Marketplace ad management often starts with the wrong kind of energy. A brand signs the agency, grants Amazon Ads access, shares a bol login, sends a MediaMarkt retail media screenshot, and everyone wants “quick wins” by Friday. That is understandable. It is also how €5K monthly ad accounts get cleaned up cosmetically while the profit problem stays exactly where it was.
The named mistake I see is the optimisation sprint without a baseline. An operator lowers bids, pauses a few expensive keywords, launches a tidy new campaign structure and sends a dashboard with a better ACOS after two weeks. Lovely. But nobody can explain whether the account is actually healthier. Did profit improve, or did the agency simply cut launch traffic? Did Amazon gain budget because it was stronger, or because bol’s margin was invisible? Did MediaMarkt underperform because the ads were weak, or because the offer was €9 more expensive than the retail shelf next to it?
My stance: the first 30 days of marketplace ad management should not be judged by how many bid changes are made. They should be judged by whether the brand now has a profit baseline: a shared operating view of which SKUs may receive spend, which campaigns are doing which job, which marketplaces deserve budget, and which decisions require commercial permission before money moves.
This matters especially for brands in the Netherlands and Belgium spending from roughly €5K per month across Amazon, bol.com and MediaMarkt. At that level, ad management is no longer “someone improves campaigns”. It is an operating system that connects media, margin, stock, pricing, listing quality and channel strategy. Without that system, the agency can be busy, responsive and technically competent while still steering the account by incomplete signals.
What competitor advice gets right
The market has become much better at explaining why brands outsource marketplace advertising. SalesDuo frames Amazon PPC outsourcing around complexity, rising CPCs, campaign reviews, bid changes, keyword updates, negative keywords, budget checks and reporting. Its useful point is that once the account reaches a certain size, PPC needs a dedicated owner instead of being handled reactively by a founder or ecommerce manager between inventory, pricing and catalog tasks.
Podean’s Amazon agency guidance makes an important operational point: Amazon specialists often have an advantage because advertising is tied to retail operations such as forecasting, replenishment, product pages and brand stores. Its solutions page goes even further by describing retail-powered media: using non-media signals such as out-of-stock data, low stock, price changes and merchandising tactics to inform media decisions.
BidX shows the tooling side well. Its agency platform promises centralised client reporting, faster campaign creation, fine-tuning rules, stock-level scheduling, Share of Voice signals and adjustment logs. That is exactly the kind of infrastructure agencies need when one person is managing multiple accounts.
Amazon Growth Lab and several agency-comparison pieces add another useful lens: a serious Amazon PPC partner should understand TACoS, campaign architecture, listing conversion, inventory, category dynamics and reporting transparency, not just Sponsored Products bids. Reddit seller discussions are less polished but very consistent: sellers worry about losing control, paying for vague activity, and outsourcing before they know what “good” actually looks like in their own account.
What most advice still misses
Most outsourcing advice asks, “When should you hire help?” That is useful, but not enough. The more expensive question is: what should the first 30 days produce?
If the answer is “a cleaned-up account”, the bar is too low. Clean campaign names, separated match types and lower wasted spend are good hygiene. They do not tell you whether Amazon should get the next €1,000 instead of bol. They do not tell you whether MediaMarkt deserves a launch push even with lower initial ROAS. They do not tell you whether a hero SKU can afford 24% ACOS after marketplace commission, fulfilment, returns and the agency fee.
The angle FiveX can own is simple: managed marketplace advertising should start with a commercial baseline before it starts scaling tactical optimisation. In other words, the first month is not just an audit. It is the moment you decide what the agency is allowed to optimise for.
The 30-day baseline model
I like splitting the first month into four operating weeks. Not because every account behaves neatly in seven-day blocks — marketplace accounts are not toddlers with calendars — but because it gives everyone a decision rhythm.
Week 1: build the SKU permission map
The first week should answer one uncomfortable question: which products are actually allowed to receive ad spend?
For each advertised SKU, the agency needs more than sales and ACOS. It needs selling price, marketplace commission, fulfilment cost, purchase cost, expected return cost, current stock, inbound stock, Buy Box or offer status, price position, review gap and listing readiness. If those inputs are missing, the first deliverable is not “optimise campaigns”. It is “complete the commercial picture”.
In FiveX, this is where product profitability and marketplace analytics earn their keep. Instead of asking finance for a spreadsheet, the team can connect order, fee, ad and stock data and label products with practical permissions: Protect, Grow, Test, Fix first or Do not advertise.
A simple example: the NovaClean Filter Set sells for €39.95 on Amazon.nl. After commission, fulfilment, COGS and expected returns, it keeps €11.20 contribution margin before ads. Its break-even ACOS is roughly 28%. It has 44 days of stock and strong review density. That SKU can sit in Grow. The NovaClean Starter Kit sells for €49.95 but keeps only €6.10 before ads because the bundle includes a heavier accessory. Its break-even ACOS is closer to 12%. If last month’s generic campaign ran at 22% ACOS, the problem is not a lazy bid. The problem is that the SKU was never commercially allowed to play in that auction.
Week 2: separate campaign roles before changing bids
Only after SKU permissions are visible should the agency decide what each campaign is meant to do. This sounds obvious. It is not how many accounts are built.
A healthy account separates at least five roles:
- Protect: defend brand and high-intent product terms where the SKU has strong margin and stock.
- Harvest: mine search terms from controlled broad, auto or discovery campaigns.
- Scale: push proven non-branded terms that already clear contribution-margin thresholds.
- Launch: buy data for new products within a strict learning budget.
- Fix: isolate campaigns where listing, price, stock or reviews must improve before spend increases.
The first 30 days should not turn every campaign into a performance campaign. That is the quiet way to kill learning. A launch campaign with €300 spend and no sales may be doing its job if the goal was to test 40 search terms and find five candidates. A branded defence campaign with 9% ACOS may be less impressive than it looks if 95% of revenue would have happened organically.
FiveX advertising automation fits naturally here because rules should follow role and SKU permission. A Grow SKU in a Scale campaign can receive bid increases when conversion and margin agree. A Fix SKU should not receive bid increases simply because ROAS looks acceptable on a small sample. Same interface, different permission.
Week 3: build the marketplace split
By week three, the agency should stop reporting Amazon, bol and MediaMarkt as separate universes. The customer, warehouse and cashflow do not live in separate universes, so the budget should not either.
The marketplace split asks: where does the next euro create the best retained contribution margin, given stock, ranking pressure and channel role?
Named example number two: LunaBake Silicone Mat spends €5,200 per month across bol and Amazon. Amazon.de reports 3.6 ROAS and bol.com reports 4.4 ROAS, so the obvious move is to shift more money to bol. But the profit baseline tells a different story. Amazon keeps €8.40 contribution margin per order after fees and expected returns; bol NL keeps €5.70; bol BE keeps €3.90 because fulfilment and return costs are heavier. Stock cover is 31 days total. The right move is not “more bol because ROAS is higher”. The right move is: protect Amazon exact campaigns, cap bol BE discovery, use bol NL only on proven high-intent terms, and hold MediaMarkt testing until replenishment lands.
This is where FiveX multi-channel reporting prevents channel politics. The Amazon operator, bol specialist and retail media partner can all be right inside their own dashboard and still wrong for the business. A marketplace split turns the debate from “which platform performed best?” into “which platform deserves the next controlled euro?” Healthier.
Week 4: lock the weekly decision cadence
The final week should produce a cadence the team can actually keep. A 45-slide monthly report is not a cadence. It is a museum. Not where work happens.
For a €5K to €25K marketplace ad account, the weekly decision table should be brutally practical:
- Which SKUs changed permission this week?
- Which campaigns are allowed to receive more budget?
- Which bids changed and why?
- Which search terms moved to winner, watchlist or negative?
- Which products are blocked by stock, Buy Box, pricing, reviews or listing quality?
- Which marketplace gets the next test budget?
FiveX AI recommendations can support that cadence when the commercial context is present. The point is not to let AI “manage the account” in a vague way. The point is to surface candidate bid moves, wasted spend, stock risks and budget shifts so the operator reviews fewer rows and makes better decisions. Automation should reduce grunt work, not remove judgement.
The three numbers I want before the first bid move
If I had to choose only three numbers before touching bids in a new managed account, I would choose these:
- Break-even ACOS by SKU: not account average, not category guess, but SKU-level contribution margin before ads divided by selling price.
- Stock cover by advertised SKU: current and inbound stock translated into days of supply at the current paid-plus-organic run rate.
- Paid dependency by marketplace: ad spend as a share of total marketplace revenue, split by branded and non-branded demand.
These three numbers catch most expensive nonsense. If break-even ACOS is 14%, a 26% ACOS campaign is not “almost there”. If stock cover is 12 days, scaling a campaign may be self-sabotage. If paid dependency is rising while organic rank is flat, the account may be buying revenue instead of building demand.
Named example: the MediaMarkt launch that should not be compared to Amazon
The third scenario is one I see with electronics and home-appliance brands. The VoltEdge USB-C Dock sells steadily on Amazon.nl with €7,800 monthly ad spend at 21% ACOS. The team wants MediaMarkt because the product fits the retailer audience and the average order value is higher. After two weeks, MediaMarkt retail media shows only 1.8 ROAS against Amazon’s 4.1 ROAS. The nervous reaction is to cut it.
The baseline view slows everyone down. MediaMarkt orders keep €18.50 contribution margin before ads because the selling price is €89.95 and return rates are lower. Amazon keeps €12.20. The MediaMarkt campaign spent €900, generated €1,620 attributed revenue and produced 18 orders. Platform ROAS looks weak. Contribution margin before ad cost is roughly €333, so it is not profitable yet. But if retail page position improves and CPC drops from €1.10 to €0.82 after relevance builds, the campaign can become viable. The correct 30-day decision is not “scale” or “kill”. It is “continue with a €600 capped test on the two best product-category placements while listing content and price parity are fixed”.
That kind of decision rarely appears in generic PPC advice. It requires ad data, margin data, marketplace context and retail readiness in one conversation.
What should be in the first 30-day report?
A good first-month report is not a victory lap. It should make the next 60 days safer and more profitable.
Include these sections:
- Account baseline: spend, revenue, ACOS, TACoS, CPC, conversion rate and paid dependency by marketplace.
- SKU permission map: each advertised SKU labelled Protect, Grow, Test, Fix first or Do not advertise.
- Campaign role map: every campaign assigned a role, with budget rules attached.
- Waste ledger: search terms, placements or products that spent past permission.
- Opportunity ledger: terms and SKUs that deserve controlled scaling.
- Operational blockers: stock, Buy Box, price, reviews, listing content and feed issues.
- Decision requests: the few choices the client must approve before the agency can move faster.
Notice what is missing: a long list of every bid change. Activity logs matter, but they are not strategy. The client needs to understand the shape of the account, not admire the volume of clicking.
How this changes the client-agency relationship
The first 30 days set the tone. If the agency starts by proving it is busy, the client learns to value activity. If the agency starts by building a profit baseline, the client learns to value decisions.
That changes meetings. Instead of “why is ACOS up?”, the question becomes “which SKU permission changed?” Instead of “can we spend more on bol?”, the question becomes “which bol campaigns clear margin after returns and stock cover?” Instead of “why is MediaMarkt ROAS low?”, the question becomes “is this a retail-readiness issue, a placement issue or a real demand problem?”
Much better conversations. Fewer dramatic Slack messages. More decisions that survive contact with finance.
When quick wins are still allowed
To be clear, I am not anti quick win. If a campaign is burning €80 per day on irrelevant search terms, pause the nonsense. If branded defence is capped while competitor campaigns waste budget, fix it. If a product is out of stock tomorrow, stop advertising it today. Operators should operate.
The trade-off is sequencing. Quick wins should protect the baseline work, not replace it. The agency can make emergency fixes in week one, but it should not pretend those fixes are the strategy. The strategy is the permission model that tells everyone what to do when the next exception appears.
The FiveX view: managed advertising needs shared truth
Advertentie Service works best when the agency and client look at the same commercial truth. That means Amazon, bol and MediaMarkt ad decisions are connected to product profitability, inventory pressure, channel contribution and automation rules — not scattered across five exports and a very tired Google Sheet.
FiveX helps teams build that shared truth by combining marketplace analytics, profit and loss tracking, inventory insights, advertising automation and AI recommendations. The human operator still makes the call. The difference is that the call is based on margin, stock and marketplace context instead of a platform dashboard trying its best with partial information.
If you are handing over €5K+ monthly marketplace ad spend, ask for more than “we will optimise campaigns”. Ask what your first 30-day baseline will contain. Ask which SKUs are allowed to scale. Ask how bol NL and bol BE will be treated differently. Ask when MediaMarkt testing deserves patience and when it deserves a hard stop. That is where managed advertising becomes a profit function instead of a prettier campaign account.