An Amazon pricing error looks like the simplest cost in this series to calculate: units sold too cheap, times the gap. That number is real, and it’s the smallest one on the bill.
This is the tenth entry in our true-cost series (returns, stockouts, aged inventory, a PPC restructure, a suppressed listing, deal incrementality, a receiving delay, a negative review, an FBA remeasurement). After managing hundreds of brands through peak, we’d put the pricing error in its own category. It is the only cost in the series where the damage keeps working after you’ve fixed the mistake. Once a wrong price clears, it stays in Amazon’s systems. It’s in the lookback that decides whether you can run a deal, the reference price that decides whether you get a strike-through, and the price history chart a shopper can open for a year.
It also has the shortest fuse. A stockout takes weeks to develop. A pricing error takes a decimal point and a Saturday night.
The scenario
A $34 hero SKU doing about $80K a month, roughly 2,350 units, around 78 units a day. Contribution before advertising is about $19 a unit, so COGS plus fees come to roughly $15.
At 1:40 a.m. on a Saturday in mid-September, a bulk price upload goes in with the decimal in the wrong place: $3.40. It might be a flat file edited in a spreadsheet that auto-formatted a column, or a repricer rule with no floor, or a VA fixing a different SKU in the same file. Every one of those paths is common. No minimum price is set on the SKU in Seller Central, so nothing stops it.
The error runs until about 6:45 a.m., when someone on the team checks the app over coffee. Five hours.
Glitch prices on known brands get picked up quickly by deal-alert communities and bots. Instead of the usual 25 or so overnight units, the SKU sells 1,100 units before the listing is pulled. We use that number because it’s in the range sellers report on forums for hero SKUs caught overnight. Your number depends on inventory depth, purchase limits and how fast the price was spotted.
Layer 1: the units (the only number anyone calculates)
At $3.40, Amazon’s referral fee is about $0.51. The FBA fulfilment fee is charged in full regardless of price, so call it $6.50. COGS is roughly $7.40.
Each unit loses about $11. The direct loss on 1,100 units is about $12,100. Measured against the $19 contribution those units would normally have earned, the swing is closer to $33,000. That’s a painful number, but a finite one.
What most brands assume next is wrong: they think they’ll just cancel the orders. On FBA you mostly can’t. Seller forums are consistent that sellers cannot cancel FBA orders once payment clears, and a case filed afterward rarely recovers anything. On FBM you can cancel, but cancellations count against your cancellation rate, and buyers who got a “pricing error” cancellation leave feedback insisting you’re required to honour the price. Pick your damage: the units or the account health metric.
The units also leave your inventory. 1,100 units is about 14 days of cover gone in five hours, a few weeks before your Prime Big Deal Days and BFCM inbound deadlines. If that cover was sized to reach your next shipment, you’ve just created the stockout from our earlier deep dive, on a SKU with nothing wrong with it.
Layer 2: the deal lookback (where it gets expensive)
Amazon’s deal eligibility looks backwards at your price. Seller guidance for 2026 describes two rules. A deal price has to be at or below the lowest price the ASIN sold for in the previous 60 days, and at least 5% below the lowest price in the past 30 days. Coupons and earlier deals count toward those floors. Prime Big Deal Days promotional prices are reported as excluded from the lookback. An accidental $3.40 sale is not a promotion.
We want to be careful here. Amazon doesn’t publish how its deal systems treat a price that was obviously a mistake, whether a case can get an error window excluded, or how quickly that happens. Sellers report mixed outcomes. Check your own deal eligibility in Seller Central rather than trusting any blog, including this one.
Now look at the calendar. An error on September 13 plus a 60-day lookback runs to November 12. That covers the BFCM deal submission deadline (around October 20) and most of the run-in to Black Friday. If the lowest-price reference now reads $3.40, your hero SKU can’t qualify for a BFCM deal at any price you’d accept.
What’s that worth? On deal-supported SKUs we’ve watched, BFCM week can easily run at 2-3x normal velocity. Take a conservative 2x on a $34 SKU across the event window: about 550 extra units, roughly $10,000 of contribution, before any ranking lift from the velocity. Then add the next part.
Layer 3: rank you don’t get to defend
Everyone else in your category is running deals, so their velocity rises all through November. Yours doesn’t. Velocity is a ranking input, so the organic position you hold at full price quietly erodes against a field that’s accelerating. You don’t fall off page one. You slide three or four positions on the queries that pay for the business, then buy that position back in January with paid support at ACOS you’d have refused in September.
On a $10K-a-month ad budget running around 30% above target for six weeks, that’s another $4,000 or so. It’s the same rank recovery math as our stockout and receiving-delay deep dives, triggered by a price instead of an empty shelf.
Layer 4: the relist and the reference price
When the team corrects the price back to $34, two things can go wrong.
The relist can trigger a pricing error flag in the other direction. Amazon deactivates listings it believes are priced too low or too high relative to recent pricing. A listing that sold 1,100 units at $3.40 and then jumps back to $34 can trip the “potential high pricing error” check. The listing sits inactive while someone opens a case or adjusts min and max prices. We’ve seen a pricing error followed by a second outage caused by the fix.
The strike-through price is exposed too, though less than people fear. Amazon’s Typical Price uses the 90-day median price customers paid. A median is robust, so five hours at $3.40 won’t move it much on its own. The risk comes from the recovery. Brands often run a coupon or temporary discount to “restart” velocity after the error, and since May 18, 2026, if more than half the days in the 90-day window sit below the non-promotional median, Amazon calculates Typical Price including promotional sales. A long recovery discount can cost you the strike-through that the error itself didn’t.
Layer 5: the price history chart
Alexa for Shopping, formerly Rufus, now offers shoppers 30, 90 and 365 days of price history. The 365-day view is still rolling out in the US, UK and India, so confirm what your own listing shows. A shopper deciding whether your October deal is real can open a chart that may include a dip to $3.40. That’s hard to price, and we won’t make up a number. But on the exact shoppers you pay most to reach in Q4, the deal-sensitive ones, it undermines the “this is a good deal” framing for a long time.
The total
On a $80K-a-month SKU, from one five-hour decimal error:
| Layer | Estimated cost |
|—|—|
| Direct unit loss (1,100 units at ~-$11) | ~$12,100 |
| Lost BFCM deal contribution (conservative 2x) | ~$10,000 |
| Rank recovery at elevated ACOS | ~$4,000 |
| Stockout risk from 14 days of cover lost | depends on inbound timing |
| Relist outage + recovery discount | depends on case speed |
| Measurable total | ~$26,000+ |
The inventory, the relist and the price history layers can each add more than the direct loss did. The units were the part everyone calculated, and the smallest.
The six guardrails to set this week
1. Set a minimum and maximum price on every active SKU in Seller Central. This is Amazon’s own guardrail. A price uploaded below the minimum deactivates the listing immediately instead of selling 1,100 units. One caveat catches people out: Seller Central’s min and max exclude shipping, and some third-party repricers include it. Mismatched ranges are the most common cause of accidental deactivations, so set both systems deliberately rather than copying one to the other.
2. Base the floor on contribution, not on last year’s lowest price. The floor should be the lowest price you’d genuinely accept selling a thousand units at. For most SKUs that’s somewhere around landed cost plus fees plus a margin, not a round number someone picked in 2023.
3. Put a validation step on every bulk price upload. Before any flat file with a price column goes live, sort that column and read the top five and bottom five values. It takes two minutes. Anything more than 30% away from current price needs a second person. That’s the Q4 change freeze rule applied to the most dangerous column in the catalog.
4. Alert on price changes, not just sales. Most dashboards alert on revenue drops. A pricing error shows up as a revenue spike, which nobody wakes up for. Set an alert on any price change over 30% on your top 20 SKUs, delivered by text rather than email or Slack.
5. Write the first-hour playbook now. Who can set quantity to zero or close the listing at 2 a.m.? Do they have current access? What gets screenshotted (order count, times, the upload file) before anything is changed? Who opens the case, and what does it ask for? Deciding this after the error is how you get the relist outage.
6. Screenshot your deal eligibility and price history before October. Record today’s baseline with the date in the filename. If something goes wrong in October, you’ll want proof of what eligibility looked like before, not a recollection.
FAQ
Can I cancel FBA orders placed at a pricing error? Generally no. Seller forum reports consistently say FBA orders can’t be cancelled once payment goes through, and cases rarely recover the loss. That’s exactly why the minimum price guardrail matters more on FBA than on FBM.
Will Amazon reimburse me for a pricing error? Not for a price you or your tool uploaded. Amazon treats the price as your decision. Reimbursement exists for things like lost or damaged inventory, not your own price.
Does one error permanently kill my strike-through price? Usually not directly. Typical Price uses a 90-day median, and a few hours at a wrong price barely moves a median. The bigger risk is a long recovery discount pushing more than half the window below the median. Recover velocity with ads and a short, defined promotion instead.
Does an error price affect my deal eligibility? It can. Deal rules look at the lowest price the ASIN sold for in the lookback window. Amazon doesn’t publish whether an obvious error gets excluded, so open a case documenting the error window and check eligibility in the deals dashboard. Don’t assume it’s fine.
What’s the single most important fix? A minimum price on every SKU, set in Seller Central and reconciled with your repricer. It turns a $26,000 problem into an inactive listing and a slightly annoying morning.
Every other cost in this series comes from a decision somebody made about inventory, bids or creative. This one comes from a missing field. The brands that get hurt aren’t reckless. They set up pricing tools in a hurry, never filled in the floor, and ran into a decimal point on a Saturday in the quarter where a price has the longest memory.
If you’re looking for a team that manages every lever — creative, advertising, and operations — Velocity Sellers works with brands doing $100K+/month on Amazon. Contact us for a free account audit.
