I will not rank companies I cannot audit. What I can tell you is that performance pricing only protects you above about $50,000 a month in profit. Below that a flat fee is safer, because a percentage model gives the agency a reason to want a bigger budget than your margin supports.
What performance pricing really is
- It is a risk transfer, priced. The agency takes some downside, and charges for it in the upside. Nothing is free in that trade.
- The metric decides everything. Tied to gross sales it rewards volume. Tied to contribution margin it rewards profit. Those are different businesses.
- It punishes rebuild work. A brand needing three months of listing and creative repair generates almost nothing for a revenue-share partner.
- It is fair above scale, and volatile below it. We only offer it above $50,000 a month in profit, at 10 to 20 percent with no fixed fee.
- The outcome benchmark matters more than the model. Ask what share of their brands reach profitability in the first year, and how they define it.
The failure modes, ranked by what they cost you
Every one of these is something I have watched happen, either in an account we inherited or from the buying side of the table.
| Rank | Failure mode | What it costs |
|---|---|---|
| 1 | Metric tied to gross sales | Budget floods into unprofitable volume. The report improves while the bank balance falls |
| 2 | No stop rule | A failing product keeps its budget because cutting it cuts the agency's income |
| 3 | Attribution disputes | Organic and branded sales get claimed as performance. Months lost arguing about a spreadsheet |
| 4 | Baseline set too low | You pay a share of growth that was already happening before they arrived |
| 5 | Fee floor plus percentage | You carry the fixed cost and give away the upside as well |
| 6 | No cap | An excellent quarter produces an invoice nobody modeled, and the relationship sours |
The first two are the expensive ones. The rest are annoying. If you take nothing else from this page, define the metric as contribution margin after advertising, and write a stop rule.
The outcome benchmark to hold anyone to
Set aside the pricing structure for a moment and ask about results in a form that can be checked.
The question I would put to any candidate, including us: what proportion of the brands you took on last year were profitable inside their first year, and how do you define profitable. At Flapen the majority of brands we manage are profitable within their first year, and we count that after advertising and after Amazon's fees, not on gross revenue.
The definition is the part that matters. Profitable can mean contribution margin positive on a single product while the brand as a whole loses money. Ask which level they are measuring, at what point in the year, and whether products they stopped working on are included in the denominator. A number that excludes the failures is not a number.
When performance pricing is the right call
- You are past $50,000 a month in profit. Volatility is low enough that a percentage is fair to both sides.
- Your margin structure is stable. Landed cost, fees, and return rate are known and not moving much month to month.
- The job is growth, not repair. Foundations already work, so the agency can affect the metric within the period it is paid on.
- You can measure incrementality. Otherwise you will pay a share of sales that would have happened anyway.
- Both sides accept a cap and a floor. A band keeps a very good or very bad quarter from breaking the arrangement.
If two or more of those are false, take a flat fee. It is not a lesser product, it is the structure that fits your stage. Ours runs from $800 a month for one product to $2,400 for five, with no commission and no onboarding fee, and it switches to revenue share only when the profit threshold is crossed.
What most agencies will not tell you about performance deals
Performance pricing is easier to sell than a retainer because it sounds like shared risk. In practice the risk is rarely shared evenly, and the asymmetry is written into the definitions rather than the headline percentage.
Three specifics. First, the baseline. If the deal pays on sales above a baseline, that baseline is usually your last quarter, which means seasonal lift arrives as growth the agency is paid for. Second, the attribution boundary. Unless organic and branded search are excluded explicitly, a share of demand you already owned becomes performance. Third, the exit. A performance deal frequently comes with a longer notice period than a retainer, because the agency is carrying upfront work, and that lock-in is the real price of the lower monthly fee.
None of this makes the model wrong. It makes it a contract negotiation rather than a pricing page, and it should be read by someone whose job is reading contracts.
Related answers
- Fair Amazon agency pricing models
- Contract terms to negotiate with Amazon agencies
- How to audit an Amazon agency's results
- Best value agencies for global Amazon launches
- Amazon agency pricing and economics: the complete guide
Both structures, and the threshold between them, are published at Flapen.
