SELLER GUIDE

Amazon FBA Profit After PPC: A Seller's Reconciliation Guide

Imagine your console says a campaign generated $5,000 in sales last month, yet the bank balance did not improve. That is not automatically a reporting error or a failed campaign. A seller deposit includes money from different transaction dates, holds, refunds and credits. The ads report assigns credit according to its reporting rules. Neither report on its own accounts for the cost of buying the inventory. The practical question is: if this SKU sells again at this price and I pay about the same ad cost to make it happen, how much money is left to cover the business?

Begin with the decision, not the dashboard

Start with a single marketplace and SKU, and keep revenue and expenses on a consistent tax and currency basis. Choose the decision you are making: pricing, reorder size, bid targets, or whether to keep advertising a thin-margin product. A precise number for the wrong decision is worse than a thoughtful range. For a time-sensitive reorder, a base case plus two plausible downside scenarios is more actionable than waiting for every return to settle.

Build the cost stack in a useful order

1. Record what you expect to retain from the item after seller-funded discount. Separate tax you collect for remittance if that matches your accounting basis. Confirm what Amazon actually uses as the referral-fee base; it does not have to be the same field you call net revenue.

2. Add product unit cost at your chosen endpoint. Include prep and freight once only. If a landed cost already reaches the FBA receiving location, do not add that same inbound transport as a second marketplace line.

3. Record actual marketplace fees where available. For an unlaunched product, use official fee resources for the marketplace, price band, category, dimensions, and packaged weight. The US and India descriptions of selling and fulfillment fees differ; there is no universal referral or fulfillment percentage.

4. Allow for inventory holding, refunds and losses. Where refunds are handled in the revenue line, reserve only additional variable loss that has not already been subtracted. Track recoverable units, unsellable units, replacement packaging and removal separately if they change your decision.

5. Add ad costs to make the final bridge. Keep fixed overhead and income tax outside contribution, or make allocations explicit when building a full operating forecast.

A common mistake is mixing the cost for a single shipped package with revenue for a bundle or multipack. Label the seller SKU, quantity basis and marketplace in your working sheet. Another mistake is using selling price from one week and ad spend or refund rates from a promotion in another week. Align the dates or label a deliberate stress scenario.

One US worksheet from beginning to end

Consider a hypothetical $32 item with $10 landed inventory, $4.80 marketplace referral, $5.50 fulfillment, $0.70 storage/prep not included in the landed quote, and $1 reserve for incremental returns. Non-ad variable cost is $22 and pre-ad contribution is $10, or 31.25%. These are assumed figures, not a current fee rate or size-tier example.

Suppose a matched campaign shows $1,600 attributed sales and $400 spend. If those sales represent 50 orders at the same $32 price and retained revenue matches the reported basis, the model assigns $8 ad cost per attributed unit. That leaves $2 per advertised sale, or 6.25%. The corresponding period contribution on the attributed sales is $500 before ads and $100 after them. You do not need to identify which exact order paid each click to use this planning model.

Now reduce the item revenue to $30 during a promotion and, for this stress case only, hold all $22 of non-ad cost and $8 of ads fixed. After-ad contribution goes to zero. If the referral fee actually changes with sale price, recalculate it rather than silently holding it constant for the live decision. Notice what this model does: it tells you which assumption to verify before you extend the promotion.

India example: separate charges and cash timing

For an Amazon.in worksheet, assume retained comparable revenue of ₹1,200, product cost ₹510, referral ₹100, closing ₹25, fulfillment/weight handling ₹140, and ₹65 of combined non-duplicated inbound, storage and expected return costs. That leaves ₹360 pre-ad contribution, or 30%. At 20% ACOS on a matching revenue basis, modelled ad cost is ₹240, leaving ₹120 or 10%. The closing, referral and handling amounts are hypothetical, not the schedule. Confirm actual tax handling, applicable fee bases and credit eligibility with your records and appropriate advisers. An amount withheld or recovered later may change the cash picture without being the same as a permanent cost in this decision model.

Reconcile the period before increasing spend

Build a three-column comparison: forecast, settled seller report, and adjustment explanation. Reconcile returned orders and reimbursements separately. Calculate testing-window contribution from retained sales and ads, then compare to the forecast. Even two weeks of representative evidence can reveal that your packaging assumption or return reserve is wrong; a rare bulk order can still distort a small sample.

Use a normal period to estimate unit economics. Add a separate cohort for discounts or influencer traffic. For a portfolio, avoid applying a profitable SKU's margin to every advertised product. A single branded campaign can report attractive ACOS while harvesting sales which might have arrived anyway, so watch total sales, costs and contribution alongside ads. Treat credited orders as attributed orders, not a causal count.

If the final result is barely positive, set a minimum contribution buffer and an inventory limit instead of raising bids to the mathematical maximum. Reassess after the attribution window and material returns mature. Keep a log of SKU dimensions, fee verification date, estimate source, material reclassifications and campaign mixes. A reliable comparison survives a handover to another employee without relying on anyone's memory.

Frequently asked questions

Should PPC be divided by all sales?

Divide by attributed units for the advertised-unit model, or by all units for a deliberately blended portfolio view. Do not call the second measure ACOS.

Why do refunds change the picture later?

A refund may be reported after the original sale. Keep dates explicit and avoid counting the same loss in revenue and a reserve.

Can a negative model still support a launch?

A limited test can be a deliberate decision, but document its budget and exit criteria rather than treating loss as proof of future gains.

How often should I refresh fee assumptions?

At minimum when product size, weight, category, price, or marketplace terms change; periodically reconcile the model with recent real charges.

Is the break-even margin my bid target?

No. It is mathematically neutral on the inputs. Allow room for uncertainty, overhead and a profit objective.

Sources

Related tools and guides