"Which products were profitable last quarter but are operating at a loss this quarter" is answered by computing contribution per SKU for both quarters and listing the ones whose sign changed. The harder part is the second question, why, and there are five usual causes: a fulfillment fee band change, a landed cost increase surfacing under FIFO, ad spend creep, a return rate spike, and price compression. Most flipped SKUs have two of the five at once, and the fix depends on which two.
Below, three illustrative SKUs, each flipped by a different pair, with the arithmetic shown so the workflow can be rebuilt in a spreadsheet before deciding whether to let an AI run it.
Fee band change. Amazon sets FBA fulfillment fees by product size tier and shipping weight. A packaging change, a heavier insert, or a dimension remeasurement moves a SKU across a threshold, and every unit after that carries the higher fee. The seller did not change the price and did not notice.
Landed cost increase under FIFO. COGS is charged from the oldest inventory layer on hand. A PO that landed in April with a freight surcharge and new duty does not touch margin until the March layer sells out, which might be July. The margin drop shows up a quarter after the cost went up. The mechanics are in why one SKU has four different costs.
Ad spend creep. Bids drift up, budgets get raised during a launch and never lowered, a new campaign is added for a keyword that never converts. Spend per unit rises while the unit count barely moves.
Return rate. A bad production batch, a listing that overpromises, or a packaging change that breaks in transit. Refunds rise, and if returned units grade unsellable, COGS is lost too.
Price compression. A competitor cuts price, the repricer follows, and the referral fee savings on the lower price cover a fraction of the lost revenue.
Illustrative numbers. Referral fee at 15 percent, Amazon's published rate for Home and Kitchen on sell.amazon.com. Fulfillment fees are illustrative.
Q2, at $19.99:
Q3, price cut to $16.99 to hold the Featured Offer, and a new retail box that moved the SKU up a weight band:
Per-unit decomposition of the $3.37 swing: price minus $3.00, referral fee saved plus $0.45, fulfillment band minus $0.95, ad spend per unit flat, refunds per unit plus $0.12. The price cut did $2.55 of net damage and the box did $0.95. Volume rose 100 units, which at a negative per-unit contribution made the loss larger, not smaller.
Q2, at $34.99:
Q3, same price, the old inventory layer sold out in late June and a PO that landed with a freight surcharge and reclassified duty began to sell through:
The swing is $5,229.84. Landed cost rose $5.50 per unit on 880 units, or $4,840.00, which is 92.5 percent of it. Nothing about this SKU's sales, fees, ads, or returns changed in any meaningful way. The cost went up in April and the P&L found out in July, which is how FIFO works and why a dashboard with a single cost field per SKU missed it.
Q2, at $14.99:
Q3, a launch-style campaign added for a Prime Day push and left running, and a production batch with a defective clasp:
The swing is $5,929.34. Extra volume added $464.15 at Q2's per-unit economics before ads and refunds. Ad spend took $3,500.00, refunds took $2,263.49, and the write-off took $630.00. Two causes, both fixable, neither visible on a Business Report that shows units up 5 percent.
Step one is the one that fails in most seller accounts, because the ledger holds Amazon as a lump-sum deposit and COGS as a monthly average. ConnectBooks reconciles each settlement into QuickBooks or Xero at the transaction level and runs FIFO per unit with landed cost, so the SKU-level profit report for each quarter exists before anyone asks the question.
Crunch, the analytics AI inside ConnectBooks, takes the question as typed: which products were profitable last quarter but are losing money this quarter. It runs steps one through three across the whole catalog, names the driver behind each line in step four, and recommends a next step in step five. For the three SKUs above it would return SKU-A with price and fee band, SKU-B with the cost layer, and SKU-C with ads and returns, each with the dollar swing attributed.
A follow-up question, "was SKU-B's margin better in any month this year," gets a month-by-month answer on the same data, which is how a seller finds the exact week the cost layer turned over.
It cannot see the PO before it posts. If SKU-B's freight surcharge was never allocated to the landed cost, the COGS line is understated and the AI reports a smaller swing than the true one. The reconciliation has to carry landed cost, or the analysis inherits the gap.
It cannot tell you why the clasp failed on SKU-C. It sees a return rate and a reason code. The batch inspection and the supplier call are yours.
It cannot judge whether SKU-A should hold $16.99 for a strategic reason. It can show the price cut cost $2.55 per unit net. Whether the Featured Offer is worth that is a decision.
It takes no action. Crunch does not raise the price, cut the campaign, or file a removal. It answers and recommends. The Seller Central login stays with the seller.
Run the three-SKU arithmetic above on your own top 20 for the last two quarters. The ones that flipped are the ones to ask about first.
Running an e-commerce business comes with plenty of challenges, but ConnectBooks is here to make your life easier. With real-time insights, seamless integrations, and detailed tracking of your profitability and inventory, you can stay ahead of the game. Whether you’re selling on Amazon, Shopify, Walmart, TikTok or eBay, ConnectBooks helps you manage your finances with 100% accuracy and confidence, so you can focus on growing your business.
Ready to level up? Start making smarter, data-driven decisions every step of the way. Try ConnectBooks Free Today or Schedule a Demo
