Marketplace ad accounts rarely lose discipline because every campaign is terrible. They lose discipline because one campaign looks wonderfully, dangerously good.
A hero SKU starts converting. Amazon Sponsored Products sits at 18% ACOS. bol Sponsored Products keeps winning high-intent search terms. MediaMarkt retail media finally proves that electronics shoppers will click when the offer is sharp. The weekly report looks calm: revenue up, ROAS up, spend fully used. Naturally, someone says the sentence that should make every operator sit up a little straighter: “Let’s put more budget behind the winner.”
Sometimes that is exactly right. Often it is only half right.
The named mistake I see in managed marketplace advertising is letting the best-looking SKU become the budget landlord. It collects the rent from every other campaign. It gets the scale budget, the emergency budget, the “we found some room” budget and eventually the budget that was supposed to test the next product. Then a stock delay, price change, Buy Box wobble, review dip or return spike turns one good campaign into an account-level dependency.
My stance: every Amazon, bol and MediaMarkt advertising service above roughly €5K monthly spend needs a marketplace ad concentration cap. Not to punish winners. To stop one winner from hiding portfolio risk. The job is to scale what works while protecting margin, stock, learning capacity and channel resilience.
This is especially important for NL and BE brands using a specialist marketplace advertising service. Once your operator manages budget across multiple marketplaces, the question is no longer “which campaign has the best ROAS?” It is “how much of this month’s commercial risk are we comfortable placing on this SKU, this marketplace and this campaign role?”
What competitor advice gets right
The best marketplace advertising advice has moved beyond basic bid tweaks. That is good news.
Podean explains the bigger retail media context well. Their solutions page talks about integrating non-media signals such as out-of-stock data, low stock levels, price changes and merchandising into media strategy. Their hidden-costs piece is also right to warn that top-selling ASINs can be weak profit contributors once returns, storage, promotions and contribution profit are included.
BidX is strong on automation mechanics. Their positioning focuses on automated keyword, bid and budget optimisation across Amazon and Walmart, with AI agents surfacing campaign decisions and execution opportunities. That matters because manual account management simply cannot keep up once campaigns, marketplaces and formats multiply.
Skai frames retail media budget allocation as a real strategic problem rather than a spreadsheet chore. With many retail media networks fighting for spend, budget allocation has to weigh ROI, measurement quality, competitive placements and strategic priorities. Emplicit’s Amazon PPC budget tools comparison adds a useful operational layer: native Amazon controls help at day level, while third-party platforms add pacing, alerts, placement controls and reporting.
Seller forums and YouTube advice tend to bring the practical pain. Operators complain about rising CPCs, target ACOS rules that ignore TACOS, campaigns running out of budget too early and agencies that report performance without explaining whether spend is building long-term demand or just harvesting existing buyers.
All of that is useful. The gap is concentration. Most advice tells you how to move budget toward the winner. Less advice tells you when the winner has become too large a share of account risk.
The concentration problem: ROAS can improve while resilience gets worse
Imagine a €12,000 monthly marketplace ad budget split across Amazon, bol and MediaMarkt. At the start of the month the plan is sensible:
- €5,500 for Amazon Sponsored Products across protect, scale and learn campaigns.
- €4,000 for bol Sponsored Products across hero SKUs and category expansion.
- €2,500 for MediaMarkt retail media around electronics visibility and conversion tests.
By week three, one Amazon monitor stand is clearly outperforming the account. It spends €3,800, generates €19,000 attributed revenue and holds 20% ACOS against a loaded break-even ACOS of 31%. The operator shifts another €1,500 from slower tests into that SKU. On paper, the account improves. Blended ACOS drops. Revenue rises. The Monday deck looks tidy.
But the account also changed shape. One SKU now carries 44% of monthly spend, 52% of attributed revenue and 61% of measured contribution profit. Stock cover is down to 18 days. The supplier lead time is 42 days. The SKU has a return rate that usually matures after 21 days, not seven. The ad dashboard says scale. The portfolio says breathe.
A concentration cap makes that tension visible. It does not say “never let a winner get big”. It says a winner needs permission to become big.
What a marketplace ad concentration cap actually controls
A concentration cap is a rule that limits how much spend, profit dependency or learning capacity one object may absorb before a human reviews the trade-off. That object can be a SKU, product family, marketplace, campaign role, keyword cluster, placement or ad format.
For managed marketplace advertising, I like five practical caps.
1. SKU spend concentration
No single SKU should exceed a set share of total marketplace ad spend without a review. For many €5K to €25K accounts, a starting cap of 25% to 35% is sensible. Higher is possible for narrow catalogues, launches or seasonal pushes, but only with explicit approval.
The review should check contribution margin, expected returns, stock cover, offer strength, organic rank and whether the campaign is defending existing demand or creating new demand.
2. SKU profit dependency
Spend share is not enough. A SKU can use 22% of spend but generate 58% of contribution profit. That sounds lovely until the SKU goes out of stock, loses delivery promise or gets hit by a competitor price move.
A profit dependency cap asks: if this SKU had to pause tomorrow, how much of the account’s expected contribution would disappear? Once one SKU crosses 40% to 50% of paid contribution, the operator should actively build a second and third profit lane.
3. Marketplace concentration
Amazon may look more efficient than bol this week. bol may have better margin because fulfilment economics are cleaner. MediaMarkt may have lower volume but valuable category credibility. A concentration cap stops the team from turning a temporary platform advantage into a permanent budget habit.
For NL/BE teams, I would review any marketplace that takes more than 60% of total spend for two consecutive weeks unless the account strategy deliberately says Amazon, bol or MediaMarkt should carry that month.
4. Campaign-role concentration
Protect campaigns usually look efficient because branded and defensive demand converts well. Learn campaigns look messy because they are buying uncertainty. If an agency only follows ROAS, protect campaigns slowly eat the learning budget. The account becomes profitable today and weaker next quarter.
A role cap reserves budget for protect, scale, learn and fix work. The exact split depends on the account, but the important rule is simple: do not let a defensive winner consume the evidence budget needed to find the next winner.
5. Operational concentration
Some SKUs are not allowed to absorb more demand even if ads look excellent. Low stock, fragile packaging, high support burden, pending content changes, supplier risk or a promotion ending soon can all create an operational cap.
This is where FiveX is useful in a very practical way. Marketplace analytics and inventory insights sit next to advertising data, so an ad decision can see stock cover, product profitability and operational signals before automation releases more budget.
Three named examples: when the cap changes the decision
Example 1: LumaDesk and the Amazon monitor-arm winner
LumaDesk is spending €18,000 per month across Amazon, bol and MediaMarkt. One Amazon monitor-arm SKU is the obvious winner: €6,200 spend, €31,400 attributed revenue, 19.7% ACOS and €5.10 contribution margin per unit after ads at the current CPC.
Without a cap, the agency would move another €2,000 into the campaign. With a cap, the operator sees that the SKU already holds 34% of total spend and 57% of paid contribution. Stock cover is 16 days, while inbound replenishment is 29 days away. The decision changes from “scale” to “hold at €290 per day, move €900 into the second monitor-arm size, and use FiveX stock alerts to reopen scale when cover returns above 28 days.”
The winner still wins. It just stops borrowing demand from inventory it does not have.
Example 2: NoaHome and the bol air-fryer dependency
NoaHome runs €7,500 monthly spend, mostly on bol Sponsored Products. A 6-litre air fryer has become the account darling. It uses €2,850 spend, drives €14,900 attributed revenue and sits at 19% ACOS against a loaded break-even ACOS of 28%.
The cap review adds context. The SKU has 24% of spend but 63% of paid profit. The branded shelf and category terms are efficient, but generic “air fryer aanbieding” clicks have a 17% return expectation and lower basket quality. If the operator simply scales the blended winner, NoaHome becomes dependent on one product during a promotion-heavy category week.
The concentration-cap decision: keep branded and high-intent category campaigns live, cap generic terms at €45 per day, reserve €1,100 for two smaller kitchen SKUs with 32% break-even ACOS and feed the search-term evidence into FiveX advertising automation rules. The account gives up some easy revenue, but it buys a healthier product mix.
Example 3: VeloCase and MediaMarkt category credibility
VeloCase sells phone and tablet accessories. Amazon produces the biggest volume, bol has the strongest margin and MediaMarkt is strategically important because shoppers trust it for electronics. In August, Amazon takes €8,400 of the €13,000 ad budget because its ROAS is 5.6 versus 4.1 on bol and 3.2 on MediaMarkt.
A pure ROAS allocation would keep feeding Amazon. The concentration cap says Amazon is already at 65% of spend and 71% of paid revenue. Meanwhile, MediaMarkt’s tablet sleeve campaign has lower ROAS but a higher average order value, only 4% returns and 38 days of stock. The operator moves €1,000 from Amazon generic accessories into MediaMarkt Sponsored Product Ads, with a stop rule if CPC rises above €0.82 or contribution margin drops below €3.40 per order.
That is the trade-off competitors often miss: sometimes the “worse” channel earns budget because it lowers dependency risk and opens a second profit lane.
How to build the cap inside a weekly managed service
A concentration cap should be boring enough to run every week. If it needs a workshop, it will die in the second busy month.
Use one table with these fields:
- Object: SKU, product family, marketplace, campaign role or keyword cluster.
- Current spend share: percentage of total marketplace ad spend.
- Paid contribution share: percentage of paid contribution profit after marketplace fees, fulfilment, expected returns and ad cost.
- Stock cover: days available at current paid and organic velocity.
- Break-even ACOS: loaded with fees, returns, discounts and service cost where relevant.
- Role: protect, scale, learn, fix or harvest.
- Cap status: green, review, capped or release.
- Next action: scale, hold, shift, reserve, test or pause.
FiveX product profitability dashboards help here because the operator does not have to rebuild contribution margin in a separate spreadsheet before every budget call. Advertising automation can then respect the cap: bids may move inside approved limits, but budget cannot cross the concentration threshold without a review. AI recommendations become more useful too, because the question changes from “where can performance improve?” to “where can performance improve without increasing portfolio risk too much?”
Good caps are not anti-growth
The pushback is predictable: “If a campaign works, why limit it?” Fair question. The answer is that a concentration cap is not a ceiling on ambition. It is a checkpoint before dependency becomes invisible.
If the SKU has 60 days of stock, stable margin, low returns, strong reviews, clean fulfilment and proven incrementality, let it exceed the cap. Document the reason and release budget. Great operators do not block winners for the sake of balance. They make winners prove they can carry the extra commercial weight.
The cap should also flex by season. During Black Friday, Prime Day, Sinterklaas or a MediaMarkt category promotion, concentration may be intentional. A hero SKU can deserve 50% of spend for five days. The rule is not “never concentrate”. The rule is “never concentrate accidentally”.
The final test for your advertising service
Ask your marketplace advertising partner one question this week: “Which SKU, marketplace or campaign role would hurt us most if it stopped working tomorrow?”
If the answer is a shrug, you do not have budget governance. You have optimisation.
A strong Advertentie Service should know where the account is concentrated, why that concentration exists, what would break it, and which second lane is being built before the first lane gets tired. That is how Amazon, bol and MediaMarkt advertising becomes less fragile as it scales.
At €5K monthly spend, concentration risk is already worth managing. At €15K, ignoring it gets expensive. At €50K, it becomes a board-level conversation with nicer charts and less patience.
Scale the winners. Absolutely. Just do not let one winner quietly become the whole business case.