Ask six: what counts as performance, measured from which baseline, attributed how, paid on revenue or profit, what happens in a month with no growth, and what would make you tell me to stop selling a product. The last one separates a partner from a vendor.
The short version
- Define performance before you agree a percentage. The definition is worth more than the rate.
- A baseline you did not agree is a baseline you will dispute. Twelve months of history, signed off by both sides.
- Growth you would have had anyway is not performance. Strip the existing trend out of the calculation.
- The kill question is the real test. Ask what would make them recommend discontinuing a product.
- Every performance deal needs a stop clause. Notice period, trailing payments, and what you keep.
The six questions, and what a weak answer sounds like
You are looking at a proposal that ties the agency's income to results. That structure can be excellent, and it can also be a way to make an ordinary retainer sound brave. These six questions tell the two apart in a single call.
| Question | Weak answer | Answer worth hearing |
|---|---|---|
| What counts as performance? | "Growth" | A named metric, with the report it comes from |
| From which baseline? | "Where you are now" | Twelve months of monthly data, agreed in writing |
| How is it attributed? | "Our dashboard shows it" | A rule you can apply yourself from your own reports |
| Revenue or profit? | Whichever sounds better in the moment | A stated basis, with the deductions listed |
| What happens in a flat month? | Silence, or a minimum charge appears | A clear statement, floor or no floor |
| What would make you tell me to stop? | "We do not give up on products" | Named criteria and a defined window |
Diagnosing the clause you are actually being offered
Performance deals fail in predictable ways. Match the language in the proposal to what it usually produces.
Symptom: performance is measured on gross revenue. The cause is that revenue is the easiest number to move and the easiest to verify. What it produces is discounting, aggressive advertising, and volume at thin margin, all of which raise the fee while lowering your profit. The fix is to define the share on profit, then spend the extra hour agreeing which costs are deducted.
Symptom: the baseline is "current run rate". The cause is convenience. What it produces is a fee charged on seasonality and on the growth trend the brand already had. The fix is twelve months of monthly history, and a written rule for how a seasonal peak is treated.
Symptom: attribution comes from a proprietary dashboard. The cause is that attribution is hard. What it produces is an argument in month five that neither side can win. The fix is a rule you can reproduce from your own Seller Central reports without anyone's software.
Symptom: no stop condition anywhere in the document. The cause is that stop conditions are uncomfortable to write during a sales process. What it produces is spending that continues long past the point anyone believes in it. This is the expensive one, and it deserves its own section.
The question that matters most
Early on, I poured money into a failing product for three months, hoping advertising would turn it around. It did not. The rating was drifting down, returns were higher than the category, and I kept funding traffic into a page that was converting badly because stopping felt like admitting the product choice was wrong.
That episode is where our kill criteria came from. We now write down, before a product is scaled, what would make us recommend stopping: rating trend, return rate, conversion rate, and acquisition cost trajectory, each over a defined window rather than judged by mood.
Ask any agency proposing a performance deal what would make them tell you to kill a product. The structure of a performance arrangement makes this question sharper, because their income depends on the product continuing to sell. If they have criteria and can name them, the incentive problem is under control. If the answer is enthusiasm, you have learned the most important thing available about how the next twelve months will go.
What most agencies will not tell you before you sign
Performance deals are often proposed because the agency wants your account and your budget cannot support a retainer. That is a reasonable commercial position, and it means the arrangement is designed around their revenue problem rather than your growth problem. Ask directly whether a flat fee is available and what it would be. The comparison tells you whether the performance framing was chosen for alignment or for the sale.
The second thing: trailing commission. Many agreements continue paying a share of sales for months after termination, on the argument that the growth was created during the engagement. That clause is rarely mentioned on a call and it is where most of the regret in this industry sits. Read the termination section before the pricing section. Ours is month to month with 30 days' notice, nothing trails afterwards, and on exit you keep the Seller Central account, the campaigns, the creative, and a written handover.
Related answers
- Questions to ask before hiring an Amazon agency
- Typical retainer vs performance-based for Amazon management
- Amazon agency revenue share models explained
- Who offers no-win-no-fee Amazon growth services
- Amazon agency pricing and economics: the complete guide
Ask us the kill question on the first call, and the answer is published anyway at Flapen.
