Multiply the number of stockout days by the daily sales velocity you would have had if you had stayed in stock. If a SKU was selling 15 units a day before it ran out and it was out for 10 days, your lost sales are about 150 units. Use the pre-stockout velocity as the baseline, and adjust it up if the product was trending up or heading into a busy season.
What velocity should I use to estimate lost sales?
Use the daily velocity from just before the stockout, not a long trailing average. A trailing average that includes the stockout days is dragged down by the zeros and understates the loss. Take the average daily units from a stable period right before you ran out, then adjust for any upward trend or seasonal ramp you were entering.
T. Brian Jones is co-founder and CTO of Inventory Hero. He leads the engineering behind its Amazon data pipeline, demand forecasting, and the AI platform that lets sellers talk to their live inventory, sales, and supplier data in plain language.
No. Lost sales are the direct units you missed while out of stock. They are the first and most measurable part of a stockout's cost, but the full cost also includes lost organic rank, a lower IPI, and the expense of recovering, which are indirect and usually larger. Calculate lost sales first, then treat it as the floor of the total cost.
Read article
To calculate lost sales from a stockout, multiply the days you were out by the daily velocity you would have had. Put another way, lost sales are the units you would have sold if you had stayed in stock: your stockout days times the baseline velocity. The short version: the formula is simple, but the accuracy lives in the baseline velocity you pick, which should be the rate from just before you ran out, adjusted for any trend or season. Below is the formula, a worked example, and how to get the baseline right.
Lost sales = stockout days x baseline daily velocity
Stockout days is how long the SKU was unavailable; see stockout days for finding that number precisely. Baseline daily velocity is the rate you would have sold at over that window. The whole estimate turns on that second input.
Say a SKU was selling 15 units a day in the stable weeks before it ran out, and it was out for 10 days:
Input
Value
Stockout days
10
Baseline daily velocity
15 units/day
Lost sales
10 x 15 = 150 units
At a 12 dollar contribution margin, that is 150 x 12 = 1,800 dollars in missed margin. That is your direct lost-sales figure, and the starting point for the full cost of the stockout.
The mistake that wrecks this estimate is using the wrong velocity:
Do not use a long trailing average that includes the stockout. The zero-sales days drag it down, so you understate the loss. Use a clean pre-stockout window.
Use a stable recent period. The average daily units from the two to four weeks before you ran out, excluding any earlier out-of-stock days, is a good baseline.
Pull it from your data, not memory. Business Reports (units ordered by ASIN, by day) gives you the clean pre-stockout rate.
The goal is the velocity you genuinely would have sold at, which is why a recent, clean window beats a blended long-run number.
A flat baseline undercounts if the product was moving:
Upward trend. A SKU whose velocity was climbing would have sold more than its recent average during the stockout. Nudge the baseline toward the recent slope, not the flat mean.
Seasonal ramp. Running out heading into Q4 or a category's peak means the velocity you lost was the higher seasonal rate, sometimes far above the trailing average. To reconstruct it, pull the same two-to-four-week window from last year's Business Reports for that ASIN, then apply your current year-over-year growth: if you were running 20 percent above last year's pace, apply that multiplier to last year's peak-window velocity. Concretely, a SKU running 15 a day in September but expected at 22 during the Q4 stockout window lost 10 x 22 = 220 units, not the 150 a flat baseline would show. That gives a grounded seasonal baseline instead of a guess.
A launch or promo you were running. If a campaign was driving demand, the lost velocity was the promoted rate, not the baseline.
These adjustments matter most exactly when the stockout hurts most, so do not skip them for the SKUs heading into their busy season.
A handful of errors show up repeatedly, and most of them understate the loss:
Estimating from memory instead of the report. Guessing your pre-stockout pace instead of pulling the clean window from Business Reports; the guess is almost always low. (The contaminated-trailing-average version of this is the baseline mistake covered above.)
Missing the stranded and suppressed days. If you only count true zero-stock days, you undercount your stockout days, and every missed day is missed sales.
Ignoring the trend or season. A flat baseline undercounts a SKU that was climbing or entering peak, which is exactly when the loss is largest.
Valuing lost units at revenue, not margin. Lost sales cost you the margin, not the full price; use contribution margin per unit so the dollar figure is real.
Forgetting it is only the floor. Stopping at lost sales ignores the larger recovery and rank cost. Note it as the direct line and carry it into the full cost.
Avoid these and your estimate reflects what the stockout actually cost, which is usually more than the first guess.
Lost sales are stockout days times the velocity you would have had, and the accuracy comes from choosing a clean, recent, trend-and-season-adjusted baseline rather than a stale average. Calculate it per SKU to size the direct hit, then remember it is the floor of the true stockout cost. For preventing the next one, see Amazon FBA stockout prevention.