Quoted margin versus banked margin on an own-brand product

The opening figure is an assumption, not a result

Subtracting landed cost from intended selling price gives a number that is useful only as a ceiling. It presumes perfect sell-through at list price, no breakage, no refunds, no clearance and no price move by a competitor. Every one of those assumptions is testable, and each one that fails takes a bite out of the same pool. Treat the opening figure as the best case against which reality is measured rather than as the plan, and expect the gap to be widest on your newest products, where you have no trading history to calibrate against and the strongest incentive to be optimistic.

Deductions that never appear on a purchase order

Retail buyers expect participation in promotions, listing support and sometimes a settlement discount, and these arrive as deductions from remittance rather than as an invoice you approve. Online channels net off commission, fulfilment, storage that rises for slow-moving stock, refund handling and the disposal of units that come back unsellable. Payment processing takes a slice of every transaction, and returns take both the margin and the unit. Ask, before signing into a route, for the complete deduction schedule and model it into the stack. Operators are rarely surprised by the headline commission; they are regularly surprised by everything charged alongside it.

Obsolescence is a margin line, and it sits on your balance sheet

The factory has been paid. Whatever fails to sell is your asset until you write it down, and the write-down is as real a cost as the freight was. Ranges accumulate a tail almost automatically: the colour that never moved, the size nobody wanted, the variant launched because the minimum was per style rather than per unit. Dated goods make the clock explicit; undated goods rot commercially instead, becoming unsellable when the pack design moves on or the trend passes. Carrying the tail at full book value keeps reported profit looking healthy while the warehouse fills with stock that will eventually clear below cost.

The cost of buying the sale belongs inside the unit economics

Where demand is created by advertising rather than by shelf presence, the spend that makes a listing move is a variable cost of each unit sold, not a general overhead to be admired separately. Treating it as marketing produces a range that shows attractive product margin and consumes cash regardless. The number worth tracking is contribution per unit after channel deductions and after acquisition spend, because that is what decides whether another production run makes you better off. A variant with a fat product margin that only sells under paid promotion may contribute less than a plainer item that sells unassisted.

A blended average conceals the variants that are losing money

Range-level margin is the most comfortable number in the business and the least informative. It averages the item that funds the operation with the item that quietly drains it, and because both share packaging, photography and a warehouse, the loser is simple to defend on the grounds that it completes the range. Rebuild the view per variant, allocating channel deductions, returns, storage and acquisition spend to the specific item that incurred them. The exercise usually identifies a small group of products earning most of the contribution, and a longer group that has been financed by them for longer than anyone realised.

Frequently asked questions

Why does my reported margin keep coming in below the margin I planned?

Usually because the plan captured only the two figures nearest to hand — landed cost and list price — while the difference is eaten by items recorded elsewhere in the accounts. Channel deductions land as reduced remittances, returns land as refunds, ageing stock lands as a write-down at period end, and promotion lands as a discount rather than a cost. Rebuild the calculation from actual cash received per unit rather than from the price you published.

Should advertising spend come out of product margin or sit in overhead?

If the sale would not have happened without it, it belongs in the product's economics. Spend that builds durable brand recognition can reasonably sit in overhead; spend that buys placement for a specific listing behaves exactly like a variable cost and should be judged as one. The practical test is what happens when you switch it off. If units stop moving, you were not marketing the brand, you were purchasing each transaction individually.

How do I decide whether a slow variant is worth keeping in the range?

Look at what it contributes after everything attributable to it, then at what it occupies: cash locked in stock, warehouse space, artwork upkeep, a share of your attention at every reorder. A variant that covers its direct costs but consumes buying capacity better spent on a proven line is still costing you. Range completeness is a genuine argument in retail settings and a much weaker one online, where nobody sees the gap.