Marketplace ad budget meetings often start with a deceptively simple question: “Where should we put the next euro?” Amazon shows a 4.8 ROAS. bol.com has a lower CPC. MediaMarkt retail media is opening more Sponsored Brand and Sponsored Product options in Europe. The brand has €8,000 left for the month and three teams making very confident arguments.
The named mistake is channel-scoreboard budgeting: lining up Amazon, bol and MediaMarkt by platform ROAS or ACOS and moving money to the prettiest number. It feels analytical. It is usually too shallow. A channel can look efficient because it is harvesting demand you already created elsewhere. Another can look expensive because it is defending a high-margin SKU at the exact moment competitors are entering. A third can look unproven because the reporting window is short, while the strategic value is shelf credibility in a retailer that matters to the category.
My stance: a marketplace ad service should not ask “which channel has the best ROAS?” first. It should run a budget decision board: a weekly operating system that gives every channel a job, checks SKU profit permission, confirms stock cover, reads search-term evidence, then decides whether the next euro should scale, defend, learn or wait.
This guide is written for ecommerce brands in the Netherlands and Belgium spending from roughly €5K per month across Amazon, bol.com and MediaMarkt. At that level, the expensive problem is not launching campaigns. The expensive problem is letting three retail media consoles compete for budget without a single commercial truth.
What current retail media advice gets right
The market is not short of retail media advice. MediaMarktSaturn’s own retail media positioning is clear: brands can reach high-intent shoppers close to the point of sale, using onsite formats such as Sponsored Product Ads, Sponsored Brand Ads, display, content, reporting and in-store or partner options. Its public materials point to more than 1,000 stores, 11 European markets and 50+ million loyalty members. RetailDetail also reported that MediaMarkt launched Sponsored Brand Ads in eight countries and aims for €45 million in retail media revenue by financial year 2025/26. That tells advertisers something important: MediaMarkt retail media is no longer a side experiment.
Podean’s marketplace positioning gets another part right. Modern marketplace marketing is not just bid management; it blends retail, media, content, analytics and strategy. Their “waste of ad spend” framing is useful because it moves the conversation beyond campaign tinkering. BidX, meanwhile, reflects the software side of the market: automation, bid optimisation, Amazon DSP, Amazon Marketing Cloud and marketplace ad management workflows. Flywheel’s agency directory shows the other end of the spectrum: broad service providers promising to simplify marketing and help clients choose where to invest advertising dollars.
Seller conversations add the uncomfortable bit. In Amazon seller communities, agency discussions often return to control, margins and accountability. Sellers worry that nobody will care about their SKU economics as much as they do. That criticism is not always fair to good operators, but the fear is rational. If the agency is paid to make the ad dashboard nicer, it may optimise the dashboard. If the agency is paid to protect profit, it needs different inputs.
What most advice misses: channels have different jobs
The missing angle is channel role. Amazon, bol and MediaMarkt should not always fight for the same euro. They often sit at different points in the buyer’s decision and in the brand’s operating model.
- Amazon often acts as demand capture and competitor defence. It has strong search intent, deep advertising tooling and a brutal habit of rewarding operational consistency. If a SKU wins there, the brand usually needs tight Buy Box, stock, review and margin controls.
- bol.com often acts as local volume and category expansion for NL/BE. It can be the place where Dutch and Belgian shopper language, local assortment and marketplace fit are tested more precisely than on a global Amazon playbook.
- MediaMarkt often acts as category credibility and high-intent electronics retail media. For electronics, accessories, appliances and connected lifestyle products, the shopper context matters. A MediaMarkt impression is not just an impression; it can validate the product inside a retailer customers already associate with technical choice.
The budget decision board starts by writing that role down. A channel without a role will always ask for more budget when its dashboard looks good. A channel with a role has to prove it is doing the job assigned to it.
The five gates before the next euro moves
Here is the practical sequence we use when reviewing marketplace ad accounts. It is intentionally a little strict. Good budget governance should feel slightly annoying before it saves you from an expensive mistake.
Gate 1: SKU contribution margin
Before channel performance, calculate the SKU’s contribution margin after marketplace fees, fulfilment, expected returns, discounts and variable handling costs. Not gross margin. Not “margin before ads because finance has not closed the month yet”. The decision needs the money that can actually fund advertising.
If a €39.95 product has €11.40 contribution before ads, its break-even ACOS is about 28.5%. A platform-reported 24% ACOS is not automatically safe if returns are rising or if the promotion coupon was not included. FiveX helps here by putting marketplace revenue, costs, fees and ad spend into one profitability view, so the operator is not copying numbers between dashboards like it is 2014.
Gate 2: stock cover and replenishment timing
Scaling ads into a stockout is one of the most polite ways to hand future demand to competitors. The board should show units on hand, average daily sales, inbound date and campaign role. A launch campaign may accept thinner cover if the purpose is learning. A defensive hero-SKU campaign should not.
In FiveX, inventory insights make this less emotional. If the SKU has 19 days of cover and the next inbound shipment lands in 28 days, the decision is not “increase bids because ROAS is good”. The decision is “protect rank, cap budget and stop prospecting until replenishment is real”. Tiny distinction, expensive difference.
Gate 3: channel role
Every budget move gets one of four labels: scale, defend, learn or wait. Scale means the SKU has profit and operational permission. Defend means the channel protects an important query, category or competitor position. Learn means the budget is buying evidence, not immediate profit. Wait means the idea may be good, but one of the gates is missing.
This is where the operator earns the retainer. A €700 MediaMarkt Sponsored Brand test can be sensible even with weaker direct ROAS if the role is category credibility for a new docking station range. The same €700 is wasteful if Amazon is out of stock in nine days and bol.com already proves the keyword converts locally. The board makes that trade-off visible.
Gate 4: search-term evidence
Budget should follow search behaviour, not campaign naming. Pull search terms, converting queries, non-converting spend, product targets and category placements into one view. The question is not only “what converted?” It is “what did each channel teach us that another channel can use?”
A bol.com query such as “usb c hub macbook hdmi” may reveal local wording that Amazon campaigns should test. Amazon product-targeting data may show which competitor ASINs deserve defensive bidding. MediaMarkt Sponsored Product performance may show whether shoppers search by brand, feature or compatibility. FiveX AI recommendations can turn those patterns into bid, keyword and negative-keyword suggestions, but the board still decides whether the action has profit permission.
Gate 5: measurement confidence
Not every channel gives the same reporting depth, attribution window or conversion path visibility. That does not make weaker measurement useless. It means the budget type should change. Low-confidence spend should be capped, time-boxed and tied to a learning question. High-confidence spend can be scaled faster.
The board should therefore include a confidence label: high, medium or low. A mature Amazon Sponsored Products exact campaign with 90 days of sales, stable stock and clean search-term history is high confidence. A first MediaMarkt Sponsored Brand placement around a seasonal TV accessory is medium or low confidence until the evidence catches up. Different labels, different budget rules.
Named example 1: NovaCharge should not move all budget to Amazon
NovaCharge sells a 65W GaN charger for €34.95. Contribution before ads is €9.80 after marketplace fees, fulfilment and a 6% return reserve. Amazon.nl reports 5.1 ROAS on €2,400 monthly spend, so the team wants to add another €1,500. Nice number. The board says: wait.
Why? Stock cover is only 16 days and the next inbound shipment is 24 days away. bol.com has lower ROAS at 3.7 on €1,100 spend, but it is converting the Dutch query “snellader usb c 65w” at a 19% ACOS and has a stronger local ranking opportunity. MediaMarkt is not yet a scale channel for this SKU; it gets a capped €350 learning budget for Sponsored Products on compatibility-led searches.
The final decision: Amazon receives only €400 extra for exact defensive terms, bol receives €700 for local query expansion, MediaMarkt receives €350 for learning, and €50 stays unallocated until the inbound shipment is confirmed. Platform-scoreboard budgeting would have moved the full €1,500 to Amazon. Profit-led budgeting keeps the account from accelerating into a stockout.
Named example 2: HomeBrew Pro needs MediaMarkt credibility, not cheap clicks
HomeBrew Pro sells a €189 espresso grinder. Amazon has a 31% ACOS against a break-even ACOS of 33%, which looks barely acceptable. bol.com has a cheaper CPC, but conversion is weak because the product has only 11 reviews and competes with discounted accessories. MediaMarkt has a higher expected CPC, but the shopper context is stronger: visitors are already comparing appliances and electronics-adjacent products.
The board gives MediaMarkt a learn role with a €900 cap for four weeks: one Sponsored Brand test around “espresso grinder” positioning and one Sponsored Product test on category searches. Amazon gets €600 defensive spend on brand and competitor terms. bol.com gets no scale budget until review count passes 25 or the product page improves conversion.
This decision will annoy a pure ROAS operator because the cheapest clicks are on bol.com. But cheap traffic to an under-proven listing is not a bargain; it is just discounted disappointment. The board protects the brand from confusing CPC efficiency with commercial readiness.
Named example 3: SkinLogic should defend bol, not chase a shiny new format
SkinLogic sells a facial cleansing device for €59.95. bol.com generates €18,000 monthly revenue at 22% ACOS and 14% TACoS. Amazon is smaller but stable at €6,500 monthly revenue. MediaMarkt just opened an attractive Sponsored Brand opportunity. The brand wants to test it because retail media is moving quickly and nobody wants to look late to the party.
The board checks the SKU economics: contribution before ads is €17.20, but return rate jumped from 7% to 12% after a recent bundle promotion. Stock cover is 42 days, so inventory is fine. Search-term evidence shows bol.com ranking slipped from position 3 to 7 on “gezichtsreiniger elektrisch”, while a competitor increased promotions. The decision: defend bol first.
MediaMarkt gets a €250 research hold, not a live push. bol.com receives €1,200 for exact and category defence with a rule that spend pauses if ACOS crosses 27% for seven days. Amazon remains at maintenance budget. FiveX advertising automation is useful here because the pause rule does not depend on someone remembering to check the account after the weekly meeting. The board makes the decision; automation keeps the promise.
The weekly board template
A good board is boring enough to survive Monday morning. Use one row per SKU-channel combination and keep the columns brutally practical:
- SKU, marketplace and campaign owner
- Channel role: scale, defend, learn or wait
- Contribution margin before ads and break-even ACOS
- Current spend, revenue, ACOS, ROAS and TACoS
- Stock cover, inbound date and stock risk label
- Top three search terms or targets driving the decision
- Measurement confidence: high, medium or low
- Decision: add, cap, move, pause or investigate
- Rule or owner responsible for follow-up
The point is not to create another report. The point is to reduce arguments. When Amazon asks for budget, the board asks for margin, stock and search evidence. When bol.com wants local expansion, the board asks whether the SKU can afford the clicks. When MediaMarkt looks strategically attractive, the board asks what the test must prove before the cap increases.
How FiveX fits into the operating model
FiveX is useful because the budget decision board needs connected facts. Marketplace ad consoles know campaign performance, but they do not naturally know landed cost, return reserve, stock cover, channel cannibalisation, repricing pressure or the reason a previous operator paused a keyword. Those details decide whether a bid change is smart.
In practice, FiveX supports the board in three places. First, profitability dashboards turn SKU-level contribution margin into the starting point for ad decisions. Second, marketplace analytics connect Amazon, bol.com, MediaMarkt and other channels so operators can compare roles instead of screenshots. Third, advertising automation and AI recommendations help translate board decisions into controlled actions: bid changes, budget caps, pause rules, keyword expansion and exception alerts.
That is also why our Advertentie Service starts from around €5K spend. Below that level, the founder or ecommerce manager can often still make judgment calls manually. Above it, small mistakes become recurring leaks. A weekly board gives the specialist operator enough structure to move fast without pretending every good-looking ROAS deserves more money.
The practical rule
If you remember one thing, make it this: budget follows permission, not performance. Performance tells you what happened inside a channel. Permission tells you whether the business can afford to do more of it.
The next euro should move only when four things are true: the SKU can fund the click, stock can support the demand, the channel role is clear, and the evidence is strong enough for the size of the bet. If one of those is missing, the decision is not necessarily no. It might be learn. It might be defend. It might be wait. But it should not be blind scale.
That is the difference between managing marketplace ads and governing marketplace growth. One changes bids. The other protects profit while the bids change.