Harrowell

For UK distributors, wholesalers and manufacturers

Order accuracy rate: formula, KPI and how to measure it

Order accuracy rate is the share of orders that reach the customer exactly as they asked: right products, quantities, price, address and day. The formula is orders delivered without an error, divided by all orders delivered, times 100, and the result is only as honest as your count of errors.

Book thirty minutes

Or send us one order and see what we would flag on it.

The formula

Order accuracy rate, line accuracy and errors per 100 orders.

Order accuracy rate
Orders delivered minus orders with at least one error, divided by orders delivered, times 100.
Line accuracy
Order lines delivered right, divided by order lines delivered, times 100. Useful when orders run to many lines, because one wrong line marks a whole order as wrong in the first measure.
Errors per 100 orders
One hundred minus the order accuracy rate. This is the figure the order desk calculator asks for.
What counts as an error
Any order the customer received differently from what they asked for, or that you had to correct after confirming it.

Worked example

One month on a distributor's order desk.

Orders delivered
2,400
Credit notes
27 orders had a credit note for a wrong item, quantity or price.
Re-deliveries
6 orders were re-delivered. 4 already had a credit note, so 2 more orders.
Phone corrections
9 orders were corrected after the customer rang about the confirmation.
Orders with an error
38
Order accuracy rate
2,362 divided by 2,400, times 100: 98.4%.
Credit notes only
2,373 divided by 2,400, times 100: 98.9%. It misses 11 of the 38 errors.

Illustration with invented data.

Counting errors

Count orders, not credit notes, and count the ones nobody wrote down.

  1. Count each order once. One order with three credit notes is one order with an error.
  2. Include re-deliveries, even when no credit was raised.
  3. Include corrections after confirmation: the customer who rings to say they asked for 2 cases, not 20, before the van leaves. These never reach a credit note.
  4. Include orders that were never entered, found when a customer chases a delivery that was never coming, and the same purchase order keyed twice.
  5. Include orders where the desk guessed at an unclear line and guessed wrong. Log the customer's own mistakes in a separate column.
  6. Keep goodwill credits and damage in transit out of the count, and log them on their own.
  7. Tag each error with where it started: keying, picking, stock or delivery. Order desk accuracy and picking accuracy are different numbers with different fixes.

Setting a baseline

Three months from your records, two weeks from a log.

Credit notes and re-deliveries are already in your ERP. Pull three months of them, match each to its order, and tag the cause. That gives a first figure quickly, but it is the 98.9% in the example: it misses the errors caught by phone.

For the rest, keep a log on the desk for two weeks: every correction after confirmation and every missed or duplicated order, with the route each order came in by. Of the UK distributors with online ordering whose websites we read, 48% still tell customers to phone, email or use a rep as well (our research note), so split the count by route. You will see whether phoned, emailed and portal orders go wrong at different rates.

Add the two together and that is the baseline. Measure it the same way every month. If the rules for what counts change, the trend means nothing.

What it costs you today

Errors per 100 orders, times what one costs to fix.

In the example, 38 errors in 2,400 orders is 1.6 in every 100. Put your own figure into the order desk calculator with what one error costs to fix: the credit, the re-delivery and the time on the phone. It works out the cost a year from your own figures and doesn't guess at savings.

Speed is the other half of the desk. Order processing time is measured from the same two-week log.

What you get from us

Order errors are one of the two numbers we report.

We build AI into one division of a business and are measured on that division's own numbers. On an order desk, emailed, PDF and photographed orders are read and checked against the customer's prices and stock, and wait as a draft order for one click. Your people approve every draft.

  1. 30 minutes

    Call with you

    How orders reach you and what goes wrong with them.

  2. 1 hour

    Session with your desk

    We go through real orders and recent credit notes with your desk. The same day, a one-page verdict: the workflow, the KPI it moves, today's baseline and a fixed price. Or a written no.

  3. A fixed number of weeks

    Build

    Connected to your inbox and your ERP. Your team approves every draft order from the first day.

  4. Every month

    Run and report

    Order errors and order processing time, counted the way this page describes, against the baseline. Stop at any report.

How orders are read and checked: AI order entry for UK distributors and wholesalers. Nothing is subcontracted.

Questions

What people ask about order accuracy.

What does order accuracy rate mean?
It is the share of orders delivered exactly as the customer asked: the right products, quantities, units, price, address and delivery day. An order with any error counts as inaccurate.
What is the order accuracy rate formula?
Orders delivered minus orders with at least one error, divided by orders delivered, times 100. With 2,400 orders delivered and 38 with an error, that is 98.4%.
What is a good order accuracy rate?
We have not found an independent, sourced benchmark for UK order desks, so we don't quote one. Compare your rate with your own baseline each month, and split it by how orders arrive.
Is order accuracy the same as picking accuracy?
No. Picking accuracy covers the warehouse. Order accuracy covers the whole order, including keying, price and delivery. Tag each error with where it started so the right team fixes it.
Should errors caught before dispatch count?
Yes, if the order had already been confirmed to the customer. A correction by phone still costs time, and leaving it out makes the rate look better than the desk is.
How often should we measure order accuracy?
Monthly, with the same counting rules every time. Keep a two-week desk log when you set the baseline, and again whenever you change how orders are entered.

Next step

Thirty minutes on your order errors.

Tell us how orders reach you and where they go wrong. If we don't find anything worth doing, we'll tell you and leave. Or send us one order and see what we would flag on it.

The Open Build. We’ll build one UK supplier an order desk system, free and in public, measured on its own numbers. Everyone who applies gets a one-page read of their own workflow. Applications close at 5pm UK time on Friday 6 November. See the Open Build.