# Fill rate, OTIF and backorder rate

> Line, unit and order fill rates, what on-time in-full measures, why trading partners define it differently and how fill rate relates to service level.

Source: https://docs.lumina-erp.com/distribution/fill-rate-and-otif/

**In short.** Fill rate is the share of demand you shipped from stock, and it has three common forms (line, unit, order) that give very different numbers for the same month. OTIF adds the delivery date, and every trading partner defines both halves a little differently, so always write down the definition next to the number.

Fill rate is the number customers feel most directly: did the thing they ordered leave on the first truck? It sounds like one metric, but it is at least three, and they can differ by 10 points in the same month. On-time in-full (OTIF) and backorder rate add delivery and follow-up to the picture.

## The measures at a glance

| Measure | Counts a success when | Sensitive to | Typical use |
|---|---|---|---|
| Line fill rate | An order line ships complete on its first shipment | Breadth of stock across many items | Warehouse and purchasing scorecards |
| Unit fill rate | Each unit ordered ships on the first shipment | Big-quantity lines | Volume and revenue views |
| Order fill rate | Every line on the order ships complete on its first shipment | Order size (more lines, more chances to miss) | Customer experience |
| OTIF | The order (or line) arrives complete and inside the agreed window | Stock and transportation together | Retailer and key-account scorecards |
| Backorder rate | A line (or unit) could not ship and was kept open to ship later | Stock position and backorder policy | Buyer workload, customer service |

Every one of these is a ratio, so the definition of the numerator and the denominator decides the answer. Two teams can report "fill rate" from the same data and both be right.

## Line fill rate

**Line fill rate** measures the share of order lines that shipped complete on the first shipment. A line for 10 units that ships 9 counts as a miss.

**Line fill rate:** `Lines shipped complete on first shipment ÷ lines requested`

It treats every line the same, whether it is one box of screws or a pallet of pipe. That makes it a good measure of how broadly you are in stock, and the one we reach for first in a distribution business with a long tail of items.

## Unit fill rate

Unit fill rate counts units instead of lines. The 9-of-10 line above contributes 9 successes and 1 miss.

**Unit fill rate:** `Units shipped on first shipment ÷ units requested`

Unit fill is usually the highest of the three, because a short line still ships most of its units. It is dominated by high-quantity lines, so a few large orders filled well can hide many small misses. Some teams weight by dollars instead of units to get a value fill rate. The same caution applies.

## Order fill rate

Order fill rate is the strictest measure. An order counts only if every line on it shipped complete on the first shipment.

**Order fill rate:** `Orders shipped complete on first shipment ÷ orders`

It falls as order size grows. If each line independently has a 95% chance of shipping complete, a 10-line order ships complete only about 60% of the time (0.95 to the 10th power). Order fill is the closest to what a customer experiences, and the hardest number to move.

## Worked example

With invented round numbers, suppose that in one month a branch received 200 orders containing 1,000 lines and 20,000 units. On the first shipment, 950 lines shipped complete, 19,400 units shipped and 170 orders shipped with every line complete.

| Measure | Calculation | Result |
|---|---|---|
| Line fill rate | 950 ÷ 1,000 | 95.0% |
| Unit fill rate | 19,400 ÷ 20,000 | 97.0% |
| Order fill rate | 170 ÷ 200 | 85.0% |

The same month and warehouse give three answers spread over 12 points. That spread is normal, so a fill rate number needs its definition next to it.

_Interactive calculator available on the web page._

## On time and in full (OTIF)

OTIF counts an order (or line) as a success only when it is both complete and delivered inside the agreed window. It combines a stock question (did we have it) with an execution question (did it get there when promised).

**OTIF:** `Orders delivered complete and inside the agreed window ÷ orders`

In the same invented example, 160 of the 170 orders that shipped complete arrived inside the delivery window. OTIF is 160 ÷ 200, or 80.0%. OTIF can never be higher than the in-full rate alone, and it is usually lower.

ASCM's SCOR Digital Standard places a broader measure, perfect order fulfillment, at the top of its reliability attribute. In paraphrase, a perfect order is one delivered in full, on the date committed, with accurate documentation and in good condition. OTIF is the first two of those four tests.

### Why OTIF definitions differ between trading partners

OTIF is not a single standard. Large retailers and key accounts publish their own supplier requirements, and the details change the number a great deal. Settle these questions with each partner.

| Question | Common variations |
|---|---|
| On time against which date? | Requested date, first promised date or the latest agreed date |
| Measured where? | Shipped from your dock, or arrived at their dock |
| How wide is the window? | Exact day, a day either side or an appointment slot |
| Is early a miss? | Many retailers count early arrival as not on time |
| In full at what level? | Order, line or units (case count), sometimes with a tolerance |
| Who owns carrier delay? | Depends on whether you or the customer arranged freight |

We keep one internal OTIF definition for our own management and calculate each partner's version separately to match their scorecard. Trying to force one number to serve both usually means neither is trusted.

:::note
When a customer shares their OTIF scorecard, reconcile a sample of their misses to your own shipment records before arguing about the total. Most disputes come down to a different date or a different unit of measure.
:::

## Backorder rate

Backorder rate measures the share of lines (or units) that could not ship and were kept open to ship later.

**Backorder rate (lines):** `Lines placed on backorder ÷ lines requested`

In the example, if every one of the 50 short lines went to backorder, the line backorder rate is 50 ÷ 1,000, or 5.0%, the mirror image of line fill. The two do not always add to 100%, because some short lines are cancelled or substituted instead of backordered. A rising backorder rate with a steady fill rate often means customer service has stopped offering alternatives.

## First shipment versus eventual fill

The formulas above all use the **first shipment rule**: a line counts as filled only if it shipped complete the first time. The alternative, eventual fill, counts a line as filled once all of it has shipped, however many shipments that took.

Eventual fill almost always looks better, and it measures something real (you did eventually deliver). But it hides the cost of the second shipment, the extra freight and the customer who waited. We report fill on the first-shipment rule and track time to fill backorders as a separate measure.

## What distorts the numbers

Order handling habits change fill rates without any change in stock. Decide how each case counts before you build the report.

| Situation | What goes wrong | A reasonable rule |
|---|---|---|
| Substitutions | The line ships a different item, so it may count as a miss, or as a hit that hides the stockout | Count as filled for the customer measure, but log it as a stockout of the original item for purchasing |
| Customer-requested partials | The customer asked for part now and the rest later, so the short first shipment looks like a miss | Measure against what the customer asked to ship on that date, not the total ordered |
| Drop-ship lines | The supplier ships, so your stock was never involved | Exclude from warehouse fill, report supplier fill separately |
| Future-dated lines | Lines not due yet appear open and unshipped | Count lines in the period they were due to ship, not the period they were entered |
| Cancelled lines | A line cancelled because you had no stock disappears | Keep stock-driven cancellations in the denominator as misses |
| Unit of measure changes | Ordered in cases, shipped in eaches, so unit counts disagree | Convert to one base unit before calculating unit fill |

Your ERP holds the order, shipment and date fields these rules need. The work is in agreeing the rules and applying them the same way every month.

## Fill rate versus cycle service level

Fill rate and cycle service level are both called "service level," but they measure different things. Mixing them up is one of the most common errors in inventory planning.

| | Cycle service level (CSL) | Fill rate |
|---|---|---|
| Question it answers | In what share of replenishment cycles do we avoid running out at all? | What share of demand do we ship from stock? |
| A cycle with one unit short | Counts as a complete failure | Counts as a small miss |
| Where it is used | Setting safety stock with a z value | Reporting to customers and managers |
| Depends on order quantity? | No | Yes, larger orders mean fewer cycles and fewer chances to run short |

Safety stock formulas are built on cycle service level: choosing 95% means you expect to avoid a stockout in 95 of every 100 replenishment cycles, using the normal distribution (see the NIST/SEMATECH e-Handbook linked below). They say nothing directly about how many units you will be short when a stockout does happen.

With invented numbers, you can see how the two diverge. Suppose demand over the lead time has a standard deviation of 40 units and you set safety stock for a 90% cycle service level. Under the usual normal assumption, the expected shortage in a cycle works out to about 1.9 units. If you reorder 400 units at a time, that is a unit fill rate of about 99.5%. If you reorder only 100 units at a time, you face the same risk four times as often, and unit fill drops to about 98.1%.

So a 90% cycle service level can deliver a fill rate well above 95%, and the gap depends on order quantity. If customers and managers judge you on fill rate, setting safety stock as if the target were a cycle service level usually overstocks. Planning texts such as Silver, Pyke and Thomas cover the method for solving safety stock directly for a fill-rate target.

:::caution
Do not type a customer's fill-rate target into a safety stock calculator that asks for a cycle service level. The two numbers are on different scales.
:::

## Setting up service reporting

1. Pick one primary fill measure for internal reporting (we suggest line fill on the first-shipment rule) and write the definition on the report.
2. Decide how substitutions, customer-requested partials, drop-ship, future-dated and cancelled lines count, and apply those rules every month.
3. Build each key customer's OTIF to their published definition, separately from your own.
4. When you set safety stock, be explicit about whether the target is a cycle service level or a fill rate. The next page, [Safety stock, reorder point and EOQ](/distribution/safety-stock-reorder-point-eoq/), shows how.

## Sources

- [SCOR Digital Standard (ASCM)](https://www.ascm.org/corporate-solutions/standards-tools/scor-ds/)
- [SCOR Digital Standard, introduction and front matter (ASCM, PDF)](https://www.ascm.org/globalassets/ascm_website_assets/docs/scor/intro-and-front-matter-scor-digital-standard-2025.pdf)
- [Normal distribution (NIST/SEMATECH e-Handbook of Statistical Methods)](https://www.itl.nist.gov/div898/handbook/eda/section3/eda3661.htm)

---

Epicor, Prophet 21, P21 and DynaChange are trademarks or registered trademarks of Epicor Software Corporation registered in the United States and other countries. Kinetic is a trademark of Epicor Software Corporation. Lumina ERP is an independent consultancy and is not affiliated with, sponsored by or endorsed by Epicor.