue over three months, nearly 7% of the whole shop's net revenue. But its contribution margin is −7,697,340 ₫, about −14.6% of net revenue, and that is before advertising.

A negative contribution margin means that after cost of goods, marketplace fees, payment fees, shipping and return costs, every unit of this SKU's revenue already loses money. More advertising only makes the loss bigger. On a sales report, MH-NEN-03 looks like a good product.

The columns a product profit and loss report needs

ColumnHow to calculateCommon mistake
Net revenuePrice × quantity − line discount − allocated voucher − refundsNot allocating order-level vouchers to lines
Cost of goodsQuantity × cost of goods on the sale dateUsing the latest purchase price for every order
Marketplace and payment feesFrom settlements, allocated to lines by net revenueOnly order-level fees, never split to SKUs
Shipping, return costsThe shop's share, allocated to linesIgnoring return shipping
Contribution marginNet revenue − the items above—
Allocated advertisingThe channel's ad spend for the day, split by net revenueCharging all ads to the advertised SKU
Profit after feesContribution margin − allocated advertising—

Two background posts if you have not done the first two steps yet: allocating order-level vouchers and cost of goods by effective date.

Sort SKUs into four groups

Take net revenue and contribution margin (%) for every SKU over the same, long enough period (8–13 weeks), split by median revenue:

GroupRevenueContribution marginWhat to do
PillarsHighPositive, highKeep stock, prioritize ads within break-even ROAS
Best-sellers that lose moneyHighNegative or very lowFind the cause now; stop pushing ads until fixed
PotentialLowPositive, highTest more visibility, bundle with pillar SKUs
Long tailLowLowConsider stopping restocks, clearing stock, merging SKUs

The second group is the most expensive one, because high revenue times a negative margin gives a big loss, and because it is usually treated like the first group.

Pin down the cause: which line makes the SKU lose money

Illustrative example (assumed) for a glass-jar candle SKU, per single-item order:

ItemPer order% of net revenue
Net revenue (after voucher)89,000 ₫100%
− Cost of goods (after a purchase price increase)52,000 ₫58.4%
− Percentage fees (20%)17,800 ₫20.0%
− Fixed fee per order3,000 ₫3.4%
− Shipping paid by the shop (fragile, heavy packing)8,000 ₫9.0%
− Expected return cost: 10% × (25,000 ₫ return shipping + 52,000 ₫ broken goods)7,700 ₫8.7%
= Contribution margin500 ₫0.6%
− Advertising (12% of revenue)10,680 ₫12.0%
= Profit after fees−10,180 ₫−11.4%

No single item looks "excessive" on its own. But three small ones together (fixed fee, shipping, returns) eat 21% of revenue, almost as much as all the marketplace fees. That is typical of low-priced, fragile or bulky SKUs: per-order costs do not shrink when the price does.

Five common causes, in the order to check them:

  1. Purchase price went up while the selling price stayed put. Compare this month's cost of goods with three months ago.
  2. High return rate (breakage, wrong description, not as pictured). Compare the SKU's return rate with the shop average.
  3. Per-order costs large relative to price: fixed per-order fee, shipping paid by the shop, packing.
  4. Vouchers or campaigns concentrated on the SKU, especially when it is the cheap item in the order.
  5. Sold mostly on a high-fee channel or a channel with a high return rate.

Losing SKUs that bring in buyers: check before you cut

Some SKUs lose money on purpose: a cheap item that draws buyers into the shop to buy more. Before cutting, check two numbers:

  1. Attach rate: what percentage of orders containing this SKU also contain other SKUs.
  2. Whole-order contribution margin for orders containing this SKU, compared with orders without it.

If orders with the losing SKU are still nicely profitable at order level, the SKU is doing its job; manage it as a budgeted customer-acquisition cost. If most orders contain only that SKU and those orders lose money, the loss is real.

Run the SKU report every month

  1. Fix the period: use 8–13 weeks, avoid a period with just one sale event.
  2. Calculate net revenue, contribution margin and profit after fees for every SKU, plus return rate and the share of revenue with a cost of goods.
  3. Move SKUs missing a cost of goods to a separate list: their profit cannot be trusted.
  4. Sort into the four groups; for "best-sellers that lose money", pin down the cause using the table above.
  5. Record a decision for each SKU (raise the price, change packaging, pull from campaigns, stop ads, keep as a traffic builder) and review after 4 weeks.

How DataxLand does this

DataxLand calculates profit per order line: order-level vouchers allocated to lines, cost of goods by sale date, marketplace and payment fees from settlements, and the channel's daily ad spend split across lines by net revenue. So the SKU table always adds up exactly to the shop's profit. The retail profit and loss dashboard gives you net revenue, contribution margin and profit after fees by product and channel, together with the share of revenue that has a cost of goods.

With Si, you just ask: "which products had high revenue but negative contribution margin this quarter", "on which channel does MH-NEN-03 lose money". The answer has a Sources section showing the metric and version (see Ask Si in the app). If you sell on Shopee, see more suggested questions on the Si for Shopee data page.

Frequently asked questions

How do I calculate profit per product when there are order-level vouchers?

Split the voucher across lines in proportion to each line's value after discounts, then calculate net revenue per line. Without the split, SKU profit is inflated; if you load it all onto one line, one SKU loses money unfairly.

Should advertising be allocated to the advertised SKU?

Only if you have reliable product-level ad data. One ad usually pulls in a whole basket, so splitting the channel's ad spend by net revenue is more stable, as long as the rule is documented.

How long a period is enough to conclude a SKU loses money?

At least a few weeks — long enough to pass the return window and have settled fees. A single sale week can make any SKU look like a loss-maker.

What about new SKUs without a cost of goods?

Put them on a separate list and enter their cost of goods first. A SKU missing a cost of goods is treated as zero cost and looks extremely profitable.

Read next: How to calculate true profit · Pricing on marketplaces from your target profit · Return rate and its true cost

Related posts

Verified dashboards from your own data, this afternoon

Free plan with no time limit · No card required · Sample workspace ready in minutes