Sales velocity is simply how fast a product sells — units per week or per day. It's the engine behind every reorder decision: velocity determines when a product hits its reorder point and how much you should buy. Get velocity right and the rest of inventory planning is arithmetic; get it wrong and every downstream number — reorder points, order quantities, days of stock — is wrong in whichever direction costs you more.
The problem with a plain average
The obvious way to measure velocity is "units sold last year ÷ 52." The trouble is that a plain average treats every week as equally informative, and your business isn't static:
- A product that's picking up — say a Pima Tee that did 2/week all year but 5/week the last month — gets planned at its stale, low rate. You reorder late and thin, and stock out just as it's taking off.
- A product that's slowing down gets planned at its old, high rate. You keep buying like it's still hot and pile up dead stock.
- A one-time spike (a feature, a flash sale) a year ago drags the average around long after it stopped saying anything about demand.
The fix is a recency-weighted average: recent weeks count more than older ones, so the number tracks what the product is doing now while still smoothing out day-to-day noise.
Three numbers, one health check
From weekly velocity you derive the two numbers you'll actually use day to day:
| Number | Formula | What it tells you |
|---|---|---|
| Daily sales | weekly velocity ÷ 7 | The drain rate — feeds all reorder math |
| Days of stock | available ÷ daily sales | How long the current pile lasts |
| Reorder point | daily sales × (lead time + safety days) | The stock level that means "order now" |
Note the input: available stock (free to sell), not on-hand — units committed to unfulfilled orders can't cover future demand (the difference explained).
Clean inputs: count demand, not noise
Velocity should reflect genuine demand, so measure it from your order history with two corrections: net out refunds that restocked, and exclude cancelled orders. Otherwise a fraud cancellation or a bulk return quietly inflates the sales rate — and your next order with it. A rolling year of history is the sweet spot: long enough to capture seasonality, short enough to stay relevant.
One velocity isn't enough for seasonal products
A product with a real busy season has two meaningful rates: its everyday pace and its peak pace. Plan year-round on the blend and you'll be short every peak; plan year-round on the peak and you'll carry bloated stock all off-season. The answer is to use the peak rate when the busy window approaches and the blended rate the rest of the year — which is what Stockwik does automatically, detecting seasonal patterns from each product's own history with nothing to configure (how it works).
Putting velocity to work
In practice, velocity turns into action through the reorder math covered in our reorder-points guide: reorder point = daily sales × (lead time + safety days), order-up-to = daily sales × days of cover. Stockwik computes a recency-weighted velocity per variant from a rolling year of Shopify history and keeps all of it current on every sync — so the M/White tee and the XS/Olive tee each get planned at their own pace, not the style average.
FAQ
What is sales velocity?
How fast a product sells, expressed in units per week or per day. It's calculated from sales history — ideally weighted toward recent weeks so it reflects current demand.
How do I calculate days of stock?
Available units ÷ daily sales. A product with 30 available selling 2 a day has 15 days of stock.
Why is my reorder suggestion suddenly bigger?
Usually one of two reasons: the product's recent sales pace picked up (recency weighting responds fast), or a seasonal product is entering its peak window and is being planned at its peak rate.
Does velocity work for new products?
Not until they've sold — no history means no rate. Pin a manual minimum based on a comparable product, then let calculated velocity take over after a few weeks of sales.
