Stock accuracy scorecard

How this is worked out

Accuracy here is measured in units, not in lines that matched.

accuracy = (1 − (units out ÷ units Shopify said)) × 100

An overcount counts the same as an undercount. Finding forty units you did not know you had is not a happy accident. It means something moved without being recorded, and the next thing it does is break an order.

A worked example

Shopify said you had 2,000 units of a line. You counted 1,960.

You are forty units out. Forty divided by 2,000 is two percent, so your accuracy is 98 percent. That grades as an A.

Two percent sounds small. On a line you sell twelve of a day it is three days of stock, which is most of a delivery window.

The grades

A, at 98 percent or better. Order against these numbers with confidence.

B, from 95 to 98 percent. Usable, but count your fast movers weekly.

C, from 90 to 95 percent. Bad orders are being placed already.

D, from 85 to 90 percent. The count is the problem, not the stock.

F, under 85 percent. Start again at goods in.

What to do about a bad score

Almost nobody has one cause. The usual mix is receiving, returns and damage that nobody wrote down. The article on why stock counts drift works through all six in order.

Start at goods in. A delivery booked in as ordered rather than as counted puts the error in on day one. If you want the gap in money rather than in units, run the same count through the shrinkage calculator.

Then count more often and smaller. Weekly cycle counting on your fastest lines catches a problem while you can still remember what caused it.