Amazon Merch on Demand, still often called Merch by Amazon, looks like the cleanest version of ecommerce. No inventory purchase. No warehouse. No pick-pack cost. No customer-service queue. Upload the design, choose the product type and price, and Amazon handles production, fulfilment, customer service, returns and fraud prevention. The seller receives a royalty when the product sells.
That simplicity is exactly why the advertising decision gets dangerous. When there is no stock to buy, teams start acting as if there is no unit economics problem either. They launch Sponsored Products, push designs into automatic campaigns, celebrate ad-attributed sales and forget that the real commercial ceiling is not revenue. It is royalty after Amazon’s production and fulfilment economics.
The named mistake is treating a royalty product like a normal retail SKU. A normal Amazon seller can sometimes absorb a higher ACOS because contribution margin includes product margin, repeat purchase value, bundle effects or inventory liquidation goals. A Merch on Demand seller has a narrower lane: the royalty per sale. If the ad click consumes that royalty, the “sale” is just a nicely reported loss.
My stance: Merch on Demand advertising needs a royalty floor before campaign scale. Not a generic ACOS target. Not “let’s spend €20 and see.” Not a lottery campaign with every design thrown into one bucket forever. A royalty floor says: this design only receives budget if the expected ad cost per sale stays below the royalty we are willing to risk, and if the design proves buyer intent before scale.
This guide is for marketplace teams and brand owners managing at least roughly €5K in monthly marketplace media across Amazon, bol and MediaMarkt. Even if Merch on Demand is only a side lane for apparel, fan products or limited-run creative, the same discipline applies to any royalty-like product: print-on-demand apparel, licensed merchandise, marketplace-exclusive bundles, creator collaborations and low-margin accessories.
What the existing Merch advertising advice gets right
The public advice is useful, especially for beginners. BidX explains the basic Merch model well: Amazon creates the listing, prints and ships the product, and the creator earns a royalty from the sale price after Amazon’s costs. It also points out that creators can promote products with Sponsored Products or Sponsored Brands.
AMZScout and SellerApp-style guides are good at explaining the beginner flow: request access, research a niche, upload designs, write listings, price carefully and promote winners. POD advertising guides, such as PODtomatic’s Amazon Ads guide, go deeper on campaign mechanics: start with Sponsored Products, use automatic campaigns for data collection, separate close match from loose match, add negative keywords and scale winners gradually.
Reddit sellers add something the polished guides often miss: fear of waste. One seller describes increasing a campaign from $5 daily budget only after it repeatedly runs out of budget and stays under a 20% ACOS over seven days. Another says they run “lottery campaigns” with many standard tees together, spending $3–5 per day and watching low ACOS. That is operator reality: small budgets, lots of designs, imperfect data and constant temptation to scale the one thing that worked yesterday.
What most guides still miss is the account-level profit control. They discuss clicks, impressions, ACOS and scaling. They rarely ask whether the royalty per unit can carry the ad cost, whether campaign structure hides the few profitable designs inside a bucket of passengers, whether the same creative is cannibalising higher-margin channels, or whether the account is buying false confidence from a short seasonal trend.
The royalty floor: the number that decides whether ads deserve to run
For a normal marketplace SKU, break-even advertising headroom starts with selling price and subtracts COGS, referral fees, fulfilment, storage, returns and other operational costs. For Merch on Demand, Amazon already handles many of those operational costs before the creator receives the royalty. That makes the calculation simpler, but less forgiving.
If a t-shirt sells for €19.99 and the royalty is €4.80, your advertising system should not behave as if it has €19.99 of revenue to work with. The maximum paid acquisition cost is anchored to €4.80. If you want at least €2.00 contribution after ads, the ad cost per order cannot exceed €2.80. At a 10% conversion rate, that implies a maximum CPC of €0.28. At a 6% conversion rate, the same royalty only supports €0.17 CPC. A bid of €0.45 may look harmless in Seller Central. It is not harmless if the royalty lane cannot carry it.
This is why ACOS can be oddly misleading for Merch. A 25% ACOS on a €19.99 shirt means about €5.00 in ad spend per attributed sale. For a private-label seller with €9.00 pre-ad contribution, that may be acceptable. For a Merch creator earning €4.80 royalty, it is already past the floor. The platform reports an order. The royalty ledger reports a leak.
The practical rule is simple:
- Royalty per sale: what Amazon pays you after production and fulfilment economics.
- Required leftover contribution: the profit you want to keep after ads.
- Allowed ad cost per sale: royalty minus required leftover contribution.
- Max CPC: allowed ad cost per sale multiplied by expected conversion rate.
FiveX can help here because ad performance does not sit alone. You want campaign spend, attributed sales, product profitability, return signals and cross-marketplace performance in one operating view. The royalty floor becomes a bid permission rule, not a spreadsheet someone forgets after the first good week.
Example 1: the €4.80 royalty shirt that could not afford a 25% ACOS
Take NorthSea Prints, a fictional Dutch apparel brand testing a “Weekend Captain” shirt on Amazon Merch on Demand. The shirt price is €19.99. The royalty is €4.80. The team wants to keep at least €1.80 after advertising because design work, trademark checks and creative refreshes still cost money.
That leaves €3.00 maximum ad cost per sale. In week one, the automatic Sponsored Products campaign spends €96, generates 192 clicks at €0.50 CPC and attributes 12 orders. Ad cost per sale is €8.00. ACOS is 40%. Everyone can see that is too high.
In week two, after cutting bids and adding negatives, the campaign spends €72, generates 240 clicks at €0.30 CPC and attributes 16 orders. ACOS improves to 22.5%. In many Amazon accounts, the operator might call that progress and scale slowly. But the royalty floor says no. Ad cost per sale is €4.50, while the allowed ad cost is €3.00. The design is still losing €1.50 per paid order before overhead.
The right move is not “increase budget because ACOS improved.” The right move is to isolate the two search terms that converted below €3.00 ad cost per sale, keep those in an exact-match test, and move the rest of the campaign back into learning mode with a hard daily cap.
Example 2: the fan hoodie with good sales and bad channel logic
PixelPaw Studio sells a licensed pet-gaming hoodie through Merch on Demand at €39.99 with a €7.60 royalty. The Amazon ad campaign has a 18% ACOS, which means roughly €7.20 ad spend per attributed sale. At first glance, that looks almost break-even and possibly useful for rank.
The catch is channel mix. The same design is also sold as a higher-quality embroidered hoodie on the brand’s Shopify store, where contribution after fulfilment is €18. The Amazon Merch version is supposed to be the low-friction discovery product, not the main margin engine. If ads push every “pixel paw hoodie” search to the Amazon royalty product, the brand may be paying to move demand away from a more profitable channel.
The royalty floor therefore needs a second field: channel role. For PixelPaw, Amazon Merch can receive budget on generic discovery terms like “gamer dog hoodie” or “pet gamer gift,” but it should not aggressively defend branded terms that Shopify already converts profitably. FiveX’s multi-channel analytics is useful here because the decision is not inside Amazon Ads alone. The operator needs to see whether Amazon media is creating incremental demand or rerouting demand from a better-margin channel.
Example 3: the seasonal design that deserved a stop date, not a bigger budget
Belgian Brew Club launches a “Christmas Beer Crew” sweatshirt in November. The royalty is €6.40. During the first two weeks of December, conversion jumps to 14%, CPC averages €0.32 and ad cost per sale sits around €2.29. Excellent. The campaign has clear permission to scale within a capped window.
The mistake happens on 19 December. The campaign still has budget, but delivery confidence and seasonal intent are changing. Shoppers clicking “Christmas sweatshirt” after the shipping cut-off are less likely to buy, more likely to return, or more likely to be disappointed. By 23 December, conversion drops to 5% while CPC remains €0.34. Ad cost per sale rises to €6.80, above the royalty itself.
A good Merch ad system does not only ask “was this design profitable last week?” It asks “has the reason for demand expired?” Seasonal print-on-demand ads need stop dates, restart conditions and intent labels. In FiveX, this is the kind of logic that belongs in campaign notes, ad logs and automation rules: scale until the shipping cut-off, then freeze broad discovery and keep only exact terms that still clear the royalty floor.
Campaign structure: do not let lottery campaigns hide the winners
Lottery campaigns are understandable. When an account has hundreds or thousands of designs, putting many products into broad automatic campaigns is a cheap way to find signals. The problem is that the campaign average becomes useless once a few designs carry the economics.
Imagine a campaign with 200 shirts, €300 spend and €900 attributed sales. Platform ACOS is 33%. Not great, but maybe acceptable for testing. Underneath that average, five designs produced €700 of the revenue, 20 designs received clicks but no sales, and the rest barely received impressions. If the operator optimises at campaign level, the winners pay for the passengers.
For Merch on Demand, campaign structure should move through four lanes:
- Discovery lane: low daily cap, automatic targeting, many designs, no scaling rights.
- Proof lane: designs that generated orders below the allowed ad cost per sale.
- Isolation lane: exact terms, ASIN targets or theme clusters that clear the royalty floor repeatedly.
- Scale lane: only designs with stable royalty economics, enough search volume and no channel conflict.
FiveX Ads AI and bid rules should never treat those lanes equally. A bid increase in the discovery lane is expensive curiosity. A bid increase in the isolation lane can be a controlled growth move. The label matters.
The three reports a €5K Merch ad account needs every week
Once spend reaches a serious level, the weekly review should be less about “which ads got sales?” and more about “which designs still deserve paid visibility?” I would run three reports.
1. Royalty-floor exceptions
List every campaign, target or search term where ad cost per attributed sale exceeded the allowed ad cost per sale. Include royalty, required leftover contribution, conversion rate, CPC and the decision: cut bid, isolate, pause, or keep learning with a cap.
2. Design concentration
Show how much spend is concentrated in the top 10 designs, and whether those designs also carry the account’s royalty profit. If one hoodie consumes 38% of spend but only contributes 12% of after-ad royalty, that is not a hero product. That is a budget leak wearing a cape.
3. Channel-role conflict
Flag terms where Amazon Merch ads may be stealing demand from higher-margin Amazon FBA listings, bol listings, MediaMarkt retail media products or the brand’s own shop. A sale can be real and still be the wrong sale.
These reports are exactly where FiveX fits naturally: marketplace ad management should combine advertising data with profit analytics, product groups, inventory or availability context, and cross-channel reporting. The operator should not need five exports to decide whether a €0.12 CPC increase is allowed.
When should you use an advertising service for Merch on Demand?
If you are spending €100 per month testing designs, keep it simple. Use small caps, learn the platform and do not over-engineer. But once Amazon Merch, Sponsored Products and adjacent marketplace ads are part of a €5K+ monthly media budget, you need an operating model.
A good advertising service should not promise to “scale every design.” That is the wrong promise. The better promise is: we will separate creative testing from profit scaling, protect the royalty floor, document why winners receive budget, and stop paid traffic when the economics expire.
For FiveX Advertentie Service, that means the work is not just bid management. It is budget permission across Amazon, bol and MediaMarkt. It is knowing when an Amazon royalty product should learn, when a bol Sponsored Product should defend, when a MediaMarkt retail media placement should wait, and when the next euro should stay in cash.
The practical checklist before you scale Merch ads
- Do we know the exact royalty per product and marketplace?
- Have we defined required leftover contribution after ads?
- Is every campaign labelled as discovery, proof, isolation or scale?
- Are search terms judged by ad cost per sale, not just ACOS?
- Do seasonal designs have stop dates and restart rules?
- Are branded terms checked for channel cannibalisation?
- Do bid increases require a fresh royalty-floor check?
- Can the team explain why each scaled design deserves budget this week?
Merch on Demand is a lovely model when you respect its constraint. You do not carry inventory risk, but you also do not own the full retail margin. That trade-off should shape every advertising decision. The operator’s job is not to make every design visible. It is to make sure paid visibility is only bought where the royalty still has room to breathe.