Amazon sales velocity: calculate units sold per day

Amazon sales velocity is units sold per day for a stated SKU and reporting window. Calculate it from your own order quantities and a complete calendar period. Keep fulfilled sales, returns and inventory availability visible so a change in pace does not acquire the wrong explanation.

The units-per-day formula

Fulfilled-unit sales velocity = Fulfilled units attributed to the period / Calendar days in the period

Net-unit sales velocity = (Fulfilled units − Matched returned units) / Calendar days in the period

Choose a date basis before summing units. This example uses the original order date and counts only quantities confirmed fulfilled at the reporting cutoff. A shipment-date report is also useful, but it measures a different timing basis. Do not use order dates for one period and shipment dates for the next.

Include calendar days with no orders. Counting only dates on which a sale occurred would describe sales per selling day and inflate the ordinary calendar-day pace. If a product was unavailable, retain the elapsed-day measure and show availability separately.

This definition differs from the opportunities, deal values and cycle lengths in the four-factor sales velocity formula. The shared word does not make the measures interchangeable.

Build the calculation from order lines

Amazon describes Seller Central as a place to manage orders and download reports to analyze operations and sales. Use the export available to your account and verify its column definitions before calculating; report names and layouts can differ.

Keep order identifier, order-line identifier, SKU, order date, ordered quantity, canceled quantity, fulfilled quantity and status. Bring a returns export if you want net movement. Match returned quantities to their original order lines rather than subtracting refund dollars from unit quantities.

A partially fulfilled line needs quantities, not just a final status. A file can also contain updates to an existing line. Retain the final authoritative quantity for the reporting cutoff instead of summing every historical status row as another sale.

The order quantities in the example have all reached fulfilled or canceled status. If pending quantities remain in your file, preserve them separately. The reconciliation then needs pending units as an additional category rather than treating them as sales or assuming that they were canceled.

Worked example: two SKUs across matched weeks

A US seller compares October 1–7, 2026, with October 8–14, 2026. Each window contains seven calendar days. The grouped export below uses a common reporting cutoff after fulfillment statuses have settled. Returns are matched back to the original order cohort.

Calculate the first week's SKU A measure step by step:

Rates are rounded to two decimal places. Keep unrounded values in the workbook when calculating changes or reconstructing totals.

With ordered units in C, cancellations in D and returns in F, fulfilled units in E2 are =C2-D2, net units in G2 are =E2-F2, and fulfilled daily velocity is =E2/7. For a reusable date window, the inclusive day count is =EndDate-StartDate+1 when both cells contain dates.

The checks that prove the figure

First prove quantity conservation. The first period contains 86 + 44 = 130 ordered units. Cancellations are 8 + 2 = 10, leaving 130 − 10 = 120 fulfilled units. Matched returns are 3 + 2 = 5, leaving 120 − 5 = 115 net units.

Then divide the totals by the common time base:

For a single period, add SKU rates because the days overlap: both products were offered during the same calendar window. Do not add their day counts and divide the company units by that result. Across consecutive periods for the same SKU, by contrast, add both units and days.

The exact first-period fulfilled rate multiplied by seven must return 120. Reconcile those units to the export's control total after applying the same date, SKU and status filters. Resolve differences before interpreting growth.

What the change does and does not say

Total fulfilled pace rose by 126 / 120 − 1 = 5.00%. SKU A rose by 91 / 78 − 1 = 16.67%, while SKU B fell by 35 / 42 − 1 = 16.67%. The total conceals the opposing product movements.

The second period's lack of recorded returns is not proof of better product quality. More recent orders may have had less time to generate a return. Compare return-adjusted cohorts at a consistent maturity before drawing that conclusion.

Suppose the availability log shows that SKU A was available for only five days in the first window. Its fulfilled pace per available day is 78 / 5 = 15.60, compared with 11.14 per calendar day. Label that alternative explicitly; it does not establish how many orders would have arrived during unavailable days.

Amazon's Best Sellers Rank explanation describes a ranking relative to similar products and distinguishes it from search ranking. An observed units-per-day change alone does not establish a ranking effect or reveal competitors' sales.

Use the retail sales velocity formula when the denominator also includes stores. The sales velocity glossary helps distinguish the broader terminology.

Where it goes wrong

• Orders counted as units. Use line quantities when orders contain multiple items.
• Repeated status rows added together. Deduplicate to the appropriate order-line snapshot before summing fulfilled quantities.
• Only sale dates counted. Include every elapsed calendar day in the chosen reporting window.
• Return dollars subtracted from quantities. Match returned units and keep financial refunds in a separate measure.
• Stockouts interpreted as weak demand. Review availability alongside the calendar-day result without inventing lost sales.

Covirage is not affiliated with Amazon or Microsoft.

Compare your own exported sales in Covirage

Bring a sample of your own order history and, when needed, returns and availability files to Covirage. Covirage maps the columns, and its tools compute daily SKU pace and check quantity totals; the external AI model explains the comparisons with every figure traced to its rows. Create a free Covirage account, then set up your business and bring your export.

See Covirage for Excel analysis for the supporting file-analysis approach.