The Amazon paradox: Growth that pays off

Amazon can help brands reach demand, support purchase decisions and generate sales. Whether those sales make good business sense depends on how the product range, product communication, costs and advertising work together. A sound Amazon strategy connects these decisions.

The campaigns are running. Orders are increasing. Yet there is less left at the end of the month than expected.

Imagine a kitchenware brand promoting its food storage containers more heavily on Amazon. An inexpensive entry-level product sells well but incurs relatively high shipping costs. A larger set has more room for a healthy margin, yet the product page barely explains it. Meanwhile, additional stock ties up cash as the marketing team celebrates the increase in revenue.

This is a hypothetical example of the Amazon paradox: A sales channel can grow while its financial contribution falls short of expectations.

Addressing it starts with a shared question for marketing, sales and operations: Which products do we want to sell through Amazon, what value do we need to communicate – and what needs to remain after the sale?

Give Amazon a clear role in your sales strategy

For one brand, Amazon may be where new customers discover a product. For another, the platform primarily captures existing demand or makes regularly needed products easy to reorder. A general market-share figure tells us little about which role makes sense.

We recommend starting with your audiences' actual paths to purchase. Where do interest and trust develop? Where do people compare products? Where do they decide to buy? And what information do they expect at each stage?

A social campaign might generate interest in our kitchenware brand's containers. Some prospects will then visit its own shop; others will search for the product on Amazon. The brand needs suitable data to understand how often this happens and where purchases take place. It does not follow that the entire range must be available on every marketplace.

A considered selection can help align the offer with the effort involved. Entry-level products that are easy to explain, financially viable sets and selected repeat-purchase items can serve different purposes. The brand's own shop can provide additional room for advice, other product categories or special services.

Clarifying these roles is particularly relevant for brands with standardised products that can be sold online. A B2B offering requiring extensive consultation calls for a different sales decision from a consumer product. That is another reason why universal claims such as “Every brand loses without Amazon” fall short.

Build a transparent calculation for each product

Revenue, contribution margin and liquidity answer different questions. Revenue shows how much has been sold. Contribution margin shows what remains after the costs included in the calculation. Liquidity describes whether payments can be made when they fall due.

For assortment decisions, we suggest a calculation for each product or meaningful product group: actual net revenue after discounts and refunds, less cost of goods, selling and fulfilment fees, attributable logistics and advertising. Returns must be treated consistently so that refunds and inventory losses are not deducted twice. The remaining contribution must also cover the other business and channel costs.

General fee percentages cannot replace this calculation. Amazon distinguishes between selling fees, costs for the chosen fulfilment method and other potential charges. Category, dimensions, weight, storage and the programmes used affect the total. The applicable market and account terms are what matter. Amazon's German pricing overview illustrates these differences.

For our kitchenware brand, this could lead to four different decisions:

This makes it easier to identify which products can support more demand and which need another issue addressed first.

Include operational quality in growth planning

An additional order depends on stock being available and delivery being reliable. Purchasing, inventory, lead times and marketing need to work together. A promotion that generates orders the team cannot fulfil creates a different problem from a product with too much stock.

Regular reviews should therefore include discrepancies. Do the fees charged match the recorded product data? Can stock movements and reimbursements be reconciled? Which return reasons keep recurring? Potential claims should be assessed against the applicable terms and supporting evidence. This does not justify promising a fixed reimbursement rate.

The same applies to logistics models. A different transport or storage arrangement may be worth considering if the total effort, delivery capability and capital tied up suit the brand. One lower freight rate says little about that overall picture. We would compare options across the entire journey and involve the people responsible for operations in the decision.

Advertising payments also belong in liquidity planning. In its April 2026 announcement, Amazon described different payment methods and a planned change for a limited group of advertisers who were contacted directly. Your planning should reflect the account settings and payment terms that actually apply. The announcement explicitly defines the affected group.

Product knowledge becomes part of AI shopping assistance

The clarity of product descriptions now also matters for conversational shopping tools. Amazon explains that Rufus in Germany and Austria draws on sources including product information, reviews and community questions. In the US, Amazon has brought Rufus and Alexa+ together under “Alexa for Shopping”. Features and rollouts need to be considered by market. Rufus in Germany and Austria, Alexa for Shopping in the US.

For our kitchenware brand, the practical task would be to answer relevant product questions reliably. Which container sizes does the set include? Which parts can go in the dishwasher? Under what conditions are the containers leakproof? Can customers buy replacement lids?

The answers belong in maintained product data, clear copy and appropriate images. A material specification, a size comparison and a recognisable use situation each serve a different purpose. Together, they help people understand the offer.

Amazon's COSMO research explores how shopping behaviour and other signals can be used to derive knowledge about potential usage intentions. The paper describes applications such as search navigation. It does not provide a simple ranking formula that brands can satisfy by adding a few keywords. The COSMO publication describes the approach and its scope.

Our conclusion for content work is to organise real use situations and substantiated product features systematically. The product page, Brand Store, own shop and campaign should communicate the same reliable product knowledge. This provides a better information base for people and machine processing; it does not promise a particular AI recommendation or search position.

Measure media against its purpose

Advertising can introduce a product, communicate a specific reason to buy or capture existing interest. These roles should be defined before the campaign so that the team can interpret its results meaningfully.

ACOS relates advertising spend to ad-attributed revenue. The ratio commonly called TACOS relates advertising spend to the total Amazon revenue of the entity being measured. Both can support decisions. The costs, products and periods included must be defined consistently. Neither ratio alone establishes how many additional sales advertising caused. Amazon explains the different reference points for ACOS and TACOS.

The Search Query Performance dashboard adds another perspective. For search queries, it shows measures including impressions, clicks, basket additions and purchases, as well as the brand's share of the respective search performance. Amazon describes its scope in this introduction to Brand Analytics.

For the kitchenware brand, this could reveal searches where the set attracts attention but is rarely purchased. That can prompt questions about product fit, price and presentation. The data does not automatically establish how many purchases would have happened without advertising.

A strong organic position is therefore a reason to examine advertising's contribution. It does not by itself prove cannibalisation. Before making a budget decision, we recommend testing a specific change with an appropriate comparison and considering the overall result, advertising costs and contribution margin together. Offers, availability and seasonal demand need to be taken into account.

For further questions about audiences and advertising contacts, Amazon Marketing Cloud can enable additional analysis. The strategic task remains the same: make a reasoned decision and explain how strong the supporting measurement is.

Creative Engineering connects the decisions

When marketing looks only at revenue, finance only at costs and the content team only at the finished product page, there is no shared basis for the next decision.

Our Creative Engineering approach starts with that connection. Strategy defines the role of the channel and the product. Creative work makes the value understandable and the brand recognisable. Technology supports consistent product data, delivery and analysis. The team connects the results with the financial and operational conditions.

For the kitchenware brand, this could become a clearly scoped project: explain the financially promising set more clearly, ensure supply and test a suitable media hypothesis. Before starting, the team agrees which costs to include and when to decide whether to expand or adjust the work.

AI can help group product questions, find contradictions in descriptions or prepare initial copy and visual concepts. Expert review and brand direction remain part of the work. We take responsibility for concept and quality.

As a starting point, we suggest four concrete outputs: a defined role for Amazon in sales, a transparent assessment of selected products' economics, prioritised improvements to product communication and a measurement plan for the next marketing decision.

This connects Performance Marketing and Content with the needs of your business. If you want to develop your Amazon strategy, we work with the people responsible in your team to identify the next useful step.

Discuss your Amazon strategy