OTIF: what it is, how to calculate it and how to improve it in distribution

Almost every distributor says it delivers well. Very few can say with what number. OTIF is the metric that turns that feeling into a figure, and the first uncomfortable figure is usually that the real service level sits 15 to 25 points below what the team believed.

In this article we look at what OTIF measures exactly, how to calculate it without cheating, how it differs from fill rate, where it breaks in a distribution operation, and what you can do with the data already in your ERP to lift it.

1. What OTIF is and why it matters

1.1 Definition

OTIF stands for On Time, In Full. It is the percentage of orders that reach the customer on the committed date and with everything requested — no shortages, no substitutions, no partial deliveries.

The key is the conjunction. An order that arrives on time but at 80% is not OTIF. A complete order that arrives three days late is not either. OTIF punishes any deviation, and that is why it is the most honest service metric there is: it is the only one that resembles what the customer actually experiences.

1.2 The formula

The base calculation is simple:

OTIF = (Orders delivered on time and in full / Total orders delivered) × 100

If in one month you shipped 1,200 orders and 948 arrived on time and complete, your OTIF is 79%.

The disaggregated version, which is the useful one for managing, separates the two components:

• On Time (OT) = Orders delivered on date / Total orders
• In Full (IF) = Orders delivered complete / Total orders
• OTIF ≈ OT × IF

1.3 The multiplicative effect almost nobody sees

Here is the finding that makes this indicator worth the trouble. Suppose your logistics team reports 95% on-date deliveries and your warehouse team reports 92% complete orders. Both numbers look good and both teams are relaxed.

The combined OTIF is 95% × 92% = 87.4%. One in every eight customers had a problem. And neither of the two managers sees it, because each is looking at their own half.

This is why OTIF is a leadership metric and not a departmental one: it exists precisely so nobody can hide behind their own indicator.

1.4 OTIF, fill rate and service level are not the same thing

The three terms get used as synonyms and they are not:

• Fill rate measures units or lines shipped over requested. It is more forgiving: an order of 10 lines with 9 delivered gives 90% fill rate and 0% OTIF.
• Service level usually refers to the probability of not stocking out on an SKU. It is an inventory metric, not a delivery one.
• OTIF measures complete, punctual orders. It is binary per order: you comply or you do not.

Measuring fill rate and calling it service level is the most common mistake. It inflates the number and hides exactly the orders that generate complaints.

2. How to measure it without cheating at solitaire

2.1 Define "on time" before measuring anything

The first decision determines everything else: on time relative to which date?

• The date requested by the customer: the most demanding and the most honest.
• The date you confirmed: the most widely used. The risk is obvious — if you always confirm late, your OTIF is perfect and your customer is angry anyway.
• The rescheduled date: the most frequent trick. Every reschedule resets the clock and the indicator stays immaculate.

The healthy practice is to measure against the requested date and track separately the percentage of orders where you could not confirm the date asked for. That second number is as informative as the first.

2.2 Define "in full"

Just as important: complete at the order, line or unit level? Measuring per order is what reflects the customer's experience. The other two are for diagnosis, not for reporting.

You also have to decide what happens with accepted substitutions, agreed partial deliveries and shortages the customer themselves asked to cancel. The practical rule: if the customer explicitly approved it before dispatch, it does not count as a failure. If they found out when the truck arrived, it counts.

2.3 The data missing from almost every ERP

The real obstacle to measuring OTIF is not the formula, it is that the promise date is not recorded anywhere. It is in an email, in a WhatsApp message or in the rep's head. Without that field, the numerator does not exist.

Before building any dashboard you have to solve three captures: the date committed at the moment of the order, the actual delivery date and the quantity actually received. If the carrier reports the last one rather than the customer, the number will be systematically inflated.

2.4 The most frequent measurement mistakes

• Measuring over invoiced orders instead of requested orders, which makes the orders you could never assemble disappear from the calculation.
• Excluding urgent or exceptional orders, which tend to be 20% of the volume and 60% of the failures.
• Counting as delivered the moment it left the warehouse rather than the moment it arrived.
• Averaging OTIF across all customers, hiding the fact that your ten largest accounts are 20 points below.

3. Where OTIF breaks in a distributor

3.1 Stockouts

This is cause number one of In Full. And it is almost always concentrated: between 5% and 10% of SKUs explain the majority of unserved lines. A turnover and coverage analysis on those specific SKUs, crossed with the stockout history, is the obligatory starting point. In purchase planning we develop how this gets anticipated.

3.2 Picking and assembly

Assembly errors, shortages detected only at loading, orders left unprioritized in the warehouse. They get detected when there is live visibility of picking, not when you count what came back.

3.3 Transport and carriers

When the last mile is outsourced, OTIF depends on a third party that does not report at your granularity. Without KPIs by carrier, route and driver, the problem becomes invisible and the conversation with the provider becomes anecdotal.

3.4 The customer's delivery window

Especially in modern trade: you arrive on time but outside the time window, or without the documentation the distribution centre demands, and they reject the unload. Formally you delivered; in the customer's paperwork, you failed.

3.5 Attributing the root cause

An OTIF with no assigned cause is a number that depresses but does not drive action. Every failure should be classified into one of five categories: stock, assembly, documentation, transport or customer. With three months of that classification you have a Pareto that says exactly where to invest.

4. How to lift OTIF

4.1 Start with the Pareto, not the average

Overall OTIF does not move. What moves is the segments. Break the indicator out by customer, by product family, by area and by carrier, and you will find that two or three combinations concentrate the majority of the failures.

4.2 Buffers on the SKUs that break deliveries

You do not need to raise inventory overall. You need to raise it on the SKUs that appear in the unserved lines of your best customers, and lower it on the ones tying up capital. It is an exercise in recomposition, not in increase, and in reducing excess inventory you have the other side of the equation.

4.3 Promise well before delivering fast

A good share of On Time failures are not execution problems but promise problems. If the rep promises 48 hours because that is what the customer wants to hear, and the operation consistently delivers in 72, the problem is not in the warehouse. A date calculated on real availability, picking capacity and route frequency fixes more OTIF than any logistics investment.

4.4 What target to set

In wholesale distribution, an OTIF measured rigorously usually starts between 70% and 85%. Modern trade demands 95% or more and penalizes financially below that. The reasonable target for the first year is to rise 10 to 15 points by attacking the two or three causes in the Pareto, not to reach 98%.

One detail worth anticipating: when you start measuring properly, OTIF goes down. It is not that the operation got worse, it is that it stopped being made up. That first number is the most valuable one of all.

Frequently asked questions

What is a good OTIF? It depends on the channel. Between 95% and 98% in modern trade, between 85% and 95% in well-managed traditional distribution. Below 85% there are lost sales and lost customers you are not yet seeing.

Is OTIF measured on orders or on lines? On orders, because that is what the customer experiences. Lines are useful for diagnosing where it fails.

How often is it measured? Weekly to manage, monthly to report. Daily only if you have the delivery data automated.

Can OTIF be measured without changing the ERP? Yes. The order, dispatch and invoice data is already there. What is usually missing is the promise date and the receipt confirmation, and both can be captured without touching the core of the system.

How nBlock measures and improves OTIF

nBlock's Logistics block runs on the data already in your ERP and builds end-to-end delivery control:

• OTIF by order, customer and carrier, with the failure cause classified.
• Live picking in the warehouse, to see the state of assembly before it is too late.
• KPIs by carrier, route and driver, to know who delivers and who does not.
• Alerts on the SKUs and customers where the failures concentrate.
• It is implemented on top of your current system, replacing nothing.

---

Want to know your real OTIF? Book a demo and we will calculate it on your own data.