Perfect order rate

How the perfect order rate is worked out

There are two ways, and they do not give the same answer.

The first is to count: take the orders in the period that had no defect of any kind, divide by the total, and multiply by one hundred. This is the honest version, and it needs a system that records the four defect types against the order.

The second is to multiply the four component rates together. It is quick, it works from figures you probably already report, and it is slightly generous, because it assumes the four failures are independent. They are not. An order that is late is more likely to be the one that was short in the first place, so the real overlap makes the counted figure a little lower than the multiplied one.

There is no good number

This figure is severe by construction, and comparing it between operations usually compares definitions rather than performance. One warehouse counts a split shipment as incomplete; another counts it as complete because everything eventually arrived. One counts a damaged carton, another only a damaged product.

So fix the definition, write it on the report, and read the trend. Then read the components, because the headline tells you only that something went wrong somewhere. Working on the strongest component is a common and expensive mistake: in a multiplication, the weakest link sets the result.

In practice

A distributor measures the four components over a month: complete 98.5 per cent, on time 97 per cent, undamaged 99.4 per cent, documents correct 99.1 per cent.

Multiplied: 0.985 times 0.97 times 0.994 times 0.991 gives 0.9416, or 94.2 per cent. Counted directly from the order records, 938 of 1,000 orders had no defect at all, so 93.8 per cent. The gap of four orders in a thousand is the overlap the multiplication cannot see.

On-time delivery at 97 per cent is the weak component, and it is worth three times as much as the other three put together. Lifting on time to 99 while leaving everything else alone takes the rate to 96.2. The figures here are an illustration.

What moves the figure

• Availability at the moment of picking. An order cannot be complete if a line is not there. This is where fill rate decides your ceiling.
• The carrier. On-time is usually the weakest of the four and the one with your least direct control. It is also the one a different service level or a second carrier can change quickly.
• Picking and checking. Picking accuracy feeds the complete component directly.
• Packaging. Damage is the quietest of the four: nobody logs a dented carton the customer accepted, and then it returns a week later as a complaint.
• Documents. Customer-specific labels, certificates, batch documentation and customs paperwork. Dull, and the cause of a surprising share of otherwise faultless orders failing.
• What counts as a defect. A tightened definition looks like a decline. Change the definition deliberately, and restate the previous period on the new basis.