Why most Amazon sellers track profit wrong

There’s a difference between “knowing what your profit is” and “having a number on a dashboard that you call profit.” Most Amazon sellers operate with the second and assume it’s the first. It usually isn’t.

The errors aren’t malicious or even careless — they’re the natural result of using whatever number Amazon hands you, plugging it into a spreadsheet, and trusting the math. But Amazon’s numbers come with timing quirks, allocation gaps, and definitional traps that make the dashboard lie in consistent, directional ways. Almost always toward “more profitable than reality.”

This article walks through the five most common mistakes, why each one happens, and how to fix it. None of them require an accountant. All of them require deciding to look.

Mistake 1: Confusing order-date profit with settlement-date profit

When did you make money on an order placed January 28th and shipped February 1st? Most sellers answer “January” because that’s when the order was placed. But Amazon’s settlement cycle pays you for it in mid-February, and the FBA fee, ad spend allocation, and returns risk all show up against the February statement.

If your accounting treats the order as January revenue but the costs hit February’s report, you’ve created a fictional January that looks more profitable than it was and a February that looks worse.

The right pattern: match revenue and costs to the same period. Either both go to order date, or both go to settlement date — but not split. For tax and bookkeeping purposes, settlement date usually wins because it matches actual cash flow.

Mistake 2: Using “average daily sales” without trend adjustment

To calculate days-of-supply, restock quantity, and forecast revenue, you need a per-day sales rate. Most sellers pull the last 30 days, divide by 30, and call it a day.

That’s correct only when sales are flat. The moment a SKU is trending up or down, a 30-day average lags the truth — and you make decisions for next month based on what happened five weeks ago.

The right pattern: use a weighted average that emphasizes recent weeks more heavily. A simple 4-week exponential weight (most recent week × 0.4, previous week × 0.3, then 0.2 and 0.1) catches trends faster without overreacting to single-day spikes.

Why this matters for profit: if you’re allocating fixed costs (storage, software, overhead) using a stale average, you’re misallocating per-unit cost on every SKU that’s been moving against trend. Fast movers look more profitable than they are. Slow movers look less so.

Mistake 3: Not allocating returns back to the originating sale

When a customer returns a product 30 days after purchase, your accounting needs to do four things:

1. Reverse the revenue
2. Reverse the FBA fee (if Amazon refunds it — they often don’t)
3. Add the return shipping cost
4. Reverse the COGS recognition

Most sellers do step 1 and stop. Some do steps 1 and 2. Almost nobody handles steps 3 and 4 correctly.

The compounding error: if your “revenue” line includes orders that were later returned, and your “cost” line doesn’t reverse the corresponding fee/shipping costs, you’re double-counting both sides. Net effect: profit looks higher than reality by approximately (return rate × fulfillment fee × order count).

Mistake 4: Treating COGS as “what I paid the supplier”

The supplier invoice is one input to landed cost — not the whole thing. The full landed cost includes:

• Supplier price per unit
• Inbound freight (sea/air/road from supplier to your warehouse or 3PL)
• Customs duties and tariffs (currently 10-25% on most categories from China)
• Inspection and QC fees
• Prep fees (poly bagging, labeling, bundling) at your 3PL or Amazon
• Inbound shipping from your warehouse to Amazon

The standard mistake: COGS = supplier invoice ÷ units. The accurate version: COGS = (supplier invoice + freight + duties + prep + inbound to FC) ÷ units delivered to FC.

The gap can be substantial. We’ve seen sellers using “supplier price” as COGS while their actual landed cost was 18-35% higher. That entire gap shows up as fake profit on every order.

Mistake 5: Ignoring time-shifted fee changes

Amazon updates FBA fees on a calendar most sellers don’t track:

• January — annual fulfillment fee adjustment (usually +3-7%)
• June — peak storage fee changes
• October — Q4 storage surcharge ($0.50–$2.40 per cubic foot, October through December)

If your profit dashboard uses a hardcoded fee schedule that hasn’t been updated for the current quarter, every calculation is wrong. The Q4 storage surcharge alone can flip slow-movers from “marginally profitable” to “loss” without anything changing about the product or sales velocity.

The right pattern: pull current fees from your Inventory and Order reports rather than from a stored table. Or set a calendar reminder to update your fee schedule on January 15, June 15, October 15.

What “right” actually looks like

Here’s what a corrected profit row looks like for a single order:

Revenue (settlement date): $24.99
Less: FBA fulfillment fee (current schedule): -$5.50
Less: Referral fee (15% of $24.99): -$3.75
Less: PPC allocated to this SKU this month: -$1.40
Less: Landed COGS (supplier + freight + duties + prep): -$7.20
Less: Returns provision (10% × ($5.50 + $0.80 shipping)): -$0.63
Less: Storage allocation (this SKU's share of monthly fee): -$0.32
Less: Software/tooling allocation: -$0.18
================================================
True net profit on order: $6.01
True margin: 24.0%

Compare to the standard calculation that gives you:

Revenue: $24.99
Less: Amazon's listed fee: -$5.50
Less: Referral fee: -$3.75
Less: COGS (supplier price only): -$5.00
================================================
"Profit": $10.74
"Margin": 43.0%

A 19-point spread between the two. Multiply by 1,000 orders a month and the difference is real money you’ve been celebrating that doesn’t exist.

The order in which to fix this

Don’t try to fix all five at once. The 80/20 sequence:

1. First: Fix landed COGS. This is the largest typical error and affects every single calculation downstream.
2. Second: Add per-SKU PPC allocation. Most leverage for active campaigns.
3. Third: Reverse returns properly. Catches month-end distortions.
4. Fourth: Use weighted-average sales velocity. Improves forecasting and per-unit cost allocation accuracy.
5. Fifth: Update fee schedules quarterly. Catches the systematic timing error.

Doing the first three of these will move your reported margin meaningfully closer to reality, often by 8-15 points on individual SKUs.

What this changes about how you run the business

The point of getting profit right isn’t intellectual purity — it’s that decisions made on bad numbers are bad decisions:

• Restock decisions — you reorder the SKUs that look most profitable, which biases toward the most over-reported ones
• PPC scaling decisions — you raise budgets on campaigns whose true ROAS is hidden
• Pricing decisions — you cut prices on SKUs that already had thinner margins than you thought
• Discontinuation decisions — you keep SKUs running long after they’ve gone negative because the dashboard still says they’re fine

Fix the math and the operational decisions get sharper automatically. You’ll be surprised which SKUs are quietly your best performers, and which ones you’ve been propping up out of habit.

(SellerPulse runs this calculation per-order, automatically, and surfaces the per-SKU truth on a profit dashboard. Flat $99/mo on the Pro plan; the first month often pays for the year just by identifying which SKUs to stop restocking.)

The actionable summary

None of these require new tools — just deciding to look. But until you do, you’re flying with an instrument that’s lying to you in known, directional ways.