Amazon Profitability Analysis

Amazon profitability analysis exists because your payout report is not a P&L. Amazon deposits a lump sum every two weeks, and inside that number, profitable SKUs quietly subsidize losers — sometimes for years. We’ve opened accounts doing $4M a year where a third of the catalog lost money on every unit, and nobody knew, because revenue kept growing and the blended margin looked survivable. The fix isn’t a better dashboard aesthetic. It’s building true unit economics per SKU, allocating every cost to the product that caused it, and then making the kill/fix/scale decision the numbers demand. This guide walks through the full cost stack, the contribution margin framework, the tools that automate it, and the cash trap that catches growing sellers most often.

The True Cost Stack: What a Unit Actually Costs You

Most sellers can quote their COGS and their referral fee. The losses hide in the other five lines. Here’s the complete stack, in the order it hits your margin.

Landed COGS. Factory cost plus freight, duties, tariffs, and inbound placement or prep fees — per unit, at current rates, not the rates from your first PO. Sellers who quote 2023 freight numbers in 2026 are starting the analysis wrong on line one.

Referral fee. 8 to 15% of the sale price depending on category — 15% for most, 8% for some electronics tiers, with category quirks throughout. Simple, visible, and the only fee most sellers get right.

FBA fulfillment fee. Charged per unit by size tier and weight band. This is where fee engineering lives: a product that measures 0.1 inch into the next size tier pays the higher rate on every unit forever. Verify Amazon’s recorded dimensions against reality — mismeasured products are common, and remeasurement requests recover real money. Our FBA fee calculator guide covers the tier breakpoints and the packaging changes that move products down a tier, and the current rates live in our FBA fee schedule.

Storage — monthly and aged. Monthly storage runs per cubic foot and roughly triples in Q4. The killer is the aged inventory surcharge: stock sitting 181-plus days accrues escalating fees that can exceed the product’s value by day 365. Allocate storage per unit honestly — slow movers consume dramatically more storage per unit sold than fast movers, which is precisely why a slow mover’s true margin is worse than its spreadsheet margin.

Returns and refunds. Take your actual refund rate from the last 90 days, per SKU, and cost it fully: refunded revenue, the retained referral-fee portion Amazon keeps aside, return processing fees in apparel and shoes, and the fraction of returns that come back unsellable (20 to 50% is typical depending on category). A 6% return rate on a marginal SKU is often the entire difference between profit and loss, and it never appears on a standard sales report.

PPC allocation. The line sellers fudge most. Allocate ad spend to the SKUs that consumed it using the advertised-product report — not smeared evenly across the catalog. A SKU doing $20K a month with $6K of dedicated ad spend and a SKU doing $20K purely organic are utterly different businesses, and blended TACoS hides the difference. If a SKU’s allocated ad spend per unit exceeds its pre-ad contribution, you’re paying customers to take it away.

Overhead allocation. Software subscriptions, prep center, staff or agency fees, storage insurance, product liability insurance. Divide by units at whatever grain is defensible — even a crude per-unit overhead number beats pretending overhead doesn’t exist. A $2,000/month tool stack across 4,000 monthly units is $0.50 a unit, which matters when contribution is $2.

Stack all seven lines and you have contribution margin per unit: the number that decides everything else.

Contribution Margin by SKU: The Kill/Fix/Scale Decision

Run the stack for every SKU, sort descending, and the catalog resolves into three buckets. This quarterly pass is the highest-leverage financial exercise in the business.

Scale. Positive contribution margin, ideally 15%+ of price, with stable or improving trend. These SKUs have earned inventory depth, ad budget, and variation expansion. The most common portfolio error isn’t keeping losers — it’s underfeeding winners because the losers are consuming the cash and attention.

Fix. Margin between roughly break-even and thin, with an identifiable cause. The fix menu, in order of typical impact: reprice (test 5% steps upward — many SKUs are underpriced out of habit, not market pressure); re-engineer fees (size tier, packaging, case pack, shipping template); cut allocated ad waste (usually the fastest win — negate the junk terms this SKU’s campaigns have accumulated); renegotiate or re-source COGS; attack the return rate (better sizing info, expectation-setting images, packaging that survives the round trip). Give a fix SKU one quarter and one named intervention. If the number doesn’t move, it migrates to the kill bucket — “fix” cannot become the bucket where losers live indefinitely.

Kill. Negative contribution with no credible fix. Stop reordering, run down stock at whatever price clears it, and if aged fees are accruing, liquidate or remove — a removal order costs less than twelve more months of storage on a product that loses money when it does sell. The hard part is psychological: sellers protect their first product, their biggest revenue line, or the SKU that “just needs better ads” long past the point the math has spoken. Revenue is vanity here. A $30K/month SKU with negative contribution is a machine that converts your inventory capital into Amazon’s fees.

A worked example makes the stakes concrete. A $24.99 SKU with $6.20 landed COGS looks like a 75% gross margin winner. Subtract the $3.75 referral fee, a $5.47 fulfillment fee, $0.40 of allocated storage, a 7% return rate costing $1.10 per net unit, $4.80 of allocated PPC, and $0.55 of overhead, and contribution is $2.72 — about 11% of price. Fine, until Q4 storage and a $0.30 CPC increase arrive, at which point it’s a break-even product wearing a winner’s gross margin.

Two rules make the framework honest. First, decide on trend, not snapshot — a winner with three quarters of eroding margin is a fix candidate before it becomes a loser. Second, track the outputs on a standing scorecard; our seller KPI dashboard template includes contribution margin per SKU alongside the operational metrics that explain its movement.

Tools: Sellerboard, SKU Economics, and a Spreadsheet

You need two layers: an automated tool for the always-on view, and a manual model you rebuild quarterly, because every automated tool has blind spots.

Sellerboard ($19–$79/month) is the standard dedicated profit tool. It ingests fees, refunds, storage, and ad spend per SKU automatically, handles COGS with date ranges (so a landed-cost change applies from the right PO forward), and its live dashboard catches margin erosion in days rather than at quarter-end. Its blind spots are the ones every API-based tool shares: it only knows the COGS and overhead you feed it.

Seller Central’s SKU Economics report (under the Analytics menu) is free and underused. It breaks out sales, fee types, fulfillment costs, storage, and ads per SKU over your chosen window, straight from the source of truth. It’s clumsy for trend analysis and knows nothing about your COGS, but as a quarterly audit against your tool’s numbers it’s exactly what you want — when Sellerboard and SKU Economics disagree, one of your inputs is stale.

The quarterly spreadsheet is where judgment lives: current landed costs, honest overhead allocation, return-rate updates per SKU, and the kill/fix/scale call itself. Tools report; the spreadsheet decides.

The Growth Trap: Revenue Up, TACoS Discipline Gone, Cash Down

Here’s the failure pattern that makes profitability analysis urgent rather than academic. A seller scales from $2M to $4M. Revenue announcements feel great. But along the way, TACoS discipline slipped — launch campaigns never got dialed back, bids crept up to defend rank, every new SKU got “temporary” aggressive spend that became permanent. TACoS drifts from 8% to 14%. Six points of margin, gone — on the whole business, roughly $240K a year at the new scale.

Meanwhile the growth itself consumes cash: doubling revenue means roughly doubling inventory, and every incremental $1M of run rate ties up $150K–$250K in stock before Amazon pays out a dollar. So the P&L weakens at the exact moment the balance sheet demands more. This is how sellers grow into insolvency — revenue up 100%, bank account down, credit lines maxed to fund POs, and no single month where anything obviously broke. The unit economics eroded a few basis points at a time while everyone watched the top line.

The defense is boring and absolute: a TACoS ceiling per SKU tied to its contribution margin, reviewed monthly, with ad spend as an input to unit economics rather than a separate “marketing” bucket. Growth is only worth funding when the contribution math says each incremental unit pays for its own capital. That principle — unit economics before scale — anchors our entire growth strategy framework, because every growth lever is just a multiplier on whatever per-unit number you feed it.

Making It Operational

Profitability analysis fails as a one-time project and works as a standing rhythm: automated SKU-level tracking daily, dashboard review monthly, full cost-stack rebuild and kill/fix/scale pass quarterly. Most sellers who do it for the first time find one to three genuinely unprofitable SKUs and 3 to 5 points of recoverable margin — money that was leaving quietly every single day.

If you’d rather have this built and run for you — the cost stack modeled, the fee errors clawed back, the ad allocation done honestly, and the quarterly decisions put in front of you with the math attached — that’s a standing part of how our full-service Amazon management engagements operate.