A retainer is the better default below about $50,000 in monthly profit, and revenue share only becomes fair above it. Choose by asking which number you need moved. If the job is fixing conversion and listings, pay a fixed fee. If the job is scaling proven demand, share the upside.
The short version
- Diagnose first, price second. The model that fits depends on what is broken, not on which invoice looks smaller.
- Retainers fund repair work. Listing rebuilds, image testing and catalog cleanup produce no revenue for weeks.
- Revenue share funds acceleration. It only pays when demand already exists to accelerate.
- We use revenue share above $50,000 per month in profit, at 10 to 20 percent with no fixed fee under it.
- Ask for the launch ACoS number and the maturity ACoS number. A single target across a whole catalog means nobody is managing by stage.
The mistake that costs the most
The expensive error is choosing the fee structure before diagnosing the account. I watch sellers pick revenue share because it feels safe, then discover the first ninety days of work are pure repair: a primary image that does not earn the click, bullets written for a different buyer, a variation family split across three parent ASINs. None of that lifts revenue quickly. Under revenue share the provider earns almost nothing during the exact period when the most valuable work happens, so the honest ones decline the engagement and the rest quietly skip the repair and pour money into ads instead.
The reverse error is cheaper but still real. Paying a flat fee on a mature, stable, profitable catalog where the only remaining job is buying more traffic can leave value on the table, because you carry all the downside of an ad experiment and share none of the upside.
Diagnostic: match the symptom to the model
| Symptom in the account | What it usually means | Which model fits |
|---|---|---|
| Traffic is fine, conversion is weak | Listing, images or price are the constraint | Retainer. The work is repair, not spend |
| Conversion is healthy, sessions are flat | Keyword coverage and channel gaps | Either. Retainer if the catalog needs restructuring first |
| Sales grow, profit does not | Ad efficiency and fee structure | Retainer, with a profit KPI attached |
| Profitable, stable, above the threshold | The job is acceleration | Revenue share on profit, capped |
| Brand new catalog, no history | There is nothing to share yet | Retainer, always |
Work down that table with your own account open. Most sellers who arrive convinced they need a performance deal are actually in row one, where a percentage of revenue would pay for something the account does not need.
The ACoS question that reveals how a provider thinks
Ask any candidate for two numbers: the ACoS target they run at launch and the ACoS target they run at maturity. If both answers are the same, or if the answer is a single company wide figure, the account will be managed by one rule regardless of what each product needs. A launching product is buying rank, review velocity and data, so the acceptable number is aggressive. A mature product with organic position is buying incremental units, so the number should be efficient. Flapen sets that target per product stage rather than per client, which is one of the reasons a flat fee works for us: our income does not change when we tell you to cut spend.
Where each model breaks
Revenue share
It breaks during a rebuild, it invites attribution arguments over whether a sale was organic, paid, or driven by a promotion, and it can quietly encourage discounting, because a lower price moves volume even when it destroys margin. Fix those by defining the shared number as profit rather than revenue, agreeing the calculation in writing before work starts, and setting a floor below which nothing is shared.
Retainer
It breaks when the fee is disconnected from capacity. A fixed number tells you nothing about how many accounts the operator is carrying or whether the work is done in house. Fix that by asking for a named operator, a written reporting cadence, and short notice terms so you can leave if the service thins out.
What a proposal will not tell you about either model
Both structures work fine when the operator has time and both fail identically when they do not. The fee is a symptom, capacity is the cause. That is why I would rather you spend the negotiation asking about headcount, brand load and reporting than shaving the monthly number.
The other unspoken part: revenue share is usually pitched as the agency taking on risk, but in most versions the risk is asymmetric. They earn nothing in a bad month and a great deal in a good one, while you carry the inventory, the ad budget and the returns in both. Genuine shared risk means the downside is shared too, and very few proposals define what that looks like.
Related answers
- Alternatives to percentage of sales pricing on Amazon
- Which pricing suits a new Amazon brand
- Amazon account management pricing vs performance
- Month to month vs annual Amazon contracts
- Hiring an Amazon agency: the complete guide
If you want the two ACoS numbers for your own catalog, ask Flapen.
