Skip to content

Profitability Review

Re-run unit economics on a fixed cadence and decide to scale, fix, or kill each product.

Why the model decays

The unit economics modeled before sourcing describe a moment that has already passed. Fees change. Advertising gets more expensive as competitors enter the category and bid against you. Freight rates move. Returns often run higher than the forecast that assumed everything would ship clean.

None of this announces itself. A product that penciled out at launch can quietly stop earning its shelf space while the top-line numbers still look fine. The purpose of a profitability review is to catch that drift on a schedule, before it compounds.

Model the product you have, not the one you sourced
The launch spreadsheet is a starting assumption, not a standing fact — refresh every input before you trust the margin it reports.

Review on a cadence, not a feeling

Run the review on a fixed cadence, not when something feels wrong. A feeling arrives late; a schedule catches the slow leaks and forces you to look at products that seem fine but no longer are.

1

Refresh every input

Pull current fees, freight, advertising cost, and actual return rate into the launch model.

2

Read the four signals as trajectories

Compare each against the last pass.

3

Make the one call

Scale, fix, or kill — from the signals, not attachment.

The refresh step, in full:

Refresh checklist
  • Refresh the fulfillment charge and category commission against the current schedule.
  • Enter the return rate you are actually seeing, not the one you forecast.
  • Load real advertising spend per unit, not the launch estimate.
  • Apply current freight and landed cost, not last season's rate.
  • Restate net margin per unit and flag any product that has slipped.

The four signals

Every decision comes from four signals read together. No single one is a verdict, but a break in any of them is a question worth answering.

Rating trend

The direction of the average rating, which reads changes in quality or expectation before the returns do.

Return rate

The share coming back, which converts silent dissatisfaction into a cost you can see.

Conversion rate

How efficiently traffic becomes orders, which tells you whether the listing still earns the clicks it pays for.

Acquisition cost trajectory

The direction of what it costs to win a customer, which usually rises as a category fills.

Read the trajectory, not the snapshot. A single reading is noise; the direction across enough operating history is the signal.

Scale, fix, or kill

Every product gets exactly one of three calls each cycle.

Scale

When the signals trend right and margin holds, increase delivery — more inventory, more advertising, wider coverage — and watch the same signals as volume rises.

Fix

When one signal is off but addressable, diagnose before spending. Work through the candidate causes in order — listing quality and imagery, conversion, ad performance and targeting, traffic channels, pricing, the product itself — and match the fix to the cause you find, not the signal alone. A rising return rate can be the product or an expectation the listing set; climbing acquisition cost can be targeting, or a category filling that no fix reverses.

Kill

When the signals do not improve within a window you defined in advance, stop. Recover what inventory you can and redirect the capital to a product that clears the bar.

Set the kill line before you need it

Knowing when to stop matters as much as knowing how to launch. The catch is that the decision to kill is hardest when it is most needed — once time and capital are committed, the product is easy to defend and hard to abandon.

So set the window before you are emotionally invested. Decide, at launch, how long a product has to reach its signals and what "reached" means. When the window closes, read the four signals against the bar you already set and act on what they say.

Define the exit at the entrance

A kill line chosen in advance is a decision; one chosen in the moment is a rationalization.

Was this page helpful?