Supplier on-time-in-full: seven ways to measure it wrong

What the metric claims to say

On-time-in-full, or OTIF, is the share of deliveries that arrived by the promised date and in the promised quantity. It sounds like a fact. It is really a ruling, because somebody has to decide what “the promised date” means, how late is late, and what counts as one delivery. Here are the seven places those rulings usually go wrong.

1. Scoring against the latest promise

A purchase order is placed for 1 March. On 20 February the supplier moves it to 15 March. On 10 March they move it to 29 March. It arrives on 29 March.

Measured against	Date	Result
The original commitment	1 March	28 days late
The first revision	15 March	14 days late
The latest revision	29 March	On time

Score against the latest promise and a supplier who slips every order can post a perfect record. The date that matters is the one they agreed to when the order was confirmed. Keep the revisions as history, because a supplier who revises often is telling you something, but do not let them overwrite the baseline.

2. A window that nobody wrote down

Is on time the exact day? A day either side? The same week? Each choice gives a different percentage from the same deliveries. Pick one, write it into the supplier terms, and apply it to everybody. Early deliveries need a rule as well. A pallet that arrives three weeks early is not “on time” in any sense your warehouse or your cash will agree with.

3. Counting at the wrong level

Take an order with 40 lines. Thirty-nine arrive complete and on time; one is short.

Both numbers are true. The order-level figure tells you the production line stopped because one part was missing. The line-level figure tells you the supplier is usually reliable. If you only look at one, you will make the wrong call about either the supplier or the problem. Report both and say which is which.

4. Averaging across suppliers

A company-wide OTIF of 92% can hide five suppliers at 99% and one at 60%. The average is never the thing you need to act on; the individual record is. Break it out by supplier, and then by part where it matters, because a supplier can be dependable on commodity lines and poor on the one custom item you cannot buy elsewhere.

5. Treating “on time” and “in full” as separate

If a supplier is 95% on time and 95% in full, what share of orders were both? It is somewhere between 90% and 95%. If the misses fall on different orders, only 90% are both on time and complete. If they fall on the same orders, it is as high as 95%. You cannot know which from the two headline figures, which is exactly why a single combined metric measured order by order is more honest than two flattering components.

6. Ignoring how far off the misses are

Two suppliers each deliver 80% of orders on time. One is late by a day on the other 20%. The other is late by five weeks. A percentage cannot distinguish them, but your safety stock and your customers can. Pair the rate with the average (or worst) days late of the misses. For planning, the spread of lead times matters more than the average; there is a worked example in how to set safety stock.

7. Rewarding the score, not the supply

Push a supplier hard on OTIF and the cheapest way for them to comply is often to stop quoting realistic dates. Padded lead times produce excellent on-time numbers and a lot of stock. On the buyer’s side, chasing the best OTIF can concentrate all the volume on two suppliers, which looks great until one has a bad quarter. Read the metric next to quoted lead time, price and your dependence on that supplier, never alone.

In the software

Procurement Control Tower measures one part of this: on-time delivery against your own commitments, the date promised versus the date shipped, per supplier. It measures dates only. Shortfalls and defects are not measured, so it does not report an OTIF figure and does not call anything by that name. See what it covers.