Marketplace promotions rarely fail because one person made one reckless discount. They fail because five perfectly reasonable teams each add a small incentive to the same order.
The Amazon lead launches a 15% coupon to protect conversion. The retail media specialist increases Sponsored Products budget because the coupon improves click-through rate. Shopify keeps a welcome-code pop-up live. The bol.com account manager joins a category campaign with a temporary price drop. Finance reviews the week later and sees a familiar magic trick: revenue up, contribution margin missing.
The named mistake I see is measuring promotion performance one channel at a time while margin is being spent across channels. Amazon shows coupon sales. bol.com shows campaign lift. Shopify shows conversion improvement. The ad dashboard shows ROAS. None of those views answers the commercial question: did this promoted demand create incremental profit after discounts, retail media, fees, returns, stock pressure and orders that would have happened somewhere else?
My stance: multi-channel brands need a promotion overlap ledger. Not a prettier campaign calendar. A SKU-level analytics layer that records every discount, deal, voucher, ad push, affiliate commission and marketplace-funded incentive that touches the same demand window. If two incentives help the same order, the ledger says so. If a marketplace funds the discount but your ads fund the traffic, the ledger says so. If a TikTok code steals stock from a higher-margin Amazon.de listing, yes, the ledger should be mildly annoying about that too.
This guide is for brand owners in the Netherlands, Belgium, Germany, France, Spain and the US selling across Amazon, bol.com, Shopify, TikTok Shop, Walmart or Mirakl retailers, usually from around €1.5K monthly ad spend or 1,000 orders per month. At that size, promotions are no longer cute conversion nudges. They are margin allocation decisions.
What current promotion advice gets right
The existing advice is useful, especially if you are trying to understand one platform at a time. MerchantSpring explains that Amazon marketplace analytics should track sales, units, AOV, COGS, profit, ACOS, ROAS, conversion rate, refund rate and Buy Box percentage. Their point about promotions is sensible: discounts can make revenue totals fuzzy, so sellers need margin context before calling a campaign successful.
DataHawk frames the bigger problem well: ecommerce data is scattered across Shopify, Amazon, Walmart, ad platforms and spreadsheets, which causes teams to debate numbers instead of making decisions. Jungle Scout positions sales analytics as a financial command center where sellers can compare periods and understand how promotions, refunds and other variables affect the bottom line. sellerboard pushes the right profit-first idea for Amazon PPC: break-even ACOS and break-even bids should come from product economics, not campaign optimism.
SellerApp’s Amazon coupon guide is strong on mechanics. It explains clippable coupons, eligibility, fees, suppression risks, conversion effects and the practical warning that coupons are not set-and-forget. The Reddit threads are even more blunt. Sellers argue about whether lowering price beats coupons, whether coupon rules collide with Amazon’s reference-price logic, and whether high-volume low-margin growth is only one fee change away from a cash problem.
All of that is helpful. The gap is overlap. Most advice treats a coupon, deal or ad push as if it lives in one clean channel. Real operators know the messy version. A shopper sees a TikTok creator video, searches the brand on Amazon, clips a coupon, clicks a Sponsored Products ad, buys a bundle that was already on discount, then returns one item twelve days later. Which campaign gets the win? Which budget paid for the demand? Which channel should be allowed to repeat it next week?
The unique problem: incentives stack faster than reports reconcile
Promotions have three timelines. The first is the shopper timeline: discount visible, click, purchase, maybe return. The second is the platform timeline: ad attribution, coupon redemption, deal fee, settlement, refund processing. The third is the management timeline: Monday report, Wednesday decision, Friday budget change.
Those timelines do not line up politely. A coupon can be funded today, the ad sale can be attributed tomorrow, the marketplace fee can settle later, the return can arrive after the campaign review, and the stock consequence can hit the next replenishment cycle. If your analytics only reads the first visible signal, the promotion looks better than it is.
That is why a promotion overlap ledger should start before the campaign goes live. Its job is to answer four questions:
- Which SKUs are in the promotion window? Include parent/child variants, bundles and marketplace-specific identifiers.
- Which incentives can touch the same order? Coupon, sale price, marketplace deal, ad spend, affiliate commission, creator commission, free shipping, loyalty code and bundle discount.
- Which cost owner pays each incentive? Seller-funded, marketplace-funded, ad budget, creator budget, margin pool or inventory clearance reserve.
- Which decision will this data approve? Scale, repeat, cap, isolate, clear stock, protect ranking or stop.
FiveX helps here by connecting marketplace analytics, advertising data, product profitability, inventory insights and AI recommendations into one operating view. The product hook is not “look at another dashboard”. It is “stop approving promotions from dashboards that do not share the same definition of profit”.
Scenario 1: Amazon coupon plus ads creates a fake winner
Imagine NorthSea Kitchen, a Dutch cookware brand selling a pan set on Amazon.de for €59.95. Before the promotion, the SKU has this unit model:
- Selling price: €59.95
- Landed product cost: €21.40
- Amazon referral and fulfilment fees: €14.20
- Expected return and support cost: €2.10
- Contribution before ads and promotion: €22.25
The team launches a 15% coupon, worth €8.99. The pre-ad contribution falls to €13.26. That is still workable if ad spend stays controlled. The campaign then spends €2,400 in a week and attributes €10,800 in revenue at 22% ACOS. On the Amazon Ads dashboard, the promotion looks useful: sales velocity is up 38%, conversion rate improves from 9.5% to 12.8%, and ACOS remains below the old 25% target.
The ledger tells a less flattering story. Of the 180 promoted units sold, 72 came through branded search where the product already ranked organically in position one or two. Those orders carried €647 in coupon cost and €410 in ad cost that likely did not need to happen together. Another 24 units sold through a generic keyword at €1.35 CPC, but after coupon, fees and expected returns, the SKU could only afford about €1.05 per order-acquiring click at the observed conversion rate.
Net result: the visible campaign says “scale”. The overlap ledger says “split the decision”. Keep the coupon for non-branded conquesting with a lower CPC ceiling. Remove coupon support from branded Sponsored Products. Put a seven-day cap on generic discovery. Let FiveX monitor SKU contribution margin, keyword-level ad spend and stock cover together so the next bid change has profit permission, not just ACOS permission.
Scenario 2: bol.com campaign steals stock from a higher-margin channel
Now take Veluwe Pets, a Belgian pet-supplies brand selling a subscription-friendly dog supplement. The product sells on bol.com, Amazon.nl and Shopify. The normal channel economics look like this:
| Channel | Price | Net contribution per unit | Return rate | Available stock |
|---|---|---|---|---|
| Shopify | €34.95 | €12.80 | 3% | 420 units |
| Amazon.nl | €33.95 | €9.90 | 4% | 310 units |
| bol.com | €32.95 | €8.40 | 5% | 360 units |
bol.com invites the brand into a category promotion. The team drops price by 12% and adds Sponsored Products support. In five days, bol.com sells 260 units instead of the usual 95. Lovely, until the stock planner notices only 830 units are available before the next production batch lands in 24 days.
A channel-only report celebrates bol.com growth. The ledger asks a better question: what did those units replace? During the promotion, Shopify subscription starts fell from 58 to 41, Amazon.nl maintained demand but had only 14 days of cover left, and bol.com ads spent €620 to accelerate the lowest-contribution channel. If 80 of those extra bol.com orders would otherwise have gone to Shopify over the next two weeks, the brand did not create growth. It converted higher-margin future demand into lower-margin immediate demand.
The right action is not “never join bol.com promotions”. That would be too tidy, and ecommerce is not known for its tidy little personality. The right action is a capacity rule: bol.com promotions can scale only when projected stock cover remains above 28 days after Amazon.nl baseline demand and Shopify subscription demand are reserved. FiveX can support that by combining channel revenue, order forecasts, product profitability and inventory insights, then alerting when a promotion becomes a growth throttle rather than a growth lever.
Scenario 3: TikTok creator code looks cheap until settlement arrives
Consider SierraBottle, a US drinkware brand testing TikTok Shop while Amazon remains the main profit engine. A creator posts a video with a 20% TikTok code. TikTok Shop shows $18,600 GMV in one weekend. The creator commission is 12%, platform-funded shipping support covers part of the offer, and the team assumes the test is a clean win because ad spend was only $450.
The overlap ledger adds the missing pieces. The product normally sells on Amazon for $39.95 with $10.80 contribution before ads. TikTok’s discounted price is $31.96. After product cost, fulfilment, creator commission, seller-funded discount share, payment fees and expected refunds, contribution lands at $2.10 per unit. Worse, Amazon branded search rises 31% during the weekend, so Sponsored Brands spend increases automatically by $760. TikTok created attention; Amazon also charged the brand to catch some of it.
The test may still be worth repeating. It bought creative learning, social proof and new-customer reach. But it should not be reported as a $18,600 revenue win. It is a mixed demand event with low TikTok unit contribution, Amazon halo spend and a refund lag still open. The next test should cap creator commission at 8%, reserve Amazon branded budget separately and measure new-to-brand or first-purchase quality before increasing discount depth.
How to build the promotion overlap ledger
1. Create one promotion ID before anything goes live
Every discount needs an ID that follows it across channels: BF26_PANSET_AMZDE_15COUPON, BOL_SUPPLEMENT_WEEK37, TTSHOP_CREATOR_SIERRA_AUG. Use it in campaign names, discount exports, creative briefs and internal notes. The ID is boring. Boring is good. Boring is how finance finds things later.
2. Map SKUs across marketplaces
Do not trust names. One product can appear as an ASIN, EAN, bol.com offer ID, Shopify SKU, TikTok seller SKU and internal ERP item. The ledger needs the family relationship: parent SKU, child variant, bundle, multipack and replacement product. FiveX’s SKU mapping and product profitability views are especially useful here because a promotion problem often hides behind variant fragmentation.
3. Record every incentive as a cost line
Separate discount depth from discount funding. A 15% customer discount may be 10% seller-funded and 5% marketplace-funded. Ads are a separate cost line. Creator commissions are another. Deal fees, voucher fees, free shipping and loyalty points belong in the same ledger. If the order needed three incentives to happen, all three should stand next to the order like they were caught sneaking into the margin cupboard.
4. Add the baseline you would have sold anyway
Incrementality is not mystical. Start simple. Use the previous four comparable weeks, adjust for seasonality, stockouts and major price changes, then estimate baseline units by channel and SKU. Promotions should be judged on incremental contribution, not total promoted revenue. If baseline demand was 500 units and the promotion sold 580, the commercial question is about the extra 80 units and the margin you gave away on the first 500.
5. Reserve stock before judging channel performance
Promotions do not only spend money. They spend inventory. Before a promotion is approved, reserve stock for higher-priority demand: subscriptions, high-margin channels, replenishment commitments, marketplace account-health protection and hero SKUs needed for organic ranking. FiveX inventory insights can turn this from a spreadsheet argument into a rule: scale only when stock cover stays above the threshold after reserved demand.
6. Close the ledger after refund lag
Do not finalise a promotion the Monday after it ends if your returns arrive two weeks later. Mark the first read as provisional. Then close it after the relevant refund window: maybe 14 days for low-return consumables, 30 days for apparel and electronics, longer where marketplace policies or category behaviour require it. FiveX can keep the provisional and closed views separate, which prevents teams from spending profit before it has actually survived returns.
The decision rules that make it operational
A ledger is only useful if it changes behaviour. I like four simple labels:
| Label | What it means | Typical action |
|---|---|---|
| Scale | Incremental contribution is positive, stock is safe and refund lag is acceptable | Repeat or increase budget with the same guardrails |
| Isolate | One incentive works, but stacking damages margin | Keep coupon or ads, not both, for the same demand segment |
| Cap | Promotion is useful but limited by stock, cash or return risk | Set spend, unit or date limits |
| Stop | Total contribution is negative or demand is mostly non-incremental | End the promotion and fix price, content or channel fit first |
The trade-off is speed. A promotion overlap ledger adds a little discipline before launch and a little patience after close. That can feel slower than “the deal is available, let’s go”. But the alternative is slower in a more expensive way: teams spend three weeks debating why Amazon, bol.com, Shopify and TikTok all claim success while the bank account remains unimpressed.
Where FiveX fits
FiveX is built for exactly this kind of multi-channel decision. Marketplace analytics shows what sold and where. Product profitability shows what each unit was worth after fees, discounts, fulfilment, ads and returns. Advertising automation and recommendations help stop campaigns from scaling beyond SKU margin. Inventory insights show whether the promoted demand is safe to create. Repricing and marketplace research add context when price position or competitive pressure changes the decision.
The practical outcome is simple: your team stops asking “which promotion produced the most revenue?” and starts asking “which promotion deserves the next euro of margin, stock and ad budget?” That is the question brand owners need once marketplace growth becomes multi-channel.
Promotions should still be used. Coupons, deals, creator codes and retail media pushes can be excellent. But they need one shared ledger. Otherwise every channel tells a slightly flattering story, and margin becomes the quiet person in the meeting who was right all along.