ACoS answers one question
Advertising cost of sales shows ad spend relative to attributed sales. It does not, by itself, show whether those sales produced contribution, whether returns changed the result, whether stock can support demand, or whether the campaign is deliberately learning. A low figure can still sit on weak unit economics; a higher figure can be rational during a controlled launch.
Start every campaign group with an objective: protect branded demand, discover queries, launch a product, clear a planned inventory position, or harvest efficient demand. Then define the economic and operational guardrails appropriate to that job.
Create a return-adjusted contribution view
At SKU level, connect attributed net sales with marketplace charges, fulfilment and logistics, expected returns, landed cost, promotions, and ad spend. Keep the attribution method visible. This produces an estimated contribution after advertising rather than treating revenue as the finish line.
Use ranges when input quality is uncertain. A transparent estimate with a documented return assumption is more useful than a precise-looking number built on incomplete fees. Reconcile the model to settlements and confirmed return outcomes as data matures.
- Campaign role and decision owner
- Attributed net sales and stated attribution window
- Return-adjusted contribution estimate
- Inventory cover and replenishment constraint
- Query or placement evidence behind the next action
Review portfolios, not isolated percentages
Group activity by role and product economics. Discovery work should be judged on useful search-term learning as well as immediate sales. Brand defence should be reviewed in the context of organic demand. Mature efficiency campaigns need tighter economic control and clear rules for waste.
Look at changes over time and record why a bid, budget, target, or placement changed. Without an experiment log, teams repeatedly relearn the same lesson and cannot separate seasonality from an actual operating improvement.
Make the weekly scorecard actionable
A practical review highlights material contribution movement, spend without useful evidence, inventory-constrained winners, return-heavy products, and campaigns whose objective no longer matches the business need. Each item ends with an owner, action, guardrail, and review date.
Marketplace reporting definitions and controls evolve. Confirm current field definitions in the relevant advertising console before automating decisions, and keep business rules easy to update when the source data changes.