OTIF improvement: How to find where your deliveries break down

Customer service has logged three complaints in a week.

Sales wants more inventory.

Purchasing blames the supplier, and the supplier says you ordered too late.

Everyone has an explanation, and OTIF is still stuck at 86%. Everyone knows the number. Nobody knows why.

If you want to improve OTIF, the first step is neither raising inventory nor putting pressure on the supplier. It is finding out where orders actually fail.

This article gives you a simple way to split the number, so you can see whether the problem sits with the supplier, in your own planning or in the warehouse. Then you can decide what to do about it.

What is OTIF and how do you calculate it?

OTIF stands for On Time In Full. An order only counts if it arrives at the agreed time and in the full quantity. If it arrives on time but is missing one order line, it is not OTIF. If it is complete but a day late, it is not OTIF either.

OTIF (%) = orders delivered on time and in full ÷ total orders × 100

The formula is simple. The definitions behind it are not, and they need to be settled before you try to improve OTIF as a KPI. Which date do you measure against: the date the customer requested or the date you confirmed? Does a delivery one day early count as a miss? And do you measure the whole order or each order line?

In manufacturing, where one missing component stops the line, measuring at line level can make sense. A wholesaler may choose order level because that is how the customer experiences the delivery. Both choices can be defended. Pick one definition, write it down and use the same one every month. Otherwise you end up debating the number instead of the cause. You can read more about the term itself in the OTIF glossary entry.

Why is OTIF so hard to improve?

Because one number covers two different failures, and they have different owners.

A late order can be caused by picking capacity, transport or a customer promise that was never realistic. An incomplete order can be caused by the item not being there when the order was packed. And the missing item can in turn come down to a supplier, a purchase order placed too late, or a stock record that said something different from the shelf.

If you only look at the total, you know something is wrong. You don't know who needs to act.

How to improve OTIF: split the number before you act

Take a constructed example. A wholesaler has 1,200 customer orders in a quarter. 1,032 of them are delivered on time and in full. OTIF is 86%. The 168 orders that miss break down like this:

• 41 were late but complete
• 97 were on time but incomplete
• 30 were both late and incomplete

Split by the two requirements, the picture changes. 94% of orders arrived on time: (1,200 − 41 − 30) ÷ 1,200. Only 89% were complete: (1,200 − 97 − 30) ÷ 1,200. In this example, "in full" is what pulls OTIF down. 127 of the 168 failed orders were missing an item.

The next question is why the item was missing. If you record the primary cause of every miss, the 127 orders might break down like this:

• 74 were missing an item because an inbound delivery was late
• 31 were missing an item because the purchase order was placed later than the supplier's lead time allowed
• 22 were missing an item that the system showed in stock, but that wasn't on the shelf

Dig into the 74 and you find that 52 of them involve items from just two suppliers.

You now have three separate tasks instead of one. Two suppliers need a specific conversation. Planning needs to find out why purchase orders are placed too late to account for the supplier's lead time. And the warehouse needs to fix stock accuracy on the items where the numbers don't match.

Raising safety stock across the board would try to solve all three problems with inventory. It may lift OTIF. But you risk tying up more working capital without fixing the 31 late purchase orders and 22 stock errors that were also part of the problem.

The example also shows how much the two suppliers matter. Had those 52 orders otherwise gone through, OTIF for the quarter would have been just over 90%: (1,032 + 52) ÷ 1,200. The numbers are constructed to show the method. The split in your own data may look very different, which is exactly why you need to break it down.

When is low OTIF the supplier's fault?

When OTIF drops, the supplier is the easiest target. Sometimes rightly so. If a supplier delivers late, partially or with defects that only show up at goods receipt, it hits your own customer orders.

But supplier OTIF depends on what you measure it against. Measure against the date you requested, and the supplier can look bad on orders you placed too late. Measure against the date the supplier confirmed, and the supplier can look fine even though the confirmation came a week after your need date.

Both numbers are useful, for different things. The confirmed date shows whether the supplier keeps its promises. The requested date shows whether the supplier can meet your needs. If there is a big gap between the two, find out why. Did you order too late? Is the lead time in the system wrong? Or can the supplier simply not deliver when you need it?

Get that distinction in place before you negotiate. Otherwise you risk asking a supplier to fix mistakes you made yourself.

How do OTIF and supplier management fit together?

A supplier OTIF of 91% can look acceptable. It just doesn't tell you what the remaining 9% hits. If those misses are items in your best-selling products, or parts in a bill of materials where one shortage stops production, 91% is a problem. If they are items where the customer waits three weeks anyway, it may not be.

So use supplier OTIF together with the consequence. Which suppliers cause the most backorders for your customers? Which ones force you to hold extra safety stock? And which ones vary so much in lead time that you never know when the goods will arrive? The last one is worth measuring on its own.

If a supplier swings between 8 and 14 days, you are building a buffer against the supplier, not just against demand.

Once you have the numbers per supplier, supplier management becomes more concrete.

A conversation about which order lines arrived late this quarter and what that meant for your customers is a different conversation from "your delivery performance needs to improve". To assess your suppliers as a whole, read the guide to supplier portfolio analysis.

Why doesn't more inventory always fix it?

Inventory is the fastest way to protect OTIF. If a supplier is unreliable, you put a little extra on the shelf. It works, and sometimes it is the right call. One example is a critical item while you work on the supplier relationship.

But inventory only helps the orders where a missing item was the problem. In the example above, extra inventory would do nothing for the 41 orders that were complete but late. And if you add buffer to every item, you tie up working capital in items that were already delivered without issues.

So don't only ask how much inventory you need to hold OTIF. Ask why you need that much inventory to deliver on time and in full. The answer may point to a supplier, a parameter or a process that is cheaper to fix than to cover with inventory.

It also means not every item needs the same OTIF target. An item that stops a customer's production may deserve high delivery reliability. A slow-moving item with a low margin and a long replenishment time might be fine with a longer customer promise, as long as the customer knows. For how to find that balance, see the guide to service level vs inventory cost.

Which KPIs should you track alongside OTIF?

OTIF can go up for good and bad reasons. If it rises while inventory grows, you may simply have bought it. Track it alongside a few other numbers:

• Fill rate, which shows how much of the ordered quantity you deliver, even when the order is incomplete
• supplier lead times and how much they vary
• backorders: how many and how long they stay open
• inventory turnover and working capital

The article on inventory KPIs shows how several of them are calculated and read together.

From one OTIF number to three concrete tasks

The breakdown above is easy to do for one quarter and a handful of suppliers. The hard part is keeping it up to date when you have hundreds of suppliers, thousands of order lines and lead times that change from order to order.

Inact's supplier performance solution pulls purchase orders, goods receipts and invoices from your ERP system and calculates OTIF, lead times and quality per supplier. You can see where deliveries break down and talk to the supplier from a scorecard you can both see.