OTIF stands for On Time In Full. It measures the share of orders delivered both within the agreed window and complete, with every line and unit present. Because it requires both conditions on the same order, OTIF always reads lower than on-time or in-full measured separately.
The word doing the work is both. An order arriving on time with two lines missing fails. An order arriving complete two days late fails. Only orders satisfying both conditions count, which is why OTIF is consistently lower than the individual measures people quote, and why it is the number large retailers use when they want an honest picture of supplier performance.
It is a demanding measure by design. That is the point — it reflects what the customer experiences rather than what any single department achieved.
What OTIF measures
On time — the delivery arrived within the agreed window. The definition matters enormously and is where most disputes originate. A window might be a specific date, a date range, a booked delivery slot, or a period of hours. Some customers count early delivery as a failure too, because an early arrival occupies dock capacity and warehouse space they had allocated to something else.
In full — every line and every unit ordered was delivered. A shipment missing one item from a twenty-line order is not in full. Partial delivery does not earn partial credit.
Both, on the same order — this is what separates OTIF from its components. You can run 95% on-time and 95% in-full and deliver considerably less than 95% OTIF, because the failures do not necessarily overlap.
Some definitions add further conditions — undamaged, correctly documented, correct paperwork — at which point the measure is usually called perfect order rate rather than OTIF.
Why “both” matters
The arithmetic is worth seeing, because it explains why OTIF surprises people the first time it is measured.
If 95% of orders arrive on time and 95% arrive in full, and the two failure types are unrelated, then the share satisfying both is roughly 0.95 × 0.95 = 90%.
So two respectable-looking measures combine into one that looks considerably worse. Nothing has got worse; you are simply measuring what the customer experiences rather than what each function achieved separately.
In practice the failures often are related — a shortage causes both a late dispatch and an incomplete one — which makes the real figure somewhat better than the pure multiplication suggests. It remains lower than either component, always.
This is precisely why customers favour it. On-time and in-full measured separately allow both sides of a business to report success whilst the customer receives neither.
The OTIF formula
OTIF = (orders delivered on time and in full ÷ total orders delivered) × 100
Simple arithmetic, and three definitional decisions do all the real work:
What counts as on time? The agreed date, the requested date, or the date you confirmed? These differ, and choosing the most favourable one produces a number that flatters you and convinces nobody.
Measured at dispatch or delivery? Dispatch is within your control and easier to measure. Delivery is what the customer experiences. Customers measuring you will use delivery, and if you measure at dispatch, your figure and theirs will never agree.
Counted by order, line, or unit? Order-level is strictest — one short line fails the whole order. Line-level is more forgiving and more diagnostic. Whichever you choose, use it consistently, because switching between them is the easiest way to appear to improve without changing anything.
A worked example
You deliver 500 orders in a month:
- 460 arrived within the agreed window — 92% on time
- 475 arrived complete — 95% in full
- 445 achieved both
OTIF = (445 ÷ 500) × 100 = 89%
Note that 89% is below both components. Of the 55 failures, 40 were late, 25 were incomplete, and 10 were both — so 55 orders failed for 65 reasons.
That breakdown is more useful than the headline. Forty late deliveries and twenty-five incomplete ones point at different problems with different owners: lateness usually sits with dispatch, carriers or capacity, whilst incompleteness usually sits with stock availability and purchasing. An OTIF figure without the split tells you that you have a problem and nothing about where it is.
OTIF vs OTD vs fill rate
Three measures that overlap and are routinely confused, usually to somebody’s advantage.
OTD — on-time delivery. Did it arrive within the agreed window? Timing only. An order arriving on the right day with half the lines missing scores a success.
Fill rate. Was it complete? Completeness only, and it can be counted by order, by line or by unit — the three give noticeably different figures on the same data. An order arriving complete three weeks late scores a success. There is more detail in the fill rate guide.
OTIF. Both conditions, on the same order. Which is why it always reads lower than either of the other two, and why it is the one customers use.
The reason this matters is that the three get quoted interchangeably in supplier conversations. A provider reporting “97% on time” and a customer measuring 88% OTIF are both telling the truth about different things. If you are being measured, establish which of the three the figure refers to before discussing whether it is good.
A fourth term you will meet is perfect order rate, which adds further conditions — undamaged, correctly documented, correctly invoiced — on top of on time and in full. It is the strictest of the family and reads lower again.
The practical arrangement is to track all three: on-time and fill rate separately for diagnosis, because they have different causes and different owners, and OTIF as the combined figure that reflects what the customer actually received.
What causes OTIF failures
Incomplete deliveries almost always trace back to stock:
- Reorder points set from quoted rather than measured supplier lead times
- Safety stock too thin on lines where demand varies
- Forecast error, particularly around promotions
- Supplier shortfalls passed straight through to your customer
- Inventory inaccuracy — the order was accepted against stock that was not there
Late deliveries usually trace to timing and capacity:
- Orders released too close to carrier cut-offs
- Picking capacity insufficient for the day’s volume
- Pick face empty whilst reserve stock exists, so replenishment stalls the pick
- Carrier collection missed or failed
- Documentation incomplete, holding a shipment at dispatch or at a border
- Customer booking slots missed, where delivery requires an appointment
Both at once is typically a shortage discovered late: the order is held whilst somebody tries to source the missing line, then ships incomplete anyway.
The diagnostic value comes from coding every failure to one of these at the time it happens. Operations that record only the OTIF percentage know they are failing and not why.
How to improve OTIF
Fix availability first. Most OTIF failure in most businesses is in-full failure, and most in-full failure is a stock problem. Reorder points from measured lead times, safety stock sized to actual demand variability, and inventory accuracy good enough that you are not accepting orders against phantom stock.
Measure supplier OTIF too. You cannot deliver in full what you did not receive. Applying the same measure upstream frequently reveals that your customer-facing problem originates with two or three suppliers.
Release orders earlier relative to cut-offs. A large share of lateness is orders picked correctly but too late for the collection. Wave planning around carrier times removes it.
Get replenishment right. An empty pick face with full pallets in reserve above it is a self-inflicted delay, and it is invisible in most reporting.
Agree the definition with the customer. Then measure it their way. A supplier reporting 96% against a customer measuring 88% has a relationship problem on top of an operational one, and the gap is almost always definitional.
Split the reporting. Track on-time and in-full separately alongside the combined figure, because they have different causes and different owners.
Watch the trade-off. OTIF improves with more stock, which costs money. The right level is a commercial decision about what service is worth, not an operational target to maximise.
OTIF in retail supply: penalties and disputes
For anyone supplying large retailers, OTIF stops being a management metric and becomes a contractual one with money attached.
Most major grocery and general merchandise retailers operate supplier performance regimes that charge for OTIF failures. The charge is typically a percentage of the value of the affected order, sometimes with escalating rates for repeated failure, and it is deducted rather than invoiced — which means it appears as a shortfall on a remittance rather than as a bill you can query easily.
The mechanics worth understanding before signing:
Delivery windows are narrow and specific. Often a booked slot of an hour or two at a named depot. Arriving outside it can fail the delivery even when the goods are correct and the day is right.
Early is a failure too. Retail depots plan labour and dock capacity against booked slots. A vehicle arriving a day early is as disruptive as one arriving a day late, and is generally scored the same way.
Measurement is at their goods-in, not your dispatch. Transit time, carrier performance and depot queuing are all inside your measured performance and outside your direct control.
Substitutions usually count as shortfalls. Sending a similar product because the ordered one was unavailable is not delivering in full.
The data is theirs. Disputes are resolved against the retailer’s records, which is why keeping your own proof of delivery, timestamps and booking confirmations matters.
Two practical defences. Reconcile every charge rather than accepting deductions — error rates in these systems are not negligible, and unchallenged charges become the baseline. And book realistically: accepting a slot you are unlikely to meet in order to win the order converts a commercial problem into an operational penalty.
Setting a target that means something
An OTIF target that nobody can influence is decoration. Making it useful takes a few decisions.
Set it from the cost of failure, not from a benchmark. What does a failed delivery actually cost — penalty, expedited freight, lost sales, the customer’s patience? That figure tells you what level of service is worth paying for, and it differs by customer and by product.
Segment it. One target across every customer is wrong in both directions. Large retail accounts with penalty regimes and short windows need a high target. A wholesale customer ordering monthly with a flexible date does not, and holding stock to serve them at the same level is money spent for nothing.
Attach owners to the components. In-full performance generally belongs with purchasing and inventory planning. On-time performance generally belongs with warehouse operations and transport. A combined figure with no owner gets discussed rather than fixed.
Review the failures, not the percentage. A monthly review that looks at the number tells you whether it moved. A review that looks at the individual failed orders and their causes tells you what to change.
Accept that 100% is not the target. Perfect service requires stock levels nobody can justify. The right figure is the one where the cost of the next percentage point exceeds what it saves, and that point arrives earlier than most targets assume.
Measuring it when your systems do not
Plenty of operations are asked for an OTIF figure and discover their systems do not produce one. It is usually assemblable from data you already hold.
What you need per order: the date promised, the date actually delivered, the quantity ordered per line, and the quantity delivered per line. Most order management and accounting systems hold all four, even if none of them reports the combination.
On time is then a comparison of two dates. The trap is which promised date you use — the one the customer requested, the one you confirmed, or the one on the dispatch note. Pick the one the customer will measure against, which is almost always the confirmed delivery date.
In full is a comparison of ordered against delivered quantity at line level, rolled up to the order. An order is in full only if every line is complete.
The combination is a simple flag: both conditions true, or not.
Two practical points. Start with a sample — a hundred recent orders scored by hand will tell you roughly where you stand and what the split between late and incomplete looks like, which is enough to decide whether the effort of automating it is justified. And exclude what you should not be measured on — orders the customer amended after placement, or where they requested a later date — but define those exclusions explicitly and sparingly, because a measure with generous exclusions stops meaning anything and everybody knows it.
If you supply a customer who measures you, ask them for their data. It is usually available, it is what disputes are settled against, and comparing it with your own is the fastest way to find a definitional gap.
Where to go deeper
- What is fill rate — the in-full component measured on its own.
- What is a stockout — the most common cause of in-full failure.
- What is safety stock — the buffer that prevents it.
- What is a backorder — what an incomplete delivery becomes.
- Warehouse KPIs — where OTIF sits among the other measures.
- What is order fulfilment — the process being measured.
The one-line takeaway
OTIF measures the share of orders arriving both on time and complete, and because it requires both it always reads lower than either component alone — split the failures into late and incomplete before trying to improve it, because they have different causes and different owners.
Frequently asked questions
What does OTIF stand for? On Time In Full. It measures the proportion of orders delivered within the agreed window with every line and unit complete.
What is the OTIF formula? Orders delivered on time and in full, divided by total orders delivered, multiplied by 100. The arithmetic is simple; the definitions of “on time” and “in full” are where the difficulty lies.
What is a good OTIF score? It depends on sector and on how the measure is defined. Large retailers commonly set supplier requirements in the mid-nineties with financial penalties below that. More useful than any benchmark is your own trend and the split between late and incomplete failures.
What is the difference between OTIF and fill rate? Fill rate measures completeness only — the share of orders or lines supplied in full. OTIF requires completeness and timeliness together, so it is always the lower of the two.
Why is my OTIF lower than my on-time and in-full figures? Because it requires both conditions on the same order. Two measures at 95% each combine to roughly 90% when their failures are unrelated. The combined figure is always lower than either component.
Should OTIF be measured at dispatch or delivery? Delivery, because that is what the customer experiences and what they will measure. Dispatch is easier to control and to track, and measuring it instead guarantees your figure and your customer’s will disagree.
How do I improve OTIF? Start with availability, since in-full failures usually dominate — reorder points from measured supplier lead times, adequate safety stock, and accurate inventory records. Then address timing: release orders earlier relative to carrier cut-offs and fix pick face replenishment.
Warehouse Wise is a free, independent library of guides on inventory management, warehousing and fulfilment, written for the people actually doing the work — new warehouse and inventory staff, small and growing retail and ecommerce businesses, and anyone who holds stock and wants to understand it properly. Every guide explains the term, shows a worked example with real numbers where it helps, and stops. We are independent: we do not sell software or equipment, we take no commission on referrals, and our guides are not ranked by who pays us.