Practical guide · By Gil Silva
OTIF in practice: measure and act
A clear definition prevents unfair comparisons and helps turn delivery failures into improvements.
In this article
What the metric aims to show
OTIF means On Time In Full. It asks whether an order met the agreed time and quantity conditions simultaneously. It reflects the customer's need to receive the expected complete order within the agreed window. Delivering part on time and the remainder later may improve punctuality alone, but is not the same outcome for a customer relying on the whole order.
Before calculating, agree on eligible orders and the reference commitment. Companies may use different units and tolerances; this requires transparency, not abandoning the metric. A percentage without a definition cannot reliably compare carriers, distribution centres or periods. Measurement rules and event quality come first.
Calculate the intersection of timeliness and completeness
This guide uses: orders delivered both on time and in full, divided by all eligible orders, multiplied by 100. Both conditions must hold for the same order. Do not add or average timeliness and completeness percentages. Do not multiply them as if failures were independent: that assumption may not fit the observed orders.
Of 100 orders, 90 arrived on time and 95 in full. If 88 met both conditions, OTIF is 88%. You need the intersection count at order level. Aggregated dashboards may not provide it. Keep records that allow each order and the events supporting its classification to be checked.
Choose a unit and preserve the rule
Operations may measure by order, delivery, line or physical unit. Each changes sensitivity: by order, one missing line fails that order; by quantity, large volumes can dilute small shortages. Make the choice visible and explain how split orders are handled.
If two views are needed, give them different names. Order-level measurement supports customer experience; line-level analysis helps locate operational problems. Do not change the unit simply to improve the number. Record the date, rationale and expected effect of legitimate changes. Unexplained comparisons across definitions can mislead performance decisions.
Define the event representing delivery
Document issue, vehicle departure, customer arrival and receipt confirmation are different events. Choose the event appropriate to the delivery agreement and document its source. Manual timestamps have limitations. Proof received days later may require a closing rule so the metric does not remain incomplete indefinitely.
Where deliveries are scheduled, define how early and late arrivals are assessed. Early is not always compliant if the customer cannot receive then. Reflect the agreement and operational need, and communicate the rule before the measured period rather than adjusting it after an unfavourable result.
Preserve the original commitment and record changes
Rescheduling is normal, but silently replacing the original date can hide failures. Keep original and revised commitments in separate fields. Record who requested the change, when and why. This distinguishes legitimate customer changes from revisions made because delivery would otherwise fail.
Tracking both original and latest valid commitments can help. The first shows initial-plan reliability; the second shows execution after revision. Transparency lets Sales, Customer Service and Logistics discuss the same facts rather than competing calendars. Keep criteria consistent for everyone.
Be careful with denominator exclusions
Cancelled, rejected, in-transit orders and those without proof need explicit treatment. Exclusions require criteria and traceability. Removing difficult cases can improve the metric without improving service; marking failure before the delivery window closes also distorts it. Eligibility must reflect stage and measurement period.
Maintain three groups: completed with evidence, awaiting evidence and out of scope under a documented rule. Show group sizes and mark incomplete results as provisional. Final results need a cut-off date and reconciliation of changes since the initial calculation, especially when used in partner discussions.
Check data quality before holding teams accountable
Duplicate orders, time-zone differences, incompatible quantity units and empty fields can alter OTIF. A duplicate can inflate the denominator or repeat an event; cases and individual units may appear inconsistent without conversion. Check that records represent the order and receipt correctly before assigning responsibility. Master-data errors waste time and erode trust.
Use simple checks: unique ID, positive quantity where applicable, a process-consistent date and a link between order and proof. Separate data errors from operational failures: the former need source correction and prevention; the latter need process investigation. Preserve that distinction when both occur. Missing reliable information does not automatically mean success or failure.
Break down failures to identify action
After calculating the total, split failures into late, incomplete and both. Drill down by region, customer, route, product or stage as needed. Detail should guide action. Dozens of charts without a clear question can hide the main issue. Start with the largest concentration and check the individual cases behind it.
Give cause categories short definitions. Stock shortages, picking delays, vehicle unavailability, scheduling and rejection are not equivalent. Do not blame transport merely because a failure appears at delivery. Late dispatch may make the required lead time impossible. Reconstruct the event sequence and identify when fulfilling the commitment became unlikely.
Compare partners in context
Carriers may serve very different profiles: regular predictable routes versus urgent, complex destinations. Overall percentages matter but do not explain the difference alone. Compare similar segments and report order counts. With small samples, one reclassified delivery can substantially change the result.
Context is not an excuse for every deviation; it makes evaluation actionable. Assess performance with coverage, lead time, cost and cargo profile. Identify changes needed for complex flows. Negotiations should produce observable commitments such as revised windows, capacity availability or better tracking, not just a generic demand for a higher percentage.
Connect measurement to action plans
Performance meetings should end with actions tied to identified causes, with owners, deadlines and completion evidence. Repeated loading delays may call for revised dock schedules and waiting-time monitoring; missing delivery information may require better event confirmation. Each action needs a verifiable connection to the deviation it targets.
Track effects in the same order segment. A better aggregate while the problem segment remains unchanged may reflect a mix shift, not effective action. Watch trade-offs: urgent trips can recover time at higher cost; excessive consolidation can cut freight costs while harming service. Obtain proper approval for extra resources or changed customer agreements.
How to use this site's simulator
The freight simulator uses target and historical OTIF in a risk rule. It does not calculate OTIF from orders or replace this guide's measurement process. Enter percentages from validated records. It combines them with delay, capacity and other inputs to support an educational discussion of alternatives.
Review evidence and plan assumptions before reading the recommendation. High risk indicates review priority, not confirmed alternative capacity. Costs depend on your inputs. Saved files preserve parameters for later review, not an approved decision. Keep the tool as discussion support and retain operational judgement.
A short dashboard definition
A dashboard definition could read: percentage of eligible orders received in full within the agreed window, based on receipt evidence and the period-closing rule. Document units, tolerance, exclusions and source in an accessible place. Match the wording to your process. If the team cannot explain an individual case, the rule or evidence may still be inadequate.
Start with a sample and manually check cases before automating the full calculation. Validate normal and exceptional situations, including partial and rescheduled deliveries. Then maintain reconciliation and cause analysis. These are educational examples. OTIF quality combines definition, evidence and action: consistent measurement builds a basis for more reliable service.