Stop Optimizing ACoS in Isolation

Advertising cost of sales (ACoS) is useful, but it answers a narrow question: how much attributed ad revenue did each euro of ad spend produce? It does not tell you whether the resulting orders created profit, supported a launch, or displaced sales that would have happened organically.

Treat ACoS as one diagnostic input. Make the decision on contribution profit, strategic purpose, and the quality of the evidence.

1. Start with the commercial job of the campaign

Two campaigns with the same ACoS can deserve opposite decisions. A mature product defending profitable demand should usually produce cash. A new product may accept weaker short-term efficiency while it gathers search-term evidence, reviews, and organic visibility. A clearance campaign may rationally trade margin for released working capital.

Write down the campaign job before judging performance: harvest profit, launch, defend, learn, or liquidate. Add a time horizon and a stop condition. “Grow sales” is not enough because it gives every result a convenient explanation after the fact.

2. Connect the ad metric to an order-level margin

Build the margin waterfall for the products receiving spend:

The remaining contribution determines how much advertising the product can support. A low-margin product can show an attractive ACoS and still lose money. A high-margin product can tolerate a higher ACoS while creating more euros of contribution.

Use a break-even ACoS as a boundary, not a target. If the contribution margin before ads is 24%, an ACoS near 24% leaves little room for overhead or attribution error. The operating target should sit below break-even with a deliberate buffer.

3. Separate signal from attribution noise

Marketplace attribution is not a perfect causal model. Orders can arrive after earlier clicks, branded demand can be captured by ads, and organic rank can move alongside paid activity. Short windows magnify noise.

Compare at least three views:

1. Campaign efficiency: spend, attributed revenue, conversions, ACoS, and ROAS.
2. Product economics: total product revenue and contribution before and after ads.
3. Incremental direction: whether total units or contribution changed when spend changed.

Do not call one day of improvement a trend. Use enough clicks and conversions to distinguish performance from randomness. For low-volume products, widen the window and change bids in smaller steps.

4. Use a decision matrix, not one universal threshold

This prevents the common mistake of cutting every high-ACoS placement. Some are bad targeting. Others reveal a conversion problem, price problem, or insufficient data. The correct action depends on the mechanism.

5. Run a controlled weekly operating loop

Use the same review order every week:

1. Confirm data freshness and the selected date window.
2. Flag spend with zero conversions and search-term irrelevance.
3. Compare product contribution after ads with the previous period.
4. Identify whether the constraint is traffic, conversion, Buy Box, price, stock, or margin.
5. Make one bounded change with a recorded reason.
6. Set a review date and rollback condition.

Avoid simultaneous bid, budget, price, and listing changes on the same product. When everything moves, attribution becomes impossible and teams repeat weak decisions.

6. Worked example: the efficient campaign that loses money

Consider a product selling for €30 before VAT treatment. After commission, fulfillment, product cost, packaging, and a returns allowance, only €4.50 remains before advertising. That is a 15% contribution margin before ads. A campaign reporting 13% ACoS can look healthy in an advertising dashboard, yet it leaves only about €0.60 before overhead when the attributed revenue and total product economics line up. Small fee or return changes can erase that remainder.

Now compare a second product with €12 contribution before ads on the same €30 revenue. A 22% ACoS costs €6.60 and still leaves €5.40 contribution. The second campaign looks worse by ACoS but creates materially more value per order. It may also have room to scale while the first campaign needs price, sourcing, or fulfillment improvement.

Use the example as a team exercise. Select three campaigns labeled efficient, add current COGS and fee evidence, calculate contribution after ads, and rank them again. Then select three campaigns labeled inefficient and test whether launch purpose, total product growth, or strong margin changes the decision. The objective is not to excuse weak advertising. It is to expose decisions that a single ratio hides.

Questions for the weekly review

• Did contribution after advertising improve in euros, not only percentage?
• Did total product sales change, or only attributed sales?
• Did stock cover, price, or Buy Box state change during the window?
• Which assumption would reverse the decision if wrong?
• What is the smallest next change that can produce useful evidence?

7. Practical checklist

• Campaign purpose and horizon are documented.
• COGS and marketplace fees are current.
• Break-even ACoS includes a safety buffer.
• Decisions use total product contribution, not attributed revenue alone.
• Low-volume campaigns receive longer evaluation windows.
• Every change has an owner, reason, review date, and rollback rule.
• Scaling stops when stock cover or Buy Box stability creates a new constraint.

The goal is not the lowest ACoS. The goal is repeatable profitable growth with evidence strong enough to support the next decision.