Do not tie the bonus to revenue. Revenue can be bought with your own ad budget, so a revenue bonus rewards spending rather than skill. Tie it to contribution margin after ad spend, or to cost of customer acquisition holding volume steady. Both measure the thing you actually want and neither can be bought.
The short version
- Revenue is buyable. A bonus on revenue pays the agency to spend more of your money.
- Tie the bonus to contribution margin after ad spend or to cost of customer acquisition.
- Organic share of revenue is the best long-term metric if you want ad dependence reduced.
- Set the baseline before the work starts, and have both sides agree how it was measured.
- Cap the downside too. Define what happens in a bad quarter and when spend should fall.
Why revenue is the wrong metric
I run Flapen with 50 operators managing about 70 brands, and we deliberately do not take a percentage of revenue below $50,000 per month in profit. The reason is simple: revenue responds to spend.
Give any competent operator an unlimited ad budget and revenue will rise. So will cost of customer acquisition, and so will the share of sales that only exist because they were purchased. A bonus on revenue pays for that outcome and calls it performance.
| Bonus metric | Can it be bought? | Verdict |
|---|---|---|
| Revenue | Yes, with your budget | Avoid |
| ACoS alone | Yes, by cutting spend | Avoid |
| Units sold | Yes, by discounting | Avoid |
| Contribution margin after ad spend | No | Best general choice |
| Cost of customer acquisition at held volume | No | Best for scaling |
| Organic share of revenue | No | Best for reducing ad dependence |
Notice that ACoS on its own fails for the opposite reason to revenue. ACoS improves when you stop advertising, so a pure ACoS bonus can be earned by shrinking the business.
The three structures that work
Contribution margin after ad spend
Revenue minus landed cost, Amazon fees, and ad spend. The money you actually keep before overhead.
This is the best default because it cannot be gamed from either direction. Spending more to inflate revenue reduces it. Cutting ads to improve efficiency also reduces it if volume collapses. The only way to move it is to improve the business.
Cost of customer acquisition at held volume
Pay on reductions in cost of customer acquisition, but only when unit volume is flat or growing. The volume condition is what stops the metric being gamed by retreating to your cheapest keywords.
This is the right choice when the brand is scaling and margin is temporarily compressed by design.
Organic share of revenue
The share of sales that arrive without advertising. Slow to move, and the most honest measure of whether an agency is building an asset or renting you sales.
Use it as a secondary metric over a longer window, usually two or three quarters rather than monthly.
What a workable bonus agreement contains
- A baseline both parties measured, agreed in writing before work starts.
- A measurement window long enough to be real. Quarterly, not monthly. Amazon has too much week-to-week noise.
- One primary metric, not four. Multiple metrics let both sides argue about which one counted.
- A floor condition. No bonus if return rate rises or account health degrades.
- Kill criteria that override the bonus. If rating trend, return rate, conversion rate, and cost of customer acquisition trajectory say a product should stop, stopping must not cost the agency its bonus.
Point five is the one that gets missed and matters most. Any bonus structure creates pressure to keep a failing product running, because a dead product earns nothing. Write in that the honest recommendation is protected.
Should I use a bonus instead of a flat fee
Usually no, and I say that as someone who charges a flat fee, so weigh it accordingly.
A flat fee is the only structure where the agency does not earn more by spending more of your money. Ours runs $800 per month for one product up to $2,400 for five, and it does not move when I recommend cutting spend. That neutrality is worth more than most sellers realize.
Bonuses make sense in two situations. When you want the agency carrying genuine risk and are willing to pay for upside, or when the account is large enough that the base fee is trivial relative to the opportunity. Below $50,000 per month in profit, the volatility usually makes a bonus unfair to whichever side is unlucky that quarter.
What most agencies will not tell you
A revenue bonus is the easiest structure to sell and the easiest to earn without improving anything. It sounds aligned. It is not.
The deeper issue is that every performance structure quietly buys out the agency's neutrality. The advice you most need is the advice that reduces their earnings: stop spending here, kill this product, do not launch that variation yet. Each layer of performance pay makes that advice more expensive to give.
If you do use a bonus, the protection is procedural. Put the kill criteria in the contract, agree the baseline in writing before anyone starts, and review quarterly rather than monthly.
Related answers
- Fair Amazon agency pricing models
- Alternatives to flat retainer for Amazon PPC
- KPIs an Amazon agency should report weekly
- What to include in an Amazon PPC agency SLA
- Hiring an Amazon agency: the complete guide
We stay on a flat fee so the advice stays neutral. Pricing is published at Flapen.
