Marketplace advertising fee structures look harmless on a proposal slide. Five percent of spend. Ten percent of managed media. A fixed retainer. A performance bonus. A software fee plus service hours. Nice and tidy.
Then the account scales, the bol Sponsored Products budget doubles, Amazon CPCs rise, MediaMarkt asks for more retail media support, and suddenly the fee model is making decisions in the background. Not the brand. Not the P&L. The contract.
That is the operator mistake I want to name: treating the agency fee as a procurement line instead of a campaign constraint. For marketplace ads, the fee is not just what you pay the agency. It changes your real break-even ACOS, your minimum ROAS, your willingness to launch, and your ability to pull budget back when margin says “stop”.
This matters most for brands spending €5K+ per month across Amazon, bol and MediaMarkt. At that level, you are usually beyond “please fix my campaigns” and into a more expensive question: which products deserve paid demand after fees, returns, fulfilment, stock risk and agency costs?
Most articles about Amazon PPC agency pricing list the usual models: 10–20% of ad spend, €1,500–€10,000 monthly retainers, hybrids and performance fees. Useful, but incomplete. The missing angle is margin translation. A fee model is only fair if the account can still make money at SKU level after the fee is loaded into the advertising economics.
The four fee models you will actually see
Marketplace advertising services usually come in four shapes. None is automatically good or bad. Each creates a different operating behaviour.
1. Percentage of ad spend
The agency charges a percentage of monthly ad spend, often somewhere between 8% and 20% depending on scope, channel complexity and minimums. It is simple to understand and it scales with workload when the account grows.
The trade-off is incentive risk. If the agency earns more when spend increases, the governance has to be strong enough to separate “we can spend more” from “we should spend more”. That distinction sounds small until a campaign with 28% ACOS sits on a SKU with only 24% contribution margin before agency fees.
2. Fixed monthly retainer
The brand pays a fixed monthly fee regardless of spend. This is cleaner for budgeting and removes the obvious incentive to inflate media spend. It can work very well for stable accounts with a clear SKU set and weekly optimisation rhythm.
The trade-off is scope pressure. A low retainer can quietly become a maintenance package: bid tweaks, a monthly call and a dashboard screenshot. For marketplace accounts, the real work is not just bids. It is SKU margin mapping, search term pruning, stock-aware budget allocation, promotion planning, placement mix, competitor movement and explaining why a campaign with a “good” ROAS should still be cut.
3. Hybrid: retainer plus spend percentage
The agency charges a base retainer plus a smaller percentage of spend. This is common when accounts need both strategic work and media operations. It gives the agency a floor for planning and a variable component when complexity rises.
The trade-off is that the brand can end up paying twice for the same growth. If the retainer already covers weekly management, what exactly does the percentage cover? Extra marketplaces? Extra SKUs? Extra campaigns? More senior attention? The contract should say.
4. Performance fee
The agency earns a bonus based on sales, growth, profit or agreed KPI improvement. In theory, this aligns everyone. In practice, it only works when the baseline, attribution window, margin definition and exclusions are painfully clear.
Marketplace advertising is noisy. Organic rank changes, Buy Box position, promos, price cuts, out-of-stock periods and competitor exits can all move sales. A performance fee based on ad-attributed revenue can reward spend that would have happened organically. A performance fee based on total revenue can reward pricing discounts that hurt margin. A performance fee based on contribution profit is better, but it requires trustworthy data.
The margin-loaded fee test
Here is the test we use mentally before accepting any fee model: load the agency fee back into the SKU economics and recalculate the ad threshold.
For a simple version, use this:
- Contribution margin before ads = selling price minus VAT impact, marketplace commission, fulfilment, COGS, returns, payment/operational costs and expected promo cost.
- Break-even ACOS before agency fee = contribution margin before ads divided by selling price.
- Management-fee-loaded ACOS = ad spend plus agency fee divided by ad-attributed revenue.
- Safe ACOS = break-even ACOS minus the margin buffer you want to keep.
The point is not accounting perfection. The point is to stop pretending that a 25% ACOS is the same under every fee model. It is not.
If your agency charges 12% of ad spend, every €1,000 in media costs €1,120 before you even discuss internal time. A campaign showing 25% platform ACOS behaves like 28% management-fee-loaded ACOS. On a high-margin supplement, that may be fine. On a low-margin electronics accessory with returns, it can turn a “green” campaign red.
Named example 1: bol kitchen bundle with LVB
Take a Dutch kitchen brand selling a bundle on bol for €39.95. After purchase cost, bol commission, LVB fulfilment, expected returns and packaging, the SKU has €10.40 contribution margin before ads. Break-even ACOS is therefore about 26%.
The current bol Sponsored Products campaign spends €6,000 per month and produces €24,000 ad-attributed revenue. Platform ACOS is 25%. Looks acceptable, if you stop there.
Now add a 15% of-spend management fee. The agency fee is €900. Total managed advertising cost becomes €6,900. The loaded ACOS becomes 28.8%.
That changes the decision. The campaign is no longer slightly below break-even; it is above break-even before you even keep a profit buffer. The right move is not “scale because ROAS is 4.0”. The right move is to split the campaign:
- keep exact terms converting below 20% platform ACOS,
- cap category discovery until the SKU has more review depth,
- move budget to the multipack variation with €13.80 margin,
- and set a loaded ACOS alert inside the weekly reporting.
This is where FiveX is useful in the background. When bol ad spend, product margin, LVB costs and SKU-level contribution margin sit in one view, the agency discussion changes from “ACOS is almost target” to “this SKU can only afford €4,950 of this current spend unless conversion improves”. Much better coffee-meeting material.
Named example 2: Amazon launch with a retainer
Now take a Belgian home appliance brand launching on Amazon.nl and Amazon.de. The agency proposes a €3,500 monthly retainer for launch planning, campaign structure, keyword harvesting, weekly optimisation and reporting. Media spend starts at €15,000 per month.
On paper, the retainer is 23.3% of media spend in month one. That sounds expensive compared with a 10% spend fee. But the comparison is too shallow.
The hero SKU sells for €89.99. Contribution margin before ads is €24.30 after referral fee, FBA, COGS and expected returns. Break-even ACOS is 27%. During launch, the brand accepts a temporary 34% platform ACOS for 60 days because the goal is ranking and review velocity.
Month one results:
- ad spend: €15,000,
- ad-attributed revenue: €44,100,
- platform ACOS: 34%,
- retainer: €3,500,
- loaded ACOS: 42%.
If this were a mature SKU, it would be a no. For a launch, it can still be acceptable if the contract defines what must happen by day 60: organic rank movement on the top 20 terms, review count, conversion rate lift, TACoS decline and a clear plan to bring loaded ACOS under 30% by month three.
The retainer is not the problem. The missing sunset clause is the problem. Launch economics need an expiry date. Otherwise “temporary investment mode” becomes the most expensive sentence in the account.
In FiveX, we would want this launch tracked as a separate role: Launch / tolerate loss, not mixed into evergreen profit campaigns. That way the dashboard can show whether the loss is buying future organic sales or simply renting expensive clicks.
Named example 3: MediaMarkt electronics and the hybrid trap
Imagine an electronics accessories seller advertising on MediaMarkt Marketplace with €22,000 monthly retail media spend. The agency fee is €2,000 plus 7% of spend. Total fee: €3,540.
The account-level ROAS is 5.2×. Many teams would be happy. But the SKU mix tells a different story:
- USB-C hub: €49.99 price, €8.20 contribution margin before ads, 18% return-adjusted break-even ACOS.
- Premium docking station: €139.99 price, €38.50 contribution margin before ads, 27.5% break-even ACOS.
- Replacement charger: €24.99 price, €3.10 contribution margin before ads, 12.4% break-even ACOS.
The replacement charger is consuming €6,500 of spend at 19% ACOS. Platform reporting says it is “not terrible”. Margin-loaded reporting says it is donating money to the marketplace. The docking station, meanwhile, is capped by budget at 21% ACOS and has 42 days of stock cover.
This is the hybrid trap. The account looks efficient, the agency fee looks reasonable, but budget is trapped in the wrong products. A fee model that does not require SKU-level budget movement can look fair while still producing poor profit.
What competitors usually cover — and what they miss
Most agency-pricing content covers the menu: flat fee, percentage of spend, hybrid, performance. Some stronger guides explain incentive risk. A few mention TACoS and break-even ACOS. That is all useful.
What is usually missing is the marketplace operator layer:
- Fee model by SKU margin band. A 12% spend fee behaves differently on a 42% gross-margin beauty item than on a 14% margin electronics accessory.
- Fee model by campaign role. Launch, defence, ranking, clearance and profit harvesting should not be judged with the same fee tolerance.
- Fee model by stock cover. Paying management fees to accelerate a SKU with 11 days of stock left is rarely clever.
- Fee model by channel. Amazon, bol and MediaMarkt do not have identical fulfilment economics, ad inventory, attribution or organic spillover.
- Fee model by decision rights. Who can increase spend? Who can pause a margin-negative campaign? Who signs off on launch losses?
That is the angle FiveX can own: not “what should an agency cost?”, but “which fee structure still lets this marketplace account make rational profit decisions?”
How to choose the right model by spend level
€5K–€15K monthly spend
At this level, a percentage-of-spend model can be acceptable if there is a sensible minimum and clear deliverables. But watch the loaded economics carefully. A €5,000 spend account with a 15% fee adds €750. If the account has only three meaningful SKUs, that fee should buy more than generic optimisation.
Best fit: fixed retainer or capped spend percentage, with SKU-margin setup included.
€15K–€50K monthly spend
This is the danger zone for lazy fee structures. The account is big enough for fees to matter, but not always big enough to absorb strategic waste. A 12% fee on €40,000 spend is €4,800 per month. That fee should come with weekly actions, search term governance, budget reallocation, SKU-level margin reporting and channel-specific recommendations.
Best fit: retainer plus capped variable fee, or retainer with defined service tiers. Add performance bonus only if contribution profit is measurable.
€50K+ monthly spend
At this point, the agency should behave like an operating partner. The fee should not rise endlessly just because media spend rises. Complexity does rise, yes, but not always linearly. Managing €100,000 across 80 SKUs is different from managing €100,000 across 8 SKUs.
Best fit: strategic retainer with channel/SKU complexity bands, optional performance component tied to contribution profit, and explicit spend-approval rules.
The seven contract questions I would ask before signing
- Do you report platform ACOS or loaded ACOS including your fee? If they only report platform ACOS, you will need to do the real math yourself.
- Can we set break-even ACOS by SKU, not one account target? One target across all products is dashboard theatre.
- What happens when you recommend increasing spend? The answer should include margin, stock cover and expected incremental sales, not only ROAS.
- How are launch campaigns separated from profit campaigns? Temporary loss-making is fine. Unlabelled loss-making is not.
- Is the variable fee capped? A cap protects both sides when spend scales faster than workload.
- What is excluded from performance calculations? Promotions, price cuts, out-of-stock periods, retail events and branded search should be handled explicitly.
- Who owns the weekly decision log? Every budget increase, pause and target change should connect to a reason the commercial team understands.
The FiveX view: fee structure is a margin control
For our Advertentie Service clients, the fee conversation belongs next to the campaign conversation. Not after it. Not hidden in procurement. Next to it.
Why? Because marketplace advertising management only works when the operator can see the full chain:
- ad spend by marketplace, campaign, target and SKU,
- product contribution margin after fees, fulfilment and returns,
- stock cover and replenishment risk,
- break-even ACOS and ROAS by product,
- TACoS and organic sales movement,
- and the weekly action label: scale, hold, fix, harvest or stop.
FiveX connects those pieces so the agency fee becomes part of the operating model instead of a surprise outside the dashboard. That does not mean every client needs the same contract. It means the contract should make profitable decisions easier, not harder.
My practical stance: if your agency fee model encourages more spend without requiring more proof, tighten it. If it rewards profit but cannot define profit, rewrite it. If it looks cheap but gives you no SKU-margin visibility, it is not cheap. It is just deferred learning.
A simple decision rule
Choose the fee structure that gives you the best answers to three questions:
- Can we see whether this SKU is still profitable after media and management cost?
- Can we reduce spend without creating a commercial conflict?
- Can we scale the right products faster than the wrong ones?
If the answer is yes, the fee model is probably workable. If the answer is no, the cheapest proposal may become the most expensive one in the P&L.
Marketplace ads are not won by paying the lowest fee. They are won by buying the right decisions. Slightly less glamorous, I know. Much better for the bank account.