EU · BERLIN
FOCUS AMAZON · O+O · AI SHELF
COMPOUND COMMERCE
019
Pillar Amazon as the engine
Published April 27, 2026
Read 12 min · 2,200 words

Your Amazon P&L is lying to you.

Top-line growth keeps the slide green. The fully loaded P&L tells a different story — the one your CFO will not see until the channel is already breaking.

By Víctor Lozano Compound Commerce · Issue № 019 Filed from Düsseldorf
TL;DR

Most large CPG brands report Amazon as a fast-growing channel because they look at gross invoiced revenue and gross margin. The trade compression, AVS fees, Subscribe & Save vendor funding, compliance chargebacks, retail media spend, agency fees, and internal headcount that actually run through the channel sit in six different cost centers. None of them roll up to one P&L. Once fully loaded, a 45 percent gross margin channel reveals itself as a low single-digit net contribution — and one bad AVN cycle from negative.

The QBR is half over. The Amazon slide goes up.

Eighteen percent revenue growth year over year. Five EU markets green. Retail media efficiency improving. The CMO nods. The CFO does not ask the question, because nobody in the room would have the answer.

The answer is in six different cost centers, and none of them roll up to this slide.

The number on the slide is real. It just does not mean what the room thinks it means.


The slide your team brings to the QBR is a lie of omission

The Amazon revenue line is gross invoiced revenue, less standard COGS. Gross margin lands somewhere between 35 and 45 percent. The slide stops there.

Everything that compresses Amazon's actual economics sits in other parts of the company. Returns and shortages live in supply chain reporting. Trade allowances live in the trade marketing budget, netted against TM&I funds, not against channel revenue. AVS fees live in shared services. Vendor compliance chargebacks live in operations as "leakage." Subscribe & Save vendor funding lives in the promo budget. Sponsored Ads live in the marketing P&L. Agency fees and tooling live under e-commerce overhead. Internal Amazon headcount lives in SG&A.

A clean Walmart P&L has every trade dollar mapped against the customer it funds. A clean Amazon P&L has the same data spread across six cost centers and never reconciled. The reason is not technical. The channel grew faster than the chart of accounts caught up.

Profitero published an audit anecdote in 2024 of a hero product showing 65 percent gross margin in the brand's own P&L and 4 percent net margin once Amazon costs were fully loaded. Sixty-five to four. That gap is the gap most brands cannot see in their own reporting today.


Why Amazon margin compresses against you whether you negotiate or not

Amazon's vendor relationship is not a static set of trade terms. It is a continuous compression engine. Four levers operate at once.

Annual Vendor Negotiations take the visible chunk. Stratably's 2024 benchmark across 100+ brands found average trade terms increased 69 basis points in a single cycle. MerchantSpring's read on the 2025 cycle put it at 91 basis points and called it "the year Amazon won margin."

Cost price asks are separate. Consulterce's 2025 vendor survey found Amazon requested an average 6.25 percent cost decrease. Stratably's 2026 preview found 55 percent of vendors received a cost decrease ask in the current cycle.

CRaP suppression operates underneath both. SKUs that fall below Amazon's internal net margin targets get quietly suppressed. Modern Retail reported in 2024 on brands losing distribution because their realized net PPM came in at 32 percent against Amazon's 35 percent target.

Chargeback creep is the fourth lever, and the one most brands cannot see. Carbon6 puts the typical leakage at 1 to 5 percent of revenue. Pattern, cited in Modern Retail, puts the broader deduction stack at 5 to 10 percent.

These compound. A vendor that holds the line on year-one trade terms still loses 60 to 90 basis points of margin per year through the other three. Over five years, that is 400 to 600 basis points transferred to Amazon. The difference between a healthy 1P channel and a value-destructive one.


CRaP is not a content problem. It is a P&L problem dressed up as one.

The Wall Street Journal first reported on Amazon's "Project CRaP" in December 2018. Coca-Cola Smartwater could not make a 6-pack at $6.99 work for Amazon's fulfillment math; Coke had to ship a 24-pack at $37.20 to stay profitable, and eventually shifted to direct fulfillment. Seventh Generation redesigned wipe packs and built a 6-pack dish soap variant to escape suppression.

That was 2018. The mechanic has only sharpened. Modern Retail's 2024 follow-up reported at least five named CPG vendors had products fully suppressed — not just discount-adjusted — for missing Amazon's net margin thresholds. One head of e-commerce told them, "we barely have any margins anymore."

The brand never sees the algorithm. There is no Vendor Central report titled "your CRaP risk score." Brands discover the problem through symptoms. Lost buy box. Items reclassified as add-on. POs going to zero on a previously fast-moving SKU. "Inactive" status with no clear cause. Sponsored ad campaigns disabled because the SKU is not retail-ready. By the time the brand asks the vendor manager what happened, the SKU has been suppressed for four to eight weeks. The volume is gone. The share is lost.

The fix lives in finance and supply chain. Pack format engineering. Multi-pack consolidation. Strategic price floors. Assortment trimming. Most brand teams treat CRaP as a content problem because that is what their tooling lets them act on. The tooling is wrong. CRaP is unit economics.


Your chargeback exposure is the metric no one in your building owns

Amazon's chargeback program operates as a tax most brands have decided is cheaper to pay than to fight. Carbon6 estimates 35 to 70 percent of chargebacks are disputable. Most brands recover only a fraction.

The structural problem is that no one inside the brand owns the number. Chargebacks are deducted from remittance, not invoiced separately. Most brands' SAP or Oracle systems book the net remittance, not the line items. Vendor Central's dispute UIs do not aggregate exposure in a way that maps to financial reporting. The data that explains why a chargeback hit — which PO, which carton, which shipment — sits in operations dashboards the finance team does not look at.

The compliance rules change continuously. Pack size, label placement, ASN field requirements, prep types — they update with little notice. The brand's distribution center SOPs are written quarterly. The mismatch is built in.

Amazon updated the unconfirmed PO chargeback in July 2025, reducing it from 10 percent to 5 percent of product cost. Brands celebrated. Few did the math on what other compliance fees moved in the opposite direction over the same period.

The metric drifts because no single role inside the company has "Amazon chargeback exposure" as a top three KPI. Compliance is a last-line cost owned by no one with P&L power. So the leakage compounds, and the finance team explains the missed forecast with "net realization is down."


You are paying Amazon to advertise products that are already losing you margin

This is the part most CFOs miss.

The brand pays Amazon trade margin. Then the brand pays Amazon again — to advertise the same product whose margin Amazon already compressed.

The math is direct. A SKU running at 18 percent net product margin, after trade and compliance, running 12 percent TACoS, is netting 6 percent. If TACoS rises by 200 basis points in a competitive quarter — Pacvue's Q4 2024 benchmark showed Sponsored Brands ROAS dropped 5.1 percent QoQ — the SKU is at 4 percent. Add a Subscribe & Save vendor-funded discount on top, and the SKU is functionally break-even or below.

The break-even ROAS depends on the underlying net product margin. At 25 percent net margin, you need 4x ROAS to break even on ad-driven revenue. At 15 percent, 6.7x. At 10 percent, 10x. Industry average ROAS hovers near 5x. Industry average net margin, after full loading, sits under 15 percent on most CPG categories.

The math does not work at the average. It works at the top quartile of net margin or the top quartile of ROAS. Almost nowhere in between.

The structural reason this stays hidden: the Sales team owns Amazon revenue and trade. The Marketing team owns ad spend across channels including Amazon. Supply chain owns chargebacks and freight. Finance owns the consolidated P&L. No one owns fully loaded Amazon contribution.

Sales is incentivized on revenue and trade efficiency. Marketing is incentivized on ROAS. Neither is incentivized on the joint metric. When the consolidated view gets built — and it almost never does — both teams have already locked their plans for the year. Sales conceded 90 basis points in the AVN to "protect distribution." Marketing committed 12 percent TACoS to "win share." Together those two decisions transferred 200 basis points of margin to Amazon. Apart, neither team made a wrong call against its own MBO.


What the gap actually looks like

Take a midpoint scenario. €50M gross invoiced Amazon EU vendor. Year five of relationship. Industry-average compliance. Average TACoS. Average AVS investment. Average S&S share. Numbers below are illustrative, anchored to the industry midpoints sourced above.

§ Illustrative · €50M CPG vendor · Year 5 Source: industry midpoint benchmarks
What leadership sees in the deck
Gross invoiced revenue€50.0M100.0%
Standard COGS(€27.5M)55.0%
Reported gross margin€22.5M45.0%
(Trade booked separately, off this P&L)(€7.0M)14.0%
"Channel contribution" the slide reports€15.5M31.0%
What is actually happening in the channel
Gross invoiced revenue€50.0M100.0%
Returns / shortages / PPV adjustments(€2.0M)4.0%
Net invoiced revenue€48.0M96.0%
Annual trade terms (AVN cycle)(€7.5M)15.0%
AVS fee(€1.25M)2.5%
Freight + damage allowance(€1.5M)3.0%
S&S vendor funding(€0.4M)0.8%
Co-op / promotional accruals(€0.5M)1.0%
Compliance chargebacks(€1.5M)3.0%
Net realized revenue€35.35M70.7%
Standard COGS(€27.5M)55.0%
True gross profit€7.85M15.7%
Sponsored Ads + DSP (12% TACoS)(€6.0M)12.0%
Agency + AMC(€0.4M)0.8%
Internal headcount (3 FTE allocated)(€0.45M)0.9%
Tools + data subscriptions(€0.15M)0.3%
Operational overhead(€0.3M)0.6%
Allocated SG&A(€0.25M)0.5%
True channel contribution€0.30M0.6%
The gap: 30.4 points. €15.2M of margin that does not show up on the slide leadership reads.

This is not a worst-case scenario. It is a midpoint at industry-average compliance, average TACoS, average AVS, and average S&S share. A weaker-than-average performer is already negative today and does not know it. One more AVN cycle at 90 basis points takes the channel below zero.


What financially mature operators track that you don't

The brands that have cracked Amazon profitability are not running a different ad strategy. They are running a different reporting model.

Net PPM per ASIN. Amazon's own measure of vendor profitability — net invoiced revenue minus all post-invoice costs. Industry benchmark thresholds: hardlines 40 to 45 percent, softlines 30 to 37 percent, consumables 27 to 35 percent. Below those, the ASIN is at CRaP risk. Most brands do not pull this number.

Contribution margin per ASIN, fully loaded. Net realized revenue minus COGS minus allocated media minus allocated trade minus chargeback reserve. The single most important number in Amazon reporting and the one nobody calculates correctly.

True ROAS including trade. Sponsored ad spend divided by net revenue, not gross, with trade and S&S funding netted out. Drops most brands' headline ROAS by 30 to 50 percent.

Cost of revenue from Amazon. Total dollars sent to Amazon as a percent of Amazon revenue. Trade plus AVS plus S&S plus media plus chargebacks plus agency. Tracked as a single number. Benchmarked against other channels. Most brands have never calculated it.

TACoS to Net Margin Ratio. TACoS divided by underlying net product margin. Above 0.6, the channel is funding Amazon's media business at the expense of its own profitability.

The list is not long. The reason most brands do not track these is not that the data is hard to get. It is that no one in the building has the cross-functional authority to demand the rollup.


The question your CFO should be asking on Monday

What is our fully loaded net contribution per euro of Amazon revenue, and how has it moved over the last three years?

That single question, asked seriously, exposes everything. That nobody owns the consolidated metric. That the trend line is down, not flat, not up. That the number is materially different from the gross margin slide. That media spend, trade, and chargebacks have all grown faster than revenue.

The follow-up: if we held Amazon to the same fully loaded contribution standard we hold a new product launch to, would we still be growing this channel?

For most CPG brands in 2026, the honest answer is no. The investment would shift from media spend to assortment trim, pack engineering, and trade renegotiation. Categories that look like growth contributors at gross margin would reveal themselves as net dilutive.

The channel keeps growing because the question never gets asked in those terms. Inside the building, growth on Amazon is treated as evidence of strategy. Outside the building, it is increasingly evidence of margin transfer.

Amazon understands this perfectly. Its retail media business — now 79.7 percent of US retail media spend, $60.3 billion in 2025 per eMarketer — is not a separate business line. It is the same business line as vendor margin compression. Two faucets running into the same bucket.

The brand's role in this system is to keep paying.


The slide is still up. The CMO is satisfied. The CFO moves to the next agenda item. Nobody asked the question.

That is where the work starts.

Frequently asked questions

Why does standard gross margin reporting overstate Amazon profitability for CPG brands?

Standard gross margin is calculated as net invoiced revenue minus COGS, but Amazon's true cost stack — annual trade allowances (12 to 18 percent), AVS fees (2 to 3 percent), Subscribe & Save vendor funding, compliance chargebacks (1 to 5 percent), retail media spend (10 to 15 percent of revenue), agency fees, and allocated headcount — sits in six different cost centers across sales, marketing, supply chain, and finance. None of them roll up to one P&L. A 45 percent reported gross margin commonly lands in the low single digits once fully loaded.

What is Amazon CRaP and why does it matter for vendor margin?

CRaP stands for "Can't Realize a Profit." It is Amazon's internal flag for SKUs unprofitable to ship at the consumer price. The Wall Street Journal first reported on it in 2018 with named brands like Coca-Cola Smartwater and Seventh Generation. The algorithm is not exposed to vendors. Brands discover the problem through symptoms: lost buy box, items reclassified as add-on, POs going to zero. The fix lives in pack format engineering, multi-pack consolidation, and assortment trimming — finance and supply chain decisions, not content refreshes.

How much do Amazon vendor chargebacks typically cost CPG brands?

Industry estimates converge on 1 to 5 percent of invoice value annually for compliance chargebacks (Carbon6, Reason Automation). Pattern's broader deduction stack figure cited in Modern Retail puts total Amazon deductions at 5 to 10 percent of vendor revenue. For a €100M Amazon vendor, that is €1M to €10M of leakage per year. Carbon6 estimates 35 to 70 percent of chargebacks are disputable, but most brands recover only a fraction because the dispute process is operationally painful.

What is the right way to evaluate Amazon retail media ROI for a CPG brand?

Not ROAS in isolation. The brand needs break-even ROAS calculated against fully loaded net product margin: at 25 percent net margin you need 4x ROAS to break even on ad-driven revenue, at 15 percent you need 6.7x, at 10 percent you need 10x. Industry average ROAS hovers near 5x. Industry average net margin after full loading sits under 15 percent on most CPG categories. The math does not work at the average. The right metric is TACoS divided by underlying net margin — above 0.6, the channel is funding Amazon's media business at the expense of its own profitability.

// COMPOUND COMMERCE // EVERY OTHER MONDAY // NO PAYWALL

One essay every two weeks. Built for operators.

Frameworks, decisions, what's actually working — written for the people moving real revenue.

01 Frameworks the team can use Monday morning.
02 Field notes from inside the corporate machine.
03 The data, the decision, the implication. No filler.
COMPOUND COMMERCE NO PAYWALL

Subscribe to read what your competitors won't.

EVERY OTHER MONDAY · NO FILLER · UNSUBSCRIBE ANYTIME