Cross-marketplace ad budget allocation sounds like a finance exercise. Take the monthly budget, split it across Amazon, bol and MediaMarkt, then ask each channel owner to hit their ACOS target. Very neat. Also very dangerous.
The named mistake I see most often is the spreadsheet split: 60% of spend goes to Amazon because Amazon is biggest, 30% goes to bol because bol is strategically important in NL/BE, and 10% goes to MediaMarkt because someone wants to “test retail media”. Nobody is being lazy. The model simply starts from the wrong question.
The right question is not: “How much should each marketplace get?”
The right question is: which marketplace has profitable capacity this week, for which SKUs, at which point in the funnel?
That distinction matters for brands spending €5K+ per month on marketplace ads. At that level, budget allocation is no longer about keeping campaigns alive. It is an operating system for moving money toward SKU-level profit, while protecting stock, price position and organic ranking. If Amazon has traffic but thin margin, bol has stronger contribution per order, and MediaMarkt has high-intent electronics shoppers but limited stock depth, a fixed percentage split will quietly move budget in the wrong direction.
Here is the FiveX stance: allocate marketplace ad budget by permission, not platform preference. A channel earns budget when the SKU economics, inventory, offer quality and incrementality case say yes. If one of those says no, the budget either waits, moves to a safer SKU, or moves to another marketplace. A little less glamorous than “full-funnel activation”. Much better for the P&L.
What the research shows competitors cover well
Before writing this, I looked at how the market talks about budget allocation. There is useful advice out there, but it usually stops one level too early for operators.
Amazon PPC guides from BidX, Eva, AdLabs, SalesDuo and sellerboard are strong on campaign structure, Sponsored Products versus Sponsored Brands, TACoS, break-even ACOS, keyword harvesting and not blindly following platform budget recommendations. Seller discussions on Reddit echo the same pain: sellers move from campaign-level ACOS to ASIN-level TACoS because the blended dashboard hides where spend is actually hurting.
Podean makes an important point on Amazon DSP: media should be connected to retail signals such as low stock, price changes, merchandising and promotional cadence. That is the right direction. DataGlass goes more mathematical, arguing that reported ROAS can misallocate budget when margin, returns and platform fees are ignored, and that portfolio allocation should move toward true contribution profit rather than rolling averages.
What most of these pieces miss is the cross-marketplace transfer decision. They tell you how to manage budget inside Amazon, or inside retail media, or inside a single ad format. But a Dutch or Belgian ecommerce operator does not live inside one platform. On Monday, the question may be whether to move €1,500 from Amazon Sponsored Brands into bol Sponsored Products. On Wednesday, it may be whether MediaMarkt deserves launch support while Amazon is stock-constrained. On Friday, it may be whether a “good” Amazon ACOS is actually worse than a boring bol campaign with stronger net margin.
That is the gap this playbook fills.
The three-budget model: protect, grow, test
A useful marketplace advertising budget is not one big pot. It needs three roles.
- Protect budget: spend that defends profitable existing demand, such as brand terms, hero SKUs, Buy Box-sensitive terms, or bol queries where you already rank and convert well.
- Grow budget: spend that expands demand for SKUs with enough margin, stock and conversion strength to absorb higher CPCs or lower short-term ACOS.
- Test budget: controlled spend for new marketplaces, new ad formats, new keywords, launch products or incrementality experiments.
The mistake is letting all three roles fight inside the same monthly number. Protect campaigns then get starved because a launch needs visibility. Launch campaigns get judged too harshly because they do not hit mature ACOS in week one. Tests keep running because “we need more data”, which is often agency-speak for “we have not defined the stop rule”.
For most €5K–€50K monthly marketplace ad accounts, I like a starting point of 55% protect, 35% grow and 10% test. Not as a rule. As a sanity check. If 80% of spend is in test mode, you are probably funding curiosity. If 90% is protect, you may be harvesting today while starving tomorrow. FiveX managed advertising uses this role-based view to stop channel debates from becoming opinion tennis. Delightful sport. Terrible budgeting method.
The permission score: four checks before a marketplace gets more spend
Before you move money from one marketplace to another, give the SKU and channel combination a simple permission score. In FiveX, this is exactly where marketplace analytics, profitability dashboards, advertising automation and inventory insights belong together. The ad decision row should show margin, stock, price, performance and recommended action in one place.
1. Contribution margin after marketplace costs
Start with contribution margin after product cost, marketplace commission, fulfilment, payment costs, expected returns, service allowance and ad spend. This is where cross-marketplace allocation gets interesting. The same SKU can have different economics on Amazon, bol and MediaMarkt because fees, delivery promises, return behaviour and pricing pressure differ.
If a SKU has €14 contribution margin before ads on Amazon and €22 on bol, an identical 20% ACOS target is not identical at all. It gives each channel the same-looking efficiency target while giving one channel far less room to make money.
2. Stock cover and replenishment risk
Paid demand without stock is a polite way to buy future disappointment. A marketplace with 9 days of stock should not automatically receive more budget because yesterday’s ACOS looked good. It may need the opposite: tighter caps, lower generic bids, or a shift toward substitute SKUs with healthier coverage.
FiveX inventory insights are useful here because the ad manager should not discover stock risk after the campaign scales. If a hero product drops below 14 days of cover, that should change budget allocation before the platform optimizes itself into a stockout.
3. Offer quality: price, delivery, Buy Box and content
Marketplace ads amplify the offer you already have. They do not repair it. If the Amazon offer is €4 more expensive than bol, delivery is slower, or the Buy Box is unstable, the budget problem is not “Amazon needs better bidding”. It may be that Amazon does not deserve incremental spend this week.
This is especially true on MediaMarkt, where electronics shoppers compare specifications, warranty signals and price quickly. A Sponsored Product Ad can create attention; it cannot make a weak product page feel trustworthy.
4. Incrementality and paid dependency
Protect spend can have low incrementality and still be useful when it defends profitable demand. Grow and test spend need a tougher standard. If Amazon Sponsored Brands are mostly capturing customers who would have bought through organic ranking anyway, while bol Sponsored Products are opening new generic queries, the budget should move even if Amazon ROAS looks prettier.
This is why TACoS, organic rank movement and new-to-brand or new-to-query indicators matter. ROAS tells you whether the campaign sold. It does not tell you whether the marketplace needed that spend to sell.
Example 1: DynamoHome air fryer — when Amazon looks bigger but bol deserves the next euro
DynamoHome sells a 6-litre air fryer across Amazon.nl and bol. The monthly ad budget is €12,000. Last month Amazon received €7,200 and bol received €4,800. On the surface, Amazon looks stronger: €36,000 attributed revenue at 20% ACOS. bol generated €21,600 attributed revenue at 22.2% ACOS.
A spreadsheet split would keep Amazon in the lead. Bigger revenue, lower ACOS, more scale. Nice and tidy.
Then the contribution margin view changes the decision. On Amazon, the selling price is €89, product cost is €42, marketplace and fulfilment costs are €21, expected returns and support are €5. That leaves €21 contribution margin before ads. At 20% ACOS, ads cost €17.80 per attributed order, leaving only €3.20 before overhead.
On bol, the selling price is €91, cost stack is slightly better because fulfilment economics and returns are cleaner, and contribution margin before ads is €27. At 22.2% ACOS, ads cost €20.20 per order, leaving €6.80 before overhead. bol has worse ACOS and better profit.
FiveX would not automatically cut Amazon. Protect spend still matters because Amazon search volume supports brand visibility. But the next €1,500 should not go to Amazon generic terms. The better move is: hold Amazon protect campaigns at €180 per day, reduce bids on broad non-brand terms by 12%, move €1,000 into bol exact and category terms, and keep €500 as a reserve for whichever marketplace keeps contribution margin above €6 per ad-attributed order after three days.
That is the operator move: not “Amazon versus bol”, but “which next euro still has margin?”
Example 2: VoltEdge USB-C hub — when MediaMarkt gets test budget, not scale budget
VoltEdge launches a USB-C hub on MediaMarkt.nl at €49.95 while also selling on Amazon and bol. The team wants to put €3,000 into MediaMarkt retail media because the audience is attractive: laptop and accessory shoppers with clear purchase intent.
The SKU economics are okay but not heroic. Product cost is €17.20, marketplace and handling costs are €8.40, expected returns and support are €2.10, leaving €22.25 contribution margin before ads. With a target of at least €8 post-ad contribution, the working ad allowance is €14.25 per order. At a €49.95 price, that means a working ACOS ceiling of 28.5%.
The first week spends €650 on MediaMarkt Sponsored Product Ads, generates €2,100 attributed revenue and lands at 31% ACOS. A platform-only read says “close enough, keep learning”. The FiveX read is more cautious. Stock cover is only 18 days, the product page has 7 reviews while Amazon has 68, and the best converting query is brand-specific. MediaMarkt has potential, but not permission to scale yet.
The budget decision: keep MediaMarkt in test mode at €75 per day, shift creative effort into content and review velocity, cap generic accessory queries, and reserve scale budget until either conversion rate improves by 20% or post-ad contribution reaches €8 for three consecutive reporting windows. Meanwhile, €1,200 moves back to bol Sponsored Products where the same hub has 41 days of stock and €10.40 post-ad contribution.
MediaMarkt is not rejected. It is sequenced. That is a big difference.
Example 3: NovaClean robot vacuum — when the best campaign loses budget because stock says no
NovaClean sells a robot vacuum on Amazon, bol and MediaMarkt. Total monthly ad budget is €18,000. The strongest campaign is Amazon Sponsored Products: €4,500 spend, €27,900 attributed revenue, 16.1% ACOS and rising organic rank on “robot vacuum with mop”. Everyone wants to scale it.
But FiveX inventory data shows 10 days of Amazon FBA stock, 34 days available for bol LVB, and 29 days for MediaMarkt. Replenishment to Amazon is delayed by 12 days. If the team scales Amazon, it may win a beautiful ranking improvement and then go out of stock. Very cinematic. Very expensive.
The cross-marketplace budget decision is to protect Amazon ranking without feeding the stockout. Daily Amazon budget drops from €300 to €170. Exact high-converting terms stay live. Broad discovery pauses. €2,000 of the monthly budget moves to bol for the same product family, where stock can support demand. €800 moves to MediaMarkt for competitor-adjacent placements because the MediaMarkt price is competitive and the product has strong specs versus category alternatives.
The lesson: budget allocation is not a reward for last week’s best ACOS. It is a forecast of next week’s profitable capacity.
A weekly budget allocation cadence for Amazon, bol and MediaMarkt
Here is the practical cadence we use with marketplace advertising service clients.
Monday: classify every campaign by role
Label campaigns as protect, grow or test. If a campaign has no role, pause the budget discussion until it gets one. “It has been running for ages” is not a role. It is a historical accident with invoices.
Tuesday: score SKUs by permission
For every meaningful SKU-marketplace combination, review contribution margin, stock cover, offer quality and incrementality signal. Give each a green, amber or red status. Green can receive incremental budget. Amber needs a constraint. Red does not scale until the blocker is fixed.
Wednesday: move budget by marginal profit
Rank the next €500, €1,000 or €2,000 by expected contribution profit, not by channel politics. This is where FiveX profitability dashboards are powerful: the team can see which SKU and marketplace still has room before margin, stock or TACoS breaks.
Thursday: check pacing and cannibalisation
Look for campaigns that spend early but do not create incremental total sales, or campaigns that shift sales from organic to paid without growing the marketplace total. If TACoS rises while total revenue is flat, the budget did not grow the business. It just made the same business more expensive.
Friday: write the decision log
Record what moved and why. “Moved €1,000 from Amazon broad to bol exact because Amazon post-ad contribution fell below €4 and bol had 42 days stock cover” is a useful decision. “Optimized campaigns” is a diary entry wearing a business suit.
Simple allocation rules you can use tomorrow
- Never scale a marketplace with red stock status unless the explicit goal is clearance and the budget is capped.
- Do not compare ACOS across marketplaces without contribution margin. A higher ACOS can be more profitable when fees, returns and fulfilment economics are better.
- Keep at least 10% of budget unallocated at the start of the month. Use it as a winner reserve, not as a leftovers bucket.
- Separate launch learning from mature efficiency. New MediaMarkt tests should not be judged by the same ACOS window as mature Amazon brand defense.
- Move budget away from product problems. If content, price, reviews or Buy Box are weak, fix the offer before funding more traffic.
How FiveX helps
Cross-marketplace budget allocation is hard because the data lives in different places. Amazon Ads shows campaign performance. bol shows sponsored placements and retail outcomes. MediaMarkT retail media adds another layer. Finance has COGS and overhead. Operations has stock and replenishment. By the time the team has stitched the spreadsheet together, the best budget decision was two days ago.
FiveX connects marketplace, advertising, inventory and financial data into one operating view. For Advertentie Service clients, that means our team can manage Amazon, bol and MediaMarkt campaigns with SKU-level guardrails: break-even ACOS, contribution margin after ads, stock cover, price position, campaign role and weekly action recommendations.
The goal is not to make every marketplace spend the same. The goal is to make every euro compete for the right to be spent.
That is the grown-up version of marketplace advertising. Slightly less dashboard theatre. Much more profit discipline.