The wrong marketplace advertising agency does not usually fail because it cannot lower bids. It fails because it optimises the number that is easiest to see while the business loses money somewhere else.
That is the uncomfortable bit. A bol Sponsored Products campaign can show a tidy ROAS while LVB fees, returns and stock cover eat the profit. An Amazon campaign can hit target ACOS while the promoted SKU loses contribution margin after referral fees, FBA, coupons and returns. MediaMarkt retail media can look exciting because the inventory is new and the audience is high-intent, but the channel still needs the same boring question: does each extra euro create profitable demand?
If you spend more than €5,000 a month across Amazon, bol or MediaMarkt, the agency selection question should not be “who has the nicest deck?” It should be: can this team prove, in the first 90 days, which SKUs deserve more visibility and which campaigns are buying revenue you cannot afford?
What competitors explain well, and where the gap sits
The agency-selection content in the market is useful, but it often stops one layer too early. Podean explains the broad Amazon agency landscape well: marketplace strategy, retail operations, content, media, analytics and global expansion. BidX is strong on automation, with claims such as campaign creation up to 12 times faster, 14 hours saved per week, 36% ROAS improvement after six weeks and 43% sales growth after four months. SellerMetrics positions Amazon PPC management around transparent retainers from $900 per month and a “human led, AI powered” process. Sequence talks about full-stack marketplace growth, DSP, AMC and multi-channel execution.
All useful. But most of this advice still helps you choose a capable advertising partner, not necessarily a profitable one.
The missing layer is the operating test between the ad console and the P&L. Agencies love campaign metrics because they are available every morning. Finance teams care about contribution margin, stock pressure, returns, marketplace fees and whether the ad strategy is building organic rank or renting it. The best agency can translate between those two worlds without needing three weeks and six exports.
That is the angle FiveX can own: marketplace advertising service should be judged by its ability to connect advertising decisions to SKU economics across channels, not by ACOS theatre.
The 90-day profit test
Before you sign a 12-month agreement, ask the agency to work through a 90-day profit test. Not a fluffy “pilot”. A real operating test with four outputs.
- SKU-level margin map. Every promoted SKU gets net revenue, marketplace fees, fulfilment costs, return allowance, cost of goods, contribution margin and break-even ACOS or ROAS.
- Campaign-to-SKU mapping. Search terms, campaigns and placements are tied back to the SKU they actually sell, not only the SKU they advertise.
- Stock and Buy Box guardrails. Budgets are reduced before the SKU runs out of stock, loses delivery promise or loses the Buy Box.
- Weekly decision log. Every budget increase, bid cut, negative keyword, SKU pause and channel shift has a commercial reason attached.
If an agency cannot produce those four outputs in the first month, it may still be good at operating campaigns. It is not yet good enough to manage marketplace advertising as a profit lever.
Named example 1: the Amsterdam cookware brand with a “good” ACOS and a bad SKU
An Amsterdam cookware brand spends €18,000 a month: €10,000 on Amazon.nl, €6,000 on bol Sponsored Products and €2,000 testing MediaMarkt. The hero SKU is a €39.95 ceramic frying pan set. Amazon shows 24% ACOS, bol shows 520% ROAS, and the campaign dashboard looks green enough to scale.
The SKU margin map tells a different story. Net selling price after VAT is €33.02. Marketplace commission and payment costs take €5.10. Fulfilment and pick-pack take €4.85. Cost of goods is €14.20. Returns and damaged packaging average €1.95. That leaves €6.92 before advertising. At a €39.95 selling price, break-even ACOS is 17.3%. A campaign running at 24% ACOS is not efficient. It is buying loss-making demand with a very polite chart.
The fix is not to slash every bid. The agency should split the work. Branded defence and exact high-conversion terms stay live because they protect ranking and convert repeat buyers. Generic “ceramic pan” terms are capped until the offer changes. The bundle version with a €54.95 price and €13.40 contribution margin gets the launch budget instead. On bol, the agency moves budget away from broad category terms into long-tail queries where the bundle converts above 8%. On Amazon, FiveX margin views make the break-even line visible next to ACOS, so the weekly discussion changes from “why did ROAS fall?” to “which SKU can absorb the next €1,000?”
Named example 2: the Antwerp accessories seller that should not scale MediaMarkt yet
A Belgian electronics accessories seller spends €7,500 a month and wants MediaMarkt retail media because the audience is perfect: shoppers already comparing laptops, phones and chargers. One USB-C hub gets a €29 average order value, €0.41 CPC, 2.8% conversion rate and 2.5 ROAS in the first three weeks. The agency deck says “promising test”.
The P&L says “not yet”. The blended gross margin is 31%, but returns sit at 11% because buyers choose the wrong port type. After marketplace fees, payment costs and return allowance, the SKU needs roughly 4.6 ROAS to break even. At 2.5 ROAS, scaling the campaign simply scales a documentation problem.
A strong agency does three things before asking for more budget. First, it fixes the product content: compatibility table, port diagrams and a clearer first image. Second, it shifts spend to a €39.95 bundle with a cable included, because that bundle has a lower return rate and higher contribution per order. Third, it uses FiveX stock and margin signals to compare MediaMarkt against Amazon and bol weekly, instead of treating every channel as a separate experiment. MediaMarkt may still be the right channel. It is just not ready for more spend until the SKU can survive the traffic.
Named example 3: the Rotterdam pet brand that ran out of ranking fuel
A Rotterdam pet supplies brand spends €12,000 a month, mostly on bol Sponsored Products and Amazon Sponsored Products. Its best-selling 12-pack dog snack has 9 days of stock cover and a 6-week replenishment lead time. The campaign has a 420% ROAS, so the previous agency keeps it at full budget to “protect momentum”.
This is a classic operator mistake. Advertising into a stockout buys a ranking spike you cannot harvest. The SKU runs out, conversion history weakens, organic position slips, and the team needs to spend again after replenishment just to recover the same ground.
The better move is boring and profitable. Cap spend on the 12-pack, move budget to the 6-pack with 38 days of stock cover, and use the final 9 days to defend only the most profitable branded and exact terms. In FiveX, the ad manager can see stock cover next to campaign performance and contribution margin, so the “best campaign” is not the one with the highest ROAS. It is the one that can actually fulfil the demand it creates.
Seven questions to ask before hiring a marketplace advertising agency
1. Which margin number will you optimise against?
If the answer is “target ACOS”, keep pushing. Target ACOS is a campaign metric, not a business model. The agency should ask for cost of goods, fulfilment costs, marketplace fees, return rates, VAT logic, coupons and channel-specific commissions. For bol, that includes LVB or own fulfilment economics. For Amazon, it includes FBA, referral fees, coupons and Subscribe & Save where relevant. For MediaMarkt, it includes the retail media cost structure and the marketplace margin reality behind the offer.
2. How do you decide when a lower ACOS is bad?
A lower ACOS is not always better. If ACOS drops because the agency cuts generic discovery terms, total sales may fall and TACoS may rise later because organic rank weakens. Reddit sellers complain about exactly this tension: some are bleeding money with high ACOS, while others accept 100% ACOS for short windows if ranking and total sales are moving. The agency needs a stance. Mature SKUs need profit protection. Launch SKUs may need controlled overspend. Clearance SKUs need cash recovery. One ACOS rule cannot serve all three.
3. How do you treat stock as an advertising input?
Stock is not an operations issue that happens after advertising. It is an advertising input. A campaign with 600% ROAS and 5 days of stock cover is not a winner. It is a risk. Ask whether the agency will reduce budgets automatically when stock cover falls below a threshold, whether it changes bids around replenishment dates, and whether it separates ranking defence from demand generation.
4. Can you compare Amazon, bol and MediaMarkt in one budget conversation?
This is where many single-channel agencies struggle. Amazon may have better attribution, bol may have stronger category intent in the Netherlands and Belgium, and MediaMarkt may give access to electronics shoppers closer to purchase. The right allocation is rarely “increase the channel with the best ROAS”. It is “increase the channel where the next euro has the best contribution margin, enough stock and a realistic ranking upside”. FiveX connects marketplace, advertising, inventory and financial data so that budget conversations are made with the same yardstick across channels.
5. What will be in the weekly decision log?
You do not need a 40-slide report. You need a decision log. What changed? Why? Which SKU, search term or campaign was affected? What result do we expect next week? What will we do if it fails? This is where the operator voice shows up. Good agencies do not hide behind dashboards. They leave a trail of commercial decisions.
6. Who owns content and offer fixes?
Advertising exposes offer problems faster than almost anything else. A low conversion rate may be a bid issue, but it may also be weak images, unclear compatibility, poor reviews, a delivery promise problem or a price gap against competitors. If the agency only changes bids, it will keep treating symptoms. Ask how content, pricing and marketplace operations are fed back into the advertising plan.
7. How will automation be supervised?
Automation is useful. Unsupervised automation is expensive theatre. Bid rules, negative keyword harvesting and budget pacing should run inside margin and stock guardrails. FiveX AdMAX is built for that kind of operating model: automation handles the repetitive bid work, while the human team sets commercial thresholds, reviews exceptions and decides where growth deserves more capital.
The fee model matters less than the incentive model
Percentage-of-spend fees are not automatically bad. Retainers are not automatically safe. Performance bonuses are not automatically aligned. The question is what behaviour the model rewards.
If an agency earns more when spend rises, you need stronger margin guardrails. If an agency is on a fixed retainer, you need clear service levels and decision cadence. If a bonus is tied to revenue, you need a profit floor. For brands spending €5,000 to €30,000 a month, the cleanest model is often a base retainer plus a quarterly bonus tied to contribution-margin-safe growth. Not revenue growth. Not ROAS improvement in isolation. Contribution-margin-safe growth.
What good looks like after 90 days
After 90 days, a good marketplace advertising agency should be able to show five things without scrambling.
- Which SKUs are allowed to scale, and their break-even ACOS or ROAS.
- Which campaigns are capped because of margin, stock, returns or weak content.
- How budget moved between Amazon, bol and MediaMarkt, and why.
- Which automations are running and which commercial guardrails control them.
- What changed in total contribution margin, not only attributed ad revenue.
If the agency can show that, you have a partner. If it can only show impressions, clicks, ACOS and attributed sales, you have campaign management. Sometimes that is enough. At €5,000+ monthly spend, it usually is not.
FiveX’s practical role in the agency relationship
FiveX does not replace good marketplace judgement. It makes that judgement easier to apply every week. The platform brings advertising performance, marketplace orders, fees, stock cover, product profitability and channel reporting into one operating view. That means your agency can optimise campaigns with margin context, your finance team can see why spend moved, and your ecommerce team can spot when a campaign is about to create a stock problem.
Three hooks matter most for managed advertising teams. First, advertising analytics show ROAS, ACOS and TACoS next to SKU performance. Second, P&L dashboards make break-even targets visible before bids are raised. Third, stock signals stop teams from advertising into products they cannot fulfil. Add AdMAX automation, and the repetitive bid work can run inside the guardrails instead of outside them.
The bottom line
Do not choose a marketplace advertising agency because it promises lower ACOS. Choose the team that can tell you when lower ACOS is the wrong move, when MediaMarkt needs content before budget, when bol spend should shift because LVB economics changed, and when Amazon growth is only profitable on the bundle SKU.
The best agency is not the one that buys the cheapest traffic. It is the one that knows which demand your business can afford to create.