Key takeaways
  • OTIF counts an order as successful only if it arrived on time and complete. Partial credit is the whole point of not giving any.
  • The score depends heavily on three definitional choices: which date, what tolerance window, and what counts as a line.
  • Measure against the customer's original requested date, not the date you promised after negotiating.
The definition that flatters everyone
Measuring against your own confirmed date

An order requested for Tuesday, confirmed for Friday and delivered Friday scores 100% against the confirmed date and 0% against the request. The second is what the customer experienced, and the gap between the two measures is often where the real problem lives.

Calculating it

OTIF is the number of orders delivered both on time and in full, divided by total orders. An order failing either test fails entirely. The strictness is deliberate: a delivery that is 95% complete still stops a customer's line if the missing 5% is the part they needed.

The three choices that move the number

  • Which date. Customer request date, your confirmed date, or the last agreed date. Request date is the honest one.
  • Tolerance window. Same day only, or a window such as one day early to zero days late. Early delivery is a failure in many supply chains because it creates storage the customer did not plan.
  • Unit of measure. Order, line, or delivery. Line-level is harsher and more informative than order-level.

What to measure alongside

  • Separate the on-time and in-full components, because the causes and owners differ.
  • Track the reason for each failure, categorised, so the data supports a Pareto.
  • Track the size of the miss, since one day late and three weeks late are not the same event even though both score zero.
  • Track the gap between requested and confirmed dates, which measures how much lateness is being absorbed at order entry.

Where failures usually originate

Rarely in despatch, which is where the failure is recorded. More often in material availability, in a schedule that was never achievable, or in a promise made without checking capacity. Recording the failure reason at the point of despatch is what allows the cause to be traced back upstream.

Using it with suppliers

If you measure your suppliers on OTIF, publish the definition you are using and the underlying data. Supplier scorecards that arrive as a single percentage with no detail generate disputes rather than improvement, and the disputes are usually about definitions rather than performance.