Portfolio budgets look like a tidy finance feature. Group campaigns, set a cap, protect the monthly spend, and let the account breathe inside the guardrail. Lovely idea. Dangerous default.
In a managed marketplace ad account, especially once Amazon, bol and MediaMarkt together pass roughly €5K a month, a portfolio budget is not just a cap. It is a decision system. The campaigns that spend fastest get more influence. The campaigns that are paused by a shared cap lose learning time. The SKU with the cleanest margin may be blocked because another campaign in the same folder ate the morning. The operator thinks finance controlled spend. In reality, the folder chose winners.
The named mistake I see is letting portfolio caps replace portfolio judgement. A brand creates one Amazon portfolio for “Kitchen Q4”, puts brand defense, generic discovery, competitor ASINs and launch campaigns inside it, and sets a €160 daily cap. At 13:00 the portfolio is almost exhausted. Brand defense stayed efficient, generic discovery spent through broad traffic, and the launch SKU never reached its afternoon buying window. The account did not overspend. It still made a poor allocation decision.
My stance: every Advertentie Service account above €5K monthly spend needs a portfolio budget court. Not another spreadsheet of campaign budgets. A weekly and daily operating layer that decides which campaign groups deserve shared spend permission, which caps are allowed to flex, and which pools must be separated because their economics are too different to share one wallet.
This guide is written for brands advertising across Amazon, bol and MediaMarkt in the Netherlands and Belgium. Amazon portfolios are the clearest example because the platform has an explicit portfolio layer, but the problem is wider. bol campaigns also compete for budget attention. MediaMarkt retail media packages also draw money from the same commercial plan. The budget court is where those pools stop fighting silently.
What competitor advice explains well
The public advice around Amazon PPC budgets has improved. Atom11 explains that daily budgets are averaged over the month and that a strong campaign can run out early because Amazon does not pace spend evenly through the day. It also gives a useful maximum CPC formula: target ACOS × conversion rate × average order value. That is a good starting point for bid discipline.
BidX gives a broader 2026 PPC framework: setup, optimisation and scaling. Its budget section correctly warns that underfunded campaigns starve decision-making data, while overfunded campaigns can spend inefficiently before anyone notices. It also calls out pacing rules and red-flag patterns such as rising spend without rising sales.
Autron focuses on overlap. Its article on internal bidding wars makes a sharp point: portfolio structure can hide conflict when top-performing campaigns lose budget while stale campaigns keep running. That is especially relevant for bigger accounts with multiple teams, match types and campaign goals.
Podean’s marketplace content tends to zoom out to strategy: retail readiness, Q4 action, full-funnel media and marketplace expansion. That perspective is useful because budget is not only a PPC setting. It is also tied to content quality, retail operations, inventory and commercial timing.
What most of this advice still misses is the portfolio governance layer. It tells you how to set budgets, when to raise or lower them, and how to avoid overlap. It rarely defines who is allowed to let a shared budget choose between brand defense, launch learning, competitor conquest, stock clearance and channel expansion. That is the gap FiveX can own.
Why portfolio budgets become political
A portfolio budget becomes political when different campaign jobs share one cap. The platform sees spend. The business sees roles.
A brand-defense campaign has a different job from a launch campaign. A competitor-ASIN campaign has a different risk profile from a branded Sponsored Brands Video campaign. A bol Sponsored Products campaign for a high-stock SKU has a different purpose from a MediaMarkt placement supporting a short electronics promotion. If these jobs share one budget pool without role rules, the fastest spender gets accidental seniority.
That is the quiet danger. Portfolio budgets are often created for neatness: by brand, category, market, season or agency team. Those folders are easy to understand in the interface, but they are not always commercially safe. A “Coffee machines NL” portfolio can contain one SKU at 38% contribution margin, one bundle at 18%, one clearance item with only 12 days of stock, and one new product that needs learning spend. One cap cannot treat those four jobs equally and still be profit-aware.
FiveX’s Advertentie Service handles this by separating the reporting object from the decision object. A campaign can still sit in a tidy platform portfolio, but the spend decision must also carry SKU margin, stock cover, channel role, learning status and owner. That is why FiveX campaign logs, profitability dashboards and advertising reports are useful together: they show not only what spent, but why that spend had permission.
The portfolio budget court: five questions before caps move
The portfolio budget court is a short decision meeting, not a committee with biscuits and theatre. For most €5K to €50K marketplace ad accounts, it can be a 20-minute weekly review plus daily exceptions when spend drifts.
1. What job is this budget pool doing? Label every pool as Defend, Harvest, Learn, Launch, Clear, Recover or Conquer. If one portfolio contains three jobs, split the permission rules even if the platform folder stays the same.
2. Which SKU economics set the ceiling? Use contribution margin after marketplace fees, fulfilment, expected returns and discounts. A campaign with 22% ACOS can be brilliant on a 35% margin SKU and reckless on a 16% margin bundle.
3. What stock and offer signals can veto spend? Portfolio caps should not flex upward when the hero SKU has nine days of Amazon FBA stock left, when bol LVB is healthier, or when MediaMarkt has the stronger delivery promise for the same item.
4. What evidence has the pool earned? Separate proven demand from paid curiosity. Brand defense with 400 orders of history does not need the same learning budget as a new competitor campaign with 37 clicks and one assisted order.
5. When does permission expire? Every budget decision needs an expiry date. Event budgets expire after the event. Launch learning expires after the evidence window. Clearance spend expires when stock drops below the planned exit quantity.
FiveX can support this rhythm because it connects ad performance with product profitability, inventory insights and channel-level reporting. The point is not to make operators slower. The point is to stop budget changes from becoming undocumented guesses.
Example 1: NorthSea Home and the overstuffed Amazon portfolio
NorthSea Home sells small kitchen appliances on Amazon and bol. The team has a €4,800 Amazon monthly budget and €2,400 on bol, so the total managed spend is comfortably above the Advertentie Service threshold. Inside Amazon, five campaigns sit in one “Blenders NL” portfolio with a €160 daily cap.
Last week the portfolio spent €1,050 and produced €4,100 ad-attributed sales, a 25.6% ACOS. On the surface, acceptable. But the portfolio court splits the work:
- Brand defense: €210 spend, €1,820 sales, 11.5% ACOS, 39% contribution margin.
- Generic “smoothie blender”: €430 spend, €1,120 sales, 38.4% ACOS, 31% contribution margin.
- Competitor ASINs: €260 spend, €510 sales, 51.0% ACOS, 31% contribution margin.
- Launch campaign for a new glass-jar model: €150 spend, €650 sales, 23.1% ACOS, but only 14 days of Amazon stock.
The wrong move is to raise the portfolio cap because total ACOS looks close enough. The better decision is more surgical: protect brand defense at €35/day, reduce competitor ASINs from €40/day to €18/day, keep generic discovery at €55/day but add a seven-day evidence expiry, and shift €32/day of launch demand to bol where the same model has 48 days of LVB stock and a stronger local conversion rate.
Nothing dramatic happened in the ad interface. The important decision happened above it. The portfolio cap stopped being one blunt ceiling and became four budget permissions.
Example 2: BrightBaby and the launch pool that stole from winners
BrightBaby launches a new stroller accessory on Amazon while its mature baby-monitor SKU funds most of the profit. The agency creates a “Baby Gear Growth” portfolio with a €220 daily cap because finance wants one clean number. The baby monitor has 36% contribution margin and stable conversion. The stroller accessory has 24% contribution margin, no review base yet and a launch target of learning 25 search terms.
By day six, the launch campaign has spent €620, found three promising terms and one bad broad-match pocket. The baby-monitor campaign has hit budget limits on four of six days, missing an estimated €1,300 in profitable ad-attributed revenue at 14% ACOS.
The portfolio court names the issue: learning spend is stealing from proven harvest spend. The fix is not “stop launching”. It is to put the launch pool on a fixed evidence ticket: €420 for seven days, max CPC €0.62, pause any term after €28 without basket activity, and require at least two terms below 32% ACOS before the next €300 is released. The baby-monitor campaign receives its own protected cap of €150/day and cannot be throttled by the launch test.
This is where FiveX Ads AI and approval workflows fit naturally. Recommendations can suggest bid changes, but the court decides whether the portfolio has earned more budget. Automation proposes. Profit permission disposes. Very elegant, very unglamorous, very useful.
Example 3: VoltPeak and the MediaMarkt package problem
VoltPeak sells headphones across Amazon, bol and MediaMarkt. A MediaMarkt retail media package becomes available for €1,200 over two weeks. The commercial team likes it because the retailer audience is high-intent. The Amazon operator is nervous because the same hero SKU already uses €95/day on Sponsored Products and bol is preparing a weekend push.
The lazy decision is to add the MediaMarkt package on top and hope total TACoS stays acceptable. The portfolio court treats it as a channel-pool trade-off.
The SKU has a €79 selling price, €24 contribution margin before ads, 620 units available, and a planned reorder in 21 days. Amazon is allowed to spend up to €665 this week because it has proven non-brand conversion below 27% ACOS. bol receives €420 because local reviews and LVB delivery are strong. MediaMarkt gets an initial €480, not €1,200, with two unlocks: click cost must stay below €0.70 and total channel contribution after expected returns must remain above €11 per unit.
If MediaMarkt passes those rules after four days, the next €360 opens. If it fails, the unused money returns to the Amazon and bol harvest pools. That is the court doing its job: not blocking retail media, but refusing to let a package price become a budget argument by itself.
How to run this without creating bureaucracy
The court needs one page. Seriously. If the workflow needs twelve tabs, operators will dodge it during busy weeks.
Use one row per budget pool. Add columns for marketplace, campaign role, SKU group, weekly cap, daily cap, contribution-margin ceiling, stock cover, evidence status, owner, decision, expiry date and next review. The decisions should be verbs: release, cap, split, shrink, protect, shift, freeze or expire.
Daily exceptions should be even shorter. Did a portfolio hit 70% of daily budget before noon? Did a protected campaign stop serving? Did a launch pool spend 50% of its learning ticket without a qualified signal? Did stock cover fall below the veto line? Those are court triggers.
FiveX’s value is that those triggers do not have to live in someone’s head. Product profitability shows the margin ceiling. Inventory insights show the stock veto. Advertising reports show spend and ACOS. Campaign logs show what changed. For managed Advertentie Service work, that means the operator can walk into the weekly budget court with evidence instead of vibes.
The practical rule
Portfolio budgets should protect strategy from platform noise. They should not become the strategy.
If a shared cap contains campaigns with different margins, stock realities, learning maturity or channel roles, it needs a court above it. Decide the job first. Set the cap second. Let automation help third. Review expiry fourth.
That order matters. When the platform chooses winners, spend follows speed. When the portfolio budget court chooses winners, spend follows profit permission. For Amazon, bol and MediaMarkt accounts above €5K, that difference is not admin. It is the operating system that keeps ad management from quietly becoming budget roulette.