OTIF (On Time In Full) is the percentage of orders that reach the customer both at the agreed time and in the full quantity. The difficulty is in the 'and': an order that arrived on time but two lines short does not count. That is why OTIF is usually lower than the separate 'on-time' and 'fill rate' figures, and that gap is often the first thing that makes teams doubt the definitions.
The first question is the counting level: order, order line or unit? Count OTIF at order level and one incomplete line fails the whole order; count at line level and the number looks better but reflects customer experience less. Neither is wrong, but it must be defined the same way with the customer and internally, or two people will see two numbers for one reality.
The second question is the 'agreed time'. Is the baseline the customer's requested date or the date your system confirmed? If it is your confirmed date, conservative promises can make OTIF look high while the customer is unhappy. The clearest approach is to measure both: OTIF against the customer request and OTIF against the confirmed promise. The gap between them is itself an important diagnosis of real responsiveness.
The third is the time window. Does 'on time' mean exactly that day? Is a day early acceptable? For a customer with limited storage, early delivery is also a problem. Define the window from the customer's real need, not from what is easiest to measure. In many industries different windows for different customers are more sensible than one rule.
Average OTIF can mislead. A 90 percent overall figure may be 99 percent for small customers and 70 percent for a key account. Always slice OTIF: by customer, product group, origin warehouse and failure cause. Only these cuts reveal whether the problem is inventory, transport capacity, order entry or over-promising.
OTIF without a failure cause is just a score. For every failed order, record the main cause from a short fixed list: stockout, supplier delay, warehouse error, transport delay, customer order change and so on. After a few weeks a Pareto chart of causes usually shows that one or two causes make up a large share of failures, and corrective effort can be focused there.
A common mistake is setting a 100 percent target or one target for all customers. Every percentage point of service level costs inventory and capacity, and the relationship is non-linear. Set targets by customer and product importance and the real cost of failure, and look at service level together with safety stock and transport capacity decisions in the S&OP meeting.
A practical starting point: this month, compute OTIF for the last three months for your top 10 customers at both order and line level, against both dates, and categorize the failure causes. This few-day exercise usually exposes hidden definitions, disagreements and the first improvement levers.