How do you calculate OTIF?

Divide the number of deliveries that arrived by the promised date with the complete quantity by the total number of deliveries in the period, and express it as a percentage. A delivery fails if it is late, or short, or both. The strictness is the point: it measures what a planner actually experiences.

What is a good receiving accuracy rate?

There is no useful industry benchmark at small scale, and any figure quoted as one should be treated with suspicion. What matters is your own trend and the spread between suppliers. A supplier fifteen points below the rest of your base is a finding regardless of what the absolute number is.

Should I measure how fast staff receive deliveries?

No. Speed at the bay is the one thing you can improve instantly by checking less, and a measure that rewards it will get exactly that. If you want deliveries handled faster, remove steps from the process rather than pressure onto the person doing it.

How often should receiving metrics be reviewed?

Monthly for the trend, quarterly for supplier conversations. Weekly review of a low-volume operation mostly measures the weather: with eight deliveries a week, one bad one moves the number twelve points and means nothing.

What is dock-to-record time?

The gap between a delivery physically arriving and its receiving record existing in a form somebody else can see. It is the least glamorous measure here and often the most diagnostic, because a long gap means every record was written from memory.

Receiving metrics that are worth measuring, and three that are not

Q: Receiving metrics that are worth measuring, and three that are not

Four receiving measures are worth the effort for a small operation: on time in full, receiving accuracy, discrepancy rate by supplier, and the time between a delivery arriving and its record existing. All four can be calculated from a receiving log. Measures of individual receiver speed are worse than useless, because they reward the exact behaviour that causes expensive errors.

On time in full, receiving accuracy, dock-to-record time and discrepancy rate, how to calculate each from records you already keep, and the popular measures that mislead.

By Hamza RiveraFounder, DockCountSeptember 4, 2026

In short

Four measures, all computable from a receiving log, all about the process rather than about people.
Dock-to-record time is the one nobody tracks and the one that predicts the others.
Never measure individual receiver speed. It improves the moment somebody stops checking, which is the opposite of what you wanted.

Measurement in a small operation goes wrong in a specific way: somebody picks metrics designed for a distribution centre with two hundred staff, cannot collect half of them, and abandons the whole idea. These four need nothing you are not already writing down.

A flat illustration of a single horizontal line running across the frame with small markers above and below it at irregular intervals. — One number a month for a year tells you more than fourteen numbers for one week.

The four

Measure	How to calculate it	What it tells you
On time in full	Deliveries both on time and complete, divided by all deliveries in the period.	What planning actually experiences. Deliberately strict, because a delivery that is early and short is not a success.
Receiving accuracy	Lines received with no variance, divided by all lines received.	How often the order and reality agree. Falling accuracy across all suppliers usually means a process problem, not a supplier problem.
Discrepancy rate by supplier	Deliveries with any exception, divided by that supplier's deliveries.	Who is actually costing you time. This is the number for a renewal conversation.
Dock-to-record time	Median hours between a delivery arriving and its record existing.	Whether your records are observations or reconstructions. Above a few hours, they are reconstructions.

Why dock-to-record matters more than it looks

Every other measure on this page inherits its quality from this one. A receiving accuracy figure built from records written the following morning is a measure of what people remembered, and what people remember is that it was fine. The number will look good and mean nothing.

Worth knowing

If you track only one thing, track this. A median dock-to-record time under an hour means your data is real. Over a day, and every supplier conversation you have based on it is built on sand.

A flat illustration of two nested rectangles, the outer one wide and lightly drawn, the inner one narrow and solid, with the gap between them emphasised. — The gap between when something happened and when it was written down is where accuracy quietly goes.

Three that mislead

1
Deliveries received per hour, per person. Improves instantly if somebody stops counting. You will get the improvement and the shortages you did not catch, and you will not connect the two for months.
2
Number of discrepancies raised. Reads as a problem when it is usually a sign the process is working. Falling discrepancy counts can mean suppliers improved or can mean people stopped looking, and the raw count cannot tell you which. Use the rate, by supplier, alongside accuracy.
3
Benchmarks from somebody else’s industry. A ninety nine percent figure quoted from a distribution centre with barcode scanning at every touch is not a target for a shop taking twelve pallets a week. Compare yourself to last quarter.
4
A single composite score. Four numbers with different causes, multiplied together into one, cannot be acted on. When it moves, nobody knows why.

Reviewing them without wasting an afternoon

Monthly, four numbers, ten minutes. Trend only: this month against the last twelve.
Quarterly, split the discrepancy rate by supplier and look only at the outliers.
Investigate a move only when it holds for two periods. One month is noise at this volume.
Never present a metric without the count behind it. Fifty percent of two deliveries is not a finding.

The columns needed to compute all four are in the receiving log template, and the supplier conversation they feed is in the vendor scorecard.

Questions people ask

How do you calculate OTIF?: Divide the number of deliveries that arrived by the promised date with the complete quantity by the total number of deliveries in the period, and express it as a percentage. A delivery fails if it is late, or short, or both. The strictness is the point: it measures what a planner actually experiences.
What is a good receiving accuracy rate?: There is no useful industry benchmark at small scale, and any figure quoted as one should be treated with suspicion. What matters is your own trend and the spread between suppliers. A supplier fifteen points below the rest of your base is a finding regardless of what the absolute number is.
Should I measure how fast staff receive deliveries?: No. Speed at the bay is the one thing you can improve instantly by checking less, and a measure that rewards it will get exactly that. If you want deliveries handled faster, remove steps from the process rather than pressure onto the person doing it.
How often should receiving metrics be reviewed?: Monthly for the trend, quarterly for supplier conversations. Weekly review of a low-volume operation mostly measures the weather: with eight deliveries a week, one bad one moves the number twelve points and means nothing.
What is dock-to-record time?: The gap between a delivery physically arriving and its receiving record existing in a form somebody else can see. It is the least glamorous measure here and often the most diagnostic, because a long gap means every record was written from memory.

DockCount does this job on a phone at the bay: count a delivery against the order it was placed on, photograph anything wrong, and the record writes itself. receiving reports you can send to purchasing.