Marketplace ad management breaks down surprisingly rarely because an operator does not understand bids. It breaks down because nobody knows who is allowed to make the uncomfortable decision.
Amazon Sponsored Products are overspending on a hero SKU with 11 days of stock left. bol Sponsored Products found a cheap search term, but the product page has a weak image and a 3.8 rating. MediaMarkt retail media is producing low-cost clicks for an electronics bundle, while finance says the return reserve just moved from 6% to 11%. The agency sees optimisation work. The brand sees operational risk. Finance sees margin leakage. Sales sees lost visibility. Everyone is partly right, which is exactly why the account keeps drifting.
The named mistake is outsourcing execution without assigning decision rights. A marketplace advertising service can manage keywords, campaigns and budgets, but it should not have to guess whether a stockout risk beats a ROAS target, whether a price promotion changes break-even ACOS, or whether MediaMarkt gets more test budget while Amazon is still below contribution margin.
My stance: once monthly marketplace ad spend passes roughly €5K, you need a simple RACI before you need another clever bid rule. RACI means Responsible, Accountable, Consulted and Informed. In normal project management it can feel like corporate furniture. In Amazon, bol and MediaMarkt ads, it is much more practical: it tells the operator which decisions they can make today, which decisions need approval, and which decisions are blocked until the commercial facts are fixed.
This guide is written for NL/BE brands using an Advertentie Service model: the brand keeps commercial ownership, while a specialist team manages marketplace media across Amazon, bol and MediaMarkt. FiveX helps by connecting ad performance to SKU margin, stock cover, product profitability, marketplace analytics and automation rules, so the RACI is not managed from gut feeling or a heroic spreadsheet.
What the outsourcing advice usually gets right
The advice you find online is not useless. Most Amazon PPC outsourcing guides correctly say that agencies bring specialist experience, save time, improve campaign structure, manage search terms and create more regular reporting. Perpetua makes a useful distinction between agency support and automation software, including the reality that European agency fees often combine a monthly minimum with a percentage of ad spend. BidX is right that campaign structure matters because budget only lives at campaign level on Amazon, so exact and broad traffic should not compete for the same daily cap. Podean positions marketplace media as part of a broader operating system, not just bid pushing. Reddit-style seller discussions are also honest about the pain: rising ACOS, opaque algorithms, competitive bids and the steep learning curve.
But most of that content stops one layer too early. It answers who can optimise the account? It does not answer who is allowed to trade off profit, inventory, price, marketplace role and learning speed when those things disagree?
That missing layer is where managed marketplace advertising earns or loses trust. A good agency does not just lower ACOS. A good agency knows when a 24% ACOS is dangerous because contribution margin fell, when a 41% ACOS is acceptable because a new SKU is buying rank evidence, and when a beautiful 7x ROAS should not receive another euro because the warehouse has ten days of stock left.
The decision-rights layer most €5K accounts are missing
At €500 per month in ad spend, informal decisions can work. The founder notices the campaigns, lowers a bid, checks stock, sends a quick message and moves on. At €5K to €50K, that breaks. There are too many SKUs, too many marketplaces and too many small decisions that compound into a margin problem.
A marketplace ad RACI should answer five questions before optimisation starts:
- Who can move bids without approval? For example, the agency may change bids within a pre-approved break-even CPC band.
- Who can move budget between Amazon, bol and MediaMarkt? Cross-channel allocation changes are commercial decisions, not only media decisions.
- Who can pause spend because of stock, Buy Box, price or listing quality? The operator needs veto rights when ads are sending demand into a broken shelf.
- Who approves learning spend? Launch and test budgets should have a known loss limit, not an awkward conversation after the invoice lands.
- Who owns the fix when ads reveal a non-ad problem? A search term may be fine while the content, price, fulfilment promise or return rate is the real issue.
FiveX is useful here because the same view can show campaign spend, SKU contribution margin, stock cover, marketplace revenue, returns and product profitability. That turns the RACI from a static document into an operating board: the agency can see whether a decision is inside permission, outside permission or waiting for an owner.
The RACI table I would use for Amazon, bol and MediaMarkt
Keep it boring. Boring is good. Boring means people actually use it on Tuesday morning.
| Decision | Responsible | Accountable | Consulted | Informed |
|---|---|---|---|---|
| Bid changes inside approved margin band | Ad operator | Marketplace lead | FiveX profitability view | Finance weekly summary |
| Budget move above 15% between channels | Ad operator proposes | Commercial owner | Finance and supply chain | Agency/client team |
| Pause SKU because stock cover is below threshold | Ad operator | Supply chain owner | Marketplace lead | Sales and finance |
| Keep launch spend above break-even ACOS | Marketplace lead | Commercial owner | Ad operator | Finance |
| Change price, coupon or promo support | Commercial owner | Commercial owner | Ad operator and finance | Agency team |
| Move a winning search term into scale budget | Ad operator | Marketplace lead | Finance if margin changed | Weekly report |
| Fix listing image, title, A+ or content gap | Content owner | Marketplace lead | Ad operator | Commercial owner |
The important part is not the table itself. The important part is that every recurring ad decision has a default owner. If the ad operator has to ask for permission on every €0.07 bid move, the service becomes slow. If the operator can move €2,000 from bol to Amazon without margin and stock approval, the service becomes risky. The RACI protects both speed and profit.
Named example 1: LumaHome and the Amazon stock veto
LumaHome is selling a cordless vacuum accessory set on Amazon.nl. The SKU sells for €39.95. After referral fees, fulfilment, VAT handling, packaging, expected returns and landed cost, the product has €10.20 contribution margin before ads. The loaded break-even ACOS is about 25.5%.
The campaign is performing nicely at 21% ACOS with €1,450 spend this month. A normal agency report would say: keep scaling. The problem is stock. FiveX inventory insights show 11 days of cover, while the next replenishment is 19 days away. If the agency scales the bid from €0.74 to €0.91 because ROAS looks strong, the brand may buy a stockout, lose organic position and then spend again to restart demand two weeks later. Lovely little circle of pain.
The RACI decision is clear: the ad operator is Responsible for noticing the stock threshold and pausing scale moves. Supply chain is Accountable for the stock veto. The marketplace lead is Consulted on whether to keep a small defend budget. Finance is Informed that profit was protected by not chasing short-term ROAS. The action is not “optimise harder.” It is: hold Amazon scale, keep exact-brand defence at €12 per day, move €500 of the planned scale budget into bol where stock cover is 43 days.
Named example 2: NordTrail and the bol listing-quality block
NordTrail sells a hiking water filter on bol.com for €32.99 with €8.80 contribution margin before ads. The brand wants to grow in Belgium, and bol Sponsored Products finds a phrase-match term at €0.31 CPC. Early numbers look exciting: 270 clicks, 13 orders, €428.87 ad revenue and €83.70 spend. That is a 19.5% ACOS, below the 26.7% loaded break-even line.
Then the operator checks the shelf. The Belgian product page has only four reviews, one weak lifestyle image, and the French-language bullet points are functional but not persuasive. The search term might be a winner, but the listing is doing the heavy work with one hand tied behind its back.
Without decision rights, this becomes a familiar argument. The agency asks for more budget because ACOS is good. The brand hesitates because the page is not ready. Two weeks pass. With RACI, the ad operator can move the term into a “prove” campaign up to €25 per day, but cannot move it into scale until the content owner adds two images, rewrites the French bullets and the marketplace lead accepts the page as scale-ready. FiveX marketplace research and product performance views make the gap visible: the paid signal is promising, but listing readiness is still yellow.
Named example 3: MysaTech and the MediaMarkt return reserve
MysaTech sells a smart plug bundle on MediaMarkt and Amazon. MediaMarkt retail media clicks are cheap at €0.22 CPC, and the campaign has generated €3,120 revenue from €390 spend. On the surface, the 8.0x ROAS looks like a green light. Amazon sits at only 4.1x ROAS for the same bundle.
The finance view changes the story. MediaMarkt return reserve moved from 6% to 11% after a compatibility issue with one router brand. The bundle still has margin, but not enough to support aggressive prospecting until the content team updates compatibility copy and customer support tags the issue correctly. Amazon has a higher CPC, but lower return drag and stronger branded search capture.
The RACI says finance is Consulted on any reserve change above three percentage points. The ad operator is Responsible for shrinking MediaMarkt prospecting from €60 to €20 per day. The commercial owner is Accountable for whether the remaining €1,200 monthly budget moves to Amazon, bol or stays in a fix-first reserve. The decision is not anti-MediaMarkt. It is pro-margin: fix the reason clicks are returning before buying more of them.
Five decision thresholds to put in the service agreement
A RACI becomes useful when it has thresholds. Otherwise everyone nods politely and still argues in the next meeting.
1. Bid autonomy band
Let the operator change bids freely inside a safe range. Example: bids may move up or down 20% if the resulting estimated CPC remains below 80% of break-even CPC and stock cover is above 21 days. Anything outside that band needs approval.
2. Budget movement band
Small reallocations should be fast. Example: the agency can move up to 10% of weekly budget between campaigns in the same marketplace. Moving more than 15% between Amazon, bol and MediaMarkt requires the commercial owner because the marketplace mix changes.
3. Stock veto
Ads should not scale demand into a stockout. Example: below 14 days of stock cover, scale campaigns freeze automatically; below 7 days, only brand defence or strategic rank protection remains active. FiveX inventory insights can flag those SKUs before the ad meeting.
4. Margin reset
If landed cost, fees, return reserve, coupon depth or sales price changes, break-even ACOS changes. Example: a €4 coupon on a €39.95 product with €10.20 contribution margin can cut ad headroom by almost 40%. The RACI should force a bid review before the promo goes live, not after the campaign has spent through the new margin.
5. Learning loss limit
Not every campaign must be profitable immediately. But every learning campaign needs a receipt. Example: a launch SKU can spend €600 above break-even over 30 days if it answers one agreed question: which three search terms deserve exact-match scale, which competitor ASINs are too expensive, or whether bol Belgium converts better than Amazon Netherlands. FiveX decision logs and advertising automation help keep that learning budget from turning into permanent leakage.
What the weekly meeting should decide
A managed service meeting should not be a tour of dashboards. It should be a decision room. I would structure it around four questions:
- Which SKUs still have profit permission? Review margin, stock, price, returns and offer eligibility before media metrics.
- Which campaigns earned more autonomy? Promote campaigns that have margin, volume and clean search-term evidence.
- Which campaigns lost permission? Reduce, isolate or pause campaigns where stock, content, price or return signals changed.
- Which decisions need a non-media owner? Assign listing fixes, stock actions, price decisions and finance reviews with dates.
This is where FiveX fits naturally into an Advertentie Service workflow. The operator can bring the ad changes. The brand can bring the commercial context. FiveX provides the shared evidence layer: profitability dashboards, marketplace analytics, inventory signals, campaign performance and automation recommendations in one place. Fewer “your ROAS says this” debates. More “this SKU has permission, this one does not” decisions.
How to know your RACI is working
You will feel it in the account. Fewer emergency Slack messages. Fewer unexplained budget shifts. Fewer campaigns scaling after a margin reset. Fewer meetings where ACOS looks fine but finance is quietly unhappy.
The practical signs are measurable:
- Every budget move above the agreed band has a written reason.
- Every paused SKU has a restart condition, not just a red label.
- Every launch budget has a learning question and a loss limit.
- Every listing or stock blocker has an owner outside the ad team.
- Every weekly report ends with decisions made, not screenshots reviewed.
That is the real work of marketplace ad management. Bids matter. Keywords matter. Campaign structure matters. But once the account is large enough to hurt, the most expensive mistake is usually not a bad bid. It is an unowned decision.
If you are hiring or reviewing an Advertentie Service partner for Amazon, bol or MediaMarkt, ask them to show the RACI they would run in week one. If the answer is only “we optimise campaigns weekly,” keep pushing. Weekly optimisation is the task. Profit-safe decision rights are the operating system.