OTIF Diagnosis: Five Checks That Find Where Your Service Level Actually Leak

OTIF has one home ground: a defined order, for a defined set of items, promised for a defined date. Picture a B2B buyer ordering five forklifts from a manufacturer. The order counts as complete only when all five arrive on the date promised, with no shelf and no store inventory standing between the order and the answer. OTIF compresses two failures into one score, and a monthly figure of 87% confirms a problem exists without saying whether the order was promised, planned, sourced, or dispatched badly. This guide explains five checks that isolate the leak before any corrective spend is approved.

Why the headline OTIF number hides the cause

Made-to-order failures rarely start on the shop floor. They start with a promise date issued before anyone confirmed that capacity, material, and sequencing could actually support it. Most organisations run one common ERP across plants and depots. Transactions are shared. Planning decisions are not, because they live in spreadsheets held by individual sites and planners. The order gets promised against a plan nobody can see end to end, and OTIF absorbs the damage a cycle later. The leak generally sits in that gap between the promise and the plan.

Check 1: Split the OTIF score into on-time and in-full

Report three numbers instead of one: OTIF, on-time rate, and in-full rate. Each split points at a different part of the chain.

Pattern	Likely origin	Where to look first
On-time high, in-full low	Availability and allocation	Component positioning, safety stock policy, material segmentation
In-full high, on-time low	Scheduling and lead time	Master production schedule, sequencing, dispatch discipline
Both low on the same customers	Order promising	Commitment logic and available-to-promise checks

Run the split at order line level. Order-level reporting lets one short line mark a full shipment as a failure.

Check 2: Test whether the promised date was achievable at commit

A missed date counts as a planning failure only when the date was reachable. Rebuild the plan as it stood on the commit date and identify the constraint that broke.

Manual overrides matter here. Across more than 60,000 forecasts from four supply chain companies, researchers found smaller judgmental adjustments often damaged accuracy, and upward adjustments were more frequently made in the wrong direction, pointing to a consistent optimism bias. Commitments built on inflated numbers surface as OTIF misses one cycle later.

Measure the share of failed OTIF lines where the promise date was already infeasible against capacity or material at the moment it was given.

Check 3: Audit the order and material records behind the failures

A promise is only as good as the data behind it. An available-to-promise check run against a stale material record, an unconfirmed supplier date, or an outdated routing produces a wrong date before the order is even confirmed. Supply chain planning software plans against those same records, so a bad input reaches the commitment unchanged.

Pull the order and material records behind last quarter’s OTIF failures and check them against what was physically true on the commit date. Where the two disagree, the fault sits in data integrity, not planning logic.

Check 4: Trace every failed OTIF line to one owning decision

Assign each failure to one of five decisions: the forecast, the inventory policy, the production sequence, the material call-off, or the dispatch plan. Build a Pareto across a full quarter. Three categories usually carry most of the volume, and they become the fix list.

The diagnosis runs at shift-level cadence. Reviewing failures once a shift corrects the plan while the period is still open, and no organisation needs minute-level refresh. Once the pattern stabilises, supply chain planning software sitting above the ERP can automate the classification and flag exceptions before a commitment goes out. Corrective steps to improve OTIF follow the diagnosis rather than preceding it.

Frequently asked questions

Why does OTIF fall while total inventory rises?

Supply chain planning software positions material against the demand signal it receives. When cover builds at the wrong nodes or on the wrong components, inventory value climbs while on-time delivery still drops on the orders that matter. The correction is positioning and segmentation, and additional volume rarely moves the OTIF number.

Should OTIF be measured at order level or line level?

Line level. Order-level scoring distorts the picture, since one short line marks an entire order as an OTIF failure while several short lines inside one order still register as a single miss. Line level keeps each failure attached to specific items and nodes.

How frequently should an OTIF diagnosis run?

Monthly reporting suits governance. The diagnosis itself works better at shift-level or daily cadence, because failed lines can still be traced back to the exact plan version that produced them. Month-end reviews strip away the context needed to name the causal decision.

Closing

A service level number becomes useful once each OTIF failure carries an owner and a cause. These five checks turn a reported shortfall into a fix list for planners, operations heads, and technology teams, using order history already inside the ERP.

Talk to our team to run this five-check OTIF diagnosis against your last four quarters of order data.