Two orders miss their promised dates on the same afternoon.
The first is a small spot order for a customer who has several alternative suppliers. It is worth $18,000, and the customer accepts a two-day delay with a brief apology and a revised booking.
The second belongs to a strategic account. It contains a launch-critical product, represents a much larger commercial relationship, and is tied to a production window at the customer’s site. It ships short and three days late. The customer has to stop a line, call an executive review, and reconsider the next contract award.
If a supply-chain dashboard counts both misses as one failure, it is technically consistent and commercially incomplete.
This is the problem with raw on-time in-full performance. It tells us how many transactions met a promise. It does not necessarily tell us which promises mattered most, how severe the miss was, whether the promise moved before failure, or what consequence the customer experienced.
The On-Time and In-Full Network Reliability Index, ONRI, exists to close that gap. It treats delivery reliability as a weighted, network-wide view of consequential commitments rather than a simple average of order lines. The objective is not to make performance look worse or to create a more complicated score. It is to make the business’s service reality more truthful.
Why average OTIF hides business impact
An average is useful when the things being averaged are reasonably comparable. Supply-chain orders rarely are.
One order may be a routine replenishment with flexible timing. Another may be a contractual commitment to a strategic customer. One may contain standard stock that can be replaced tomorrow. Another may be the final component needed to keep a customer’s production line running. One late line may cost a few extra dollars in freight. Another may create lost revenue, a penalty, a quality escalation, or a damaged relationship.
Yet many organizations calculate OTIF as a simple ratio:
This ratio is easy to explain and easy to compare. It also assigns the same numerical importance to unlike commitments.
Suppose a network delivers 990 of 1,000 order lines on time and in full. The result is 99 percent. If the ten misses are low-value, recoverable orders, the number may accurately describe a healthy service position. If one of those ten misses stops a strategic customer and the other nine are harmless administrative lines, the average hides the one event leadership most needs to understand.
The issue is not that the ratio is mathematically wrong. The issue is that the denominator is doing too much work. It assumes that every transaction carries equal business significance.
Average OTIF also hides concentration. A plant may report 96 percent performance across a region while missing nearly every promise for one customer segment. A business may show improvement because it fulfilled a large volume of easy orders while a smaller group of complex, high-consequence orders became less reliable.
The leadership question should therefore move from “What percentage did we deliver?” to “Which promises did we protect, which did we miss, and what did those misses mean?”
This shift is especially important during constrained periods. When capacity is plentiful, most orders can be fulfilled without an obvious choice. When material, production time, or transport space is scarce, the organization has to decide which commitments receive protection. A metric that treats every line equally cannot help much with that decision. ONRI makes the priority logic visible before the shortage becomes a customer escalation.
It turns prioritization from an informal argument into a governed business conversation.
That clarity is valuable when every available response has a cost.
It also makes the reasoning reviewable later.
The question ONRI answers
ONRI answers a more consequential question:
Are we reliably protecting the customer commitments that matter most to the business?
That question has three parts.
First, it recognizes that not all demand has equal criticality. Customer, product, contract, revenue, safety, regulatory, and operational context can change the importance of a delivery.
Second, it looks across the network. A promise may be created by sales, constrained by planning, fulfilled by a plant, held by quality, and delivered by logistics. Reliability belongs to the customer journey, not one department.
Third, it distinguishes the fact of a miss from its severity. A one-day delay on a flexible replenishment order is not equivalent to a short shipment that stops a production line. ONRI should make that difference visible.
ONRI is best understood as a management lens. It does not replace raw OTIF. Raw OTIF remains useful for process monitoring, operational discipline, and benchmarking. ONRI adds the business weighting and context needed for executive decisions.
A transparent ONRI can be expressed conceptually as:
The actual implementation may include order value, customer criticality, product importance, schedule-line weight, fulfilment completeness, date stability, and miss severity. The important design principle is that each factor must be defined, versioned, and explainable.
Order value is useful, but it is not enough
The first instinct when weighting service is to use order value. That is understandable. A $1 million order appears more important than a $1,000 order.
Value matters, but value alone creates new distortions.
A small spare part may be worth very little and still be essential to a customer’s maintenance shutdown. A low-value component may stop a high-value assembly. A strategically important emerging customer may place small orders today while representing significant future growth. A regulated product may have modest revenue but a large compliance consequence.
Order value should therefore be one input, not the definition of criticality.
A practical criticality model may consider:
- Commercial value and margin contribution.
- Strategic-account status and relationship importance.
- Contractual service obligations or penalties.
- Customer production-line or shutdown dependency.
- Product launch, project, or seasonal timing.
- Safety, quality, regulatory, or reputational consequence.
- Availability of substitutes or alternate suppliers.
- Recovery time if the promise is missed.
The model should be simple enough for people to understand and stable enough to govern. If a planner cannot explain why one order has a higher weight than another, the score will not be trusted.
Criticality should also be assigned at the right level. A strategic customer may have routine orders and exceptional orders. A single customer label applied to every line can overstate some commitments and understate others. The useful unit may be the customer-product-period combination, the schedule line, or the specific promise event.
A practical criticality example
Consider two orders.
Order A is a $12,000 spot purchase of standard material. It is delivered two days late, but the customer has stock and no production impact. The account team records dissatisfaction, but the commercial relationship is stable.
Order B is a $220,000 delivery to a strategic account. The order is delivered three days late and 20 percent short. The customer has no immediate substitute, and the missing quantity affects its own production schedule.
Raw OTIF sees two failures.
A weighted view may assign Order B much greater importance because its value, criticality, completeness, and consequence are all higher. That does not mean Order A should be ignored. It means the organization should not allocate the same leadership attention to both events.
The distinction improves resource allocation. The strategic miss may require a cross-functional recovery team and executive communication. The spot-order miss may require a process correction within customer service or transport. Counting failures is a useful starting point; understanding consequence tells the organization where to intervene.
Why schedule lines matter
Many supply chains measure at the order level even though the customer actually receives a schedule of commitments.
An order may contain multiple products, quantities, dates, destinations, or release windows. A single order can be partly on time and partly late. It can be delivered in several shipments. It can be accepted as complete only if all schedule lines arrive within defined tolerances.
If the system collapses those details into one order-level status, it can hide the operational pattern.
Schedule-line measurement helps answer:
- Which promised quantity was due on which date?
- Was the line delivered on time, early, late, or not at all?
- Was the full quantity delivered, or was the customer forced to accept a partial?
- Did the customer need all lines together for a usable outcome?
- Was the schedule changed after the original promise?
This matters particularly in configure-to-order, project, industrial, and make-to-order environments. The customer may not care that nine of ten lines arrived if the tenth line is the component that makes the shipment usable.
Schedule lines also improve root-cause analysis. They allow leaders to separate a single late order from a pattern of late releases, repeated partials, or a specific product family that is consistently promised beyond the network’s capability.
Date changes and date stability
A delivery can be on time against the latest date and still represent a broken promise.
Suppose an order was originally promised for 10 June. On 7 June, the date is moved to 14 June because material is unavailable. The order arrives on 14 June. A system that measures only against the latest confirmed date records an on-time delivery.
The customer may experience it differently. The original commitment was not kept; the organization revised the commitment before the failure became visible.
Date changes are not automatically bad. Customers sometimes request them. Commercial circumstances change. A transparent replan can be better than an unrealistic promise. But a date change should not erase the history of the promise.
ONRI should distinguish at least three concepts:
1. Original promise reliability: whether the first committed date was achieved.
2. Current promise reliability: whether the latest agreed date was achieved.
3. Date stability: how often and how materially the promise moved.
These measures tell different stories. A business may have good current-date OTIF but poor original-date reliability. That pattern suggests the organization is managing the dashboard by moving goalposts rather than improving execution.
Date stability also matters to customers even when the final delivery is technically acceptable. Planning confidence declines when promises change repeatedly. The customer may have to reschedule labor, transport, production, or inventory because the date is not dependable.
The governance principle is simple: preserve the original promise, record every approved change, and report the reason for the change. A revised date can be legitimate without becoming invisible.
Separating reliability from miss severity
Reliability answers whether the promise was met. Severity answers what happened when it was not.
A useful ONRI design keeps those dimensions visible instead of hiding them inside one opaque weight.
Miss severity may include:
- Days or hours late.
- Percentage of quantity short.
- Whether the shipment was usable without the missing quantity.
- Customer production or service impact.
- Financial penalty or lost margin.
- Premium freight or recovery cost.
- Whether the miss affected a launch, shutdown, or regulated commitment.
- Time available to recover before the promise became critical.
This creates more useful categories than simply pass or fail. A late delivery can be minor, material, critical, or catastrophic depending on context.
The distinction also protects teams from an unhelpful incentive. If every miss is treated identically, people may focus on avoiding the appearance of failure rather than preventing consequential failure. A transparent severity model encourages the organization to learn from both frequent small misses and rare major events.
It is important not to use severity weights to hide poor performance. A business should still report the raw count and rate of all misses. ONRI is an additional view, not a way to make the result look better by assigning low weights to inconvenient events.
How leaders should use ONRI
Leaders should use ONRI to direct attention, test trade-offs, and improve promises; not to create another isolated target.
Use ONRI to focus the operating review
Begin with the weighted result, then open the gap between ONRI and raw OTIF. If raw OTIF is 96 percent and ONRI is 88 percent, the difference is meaningful. It suggests the misses are concentrated among more consequential commitments.
If raw OTIF is 88 percent and ONRI is 95 percent, the organization may be missing many low-criticality orders while protecting strategic commitments. That may be acceptable for a period, but it still deserves a process improvement plan.
Use ONRI to prioritize recovery
When capacity, inventory, or transport is constrained, ONRI can support explicit prioritization. It can help answer which order should receive scarce stock, which customer should receive a communication first, and where premium freight is justified.
The decision should never be made by the index alone. The index makes the business consequence visible; authorized leaders decide the trade-off.
Use ONRI to examine promise design
Repeated misses are not always an execution problem. The commercial organization may be promising dates without checking capacity, quality lead time, or logistics constraints. ONRI trends can reveal that a product family or customer segment is being promised beyond what the network can reliably deliver.
The answer may be better available-to-promise logic, clearer customer segmentation, different lead-time rules, or a conversation about service levels and price.
Use ONRI to learn from interventions
Record how a high-criticality promise was protected. Did the team reallocate inventory, expedite, resequence, substitute, split the shipment, or reset the date? What did the intervention cost? Did it create another miss?
This turns service performance into organizational learning. Over time, the organization can identify which recovery actions are effective, which are expensive, and which merely move the problem downstream.
Governance: do not move the goalposts
Any weighted measure creates governance risk. If the weights are changed whenever the result is uncomfortable, the index becomes a communication device rather than a decision instrument.
ONRI needs a documented definition with version control. The definition should specify:
- The unit of measurement: order, line, schedule line, shipment, or commitment.
- The time boundary and time zone.
- The treatment of customer-requested changes.
- The rules for partial shipments and early deliveries.
- The source of customer and product criticality.
- The weight ranges and approval authority.
- The treatment of cancellations, holds, force majeure, and disputed dates.
- The method for reporting missing or uncertain data.
- The effective date of each definition version.
Criticality weights should be reviewed periodically, but not silently. A change from one scoring version to another should be visible in reporting. Historical results should not be recalculated without an explanation, because trend interpretation depends on a stable basis.
There is also a temptation to make the index too precise. A score of 87.43 can create false confidence if the underlying criticality data is incomplete or manually maintained. In many environments, a banded result: high, medium, or low reliability, may be more honest than a decimal-level score.
Governance should include the people whose work the metric affects. Customer service, commercial, planning, operations, logistics, finance, and data owners should agree on the meaning before the target is attached. Otherwise the index will become a negotiation over weights rather than a shared view of reliability.
Questions for a leadership review
The following questions help leaders use ONRI as a decision tool:
- Which customers, products, or commitments drive the gap between raw OTIF and ONRI?
- Are high-criticality misses concentrated in one plant, supplier, route, product family, or promise type?
- How much of the performance depends on revised dates rather than original promises?
- Which schedule lines were technically delivered but operationally unusable?
- What percentage of high-criticality misses were late, short, or both?
- Which recovery actions protected service, and what did they cost?
- Are criticality weights transparent to the teams making daily allocation decisions?
- What customer or commercial consequences are not represented in the current score?
- Has the definition changed, and can we compare this period fairly with the last one?
- What decision should change because of the ONRI result?
The last question is the most important. A metric that does not change attention, allocation, promise design, or improvement work is only reporting.
Start small and make the weighting visible
An organization does not need a perfect enterprise model to begin. Choose one customer segment, product family, or order journey where the limitations of average OTIF are already obvious.
For the first version, retain raw OTIF and add three transparent dimensions: customer criticality, order or line value, and miss severity. Preserve original and revised dates. Review the result with the teams who know the exceptions. Test whether the weighted ranking identifies the events leaders actually care about.
Then refine. Add schedule-line logic where orders are being collapsed. Add fulfilment conformance where partial shipments are common. Add recovery cost where service is being protected through premium actions. Add customer impact where the operational result does not capture the commercial consequence.
The aim is not to create a perfect score. It is to create a better question and a more honest conversation.
Reliability should reflect the promises that matter
Supply chains do not exist to complete equal transactions. They exist to create and protect value for customers and the business under real constraints.
Raw OTIF tells us how many promises were met. ONRI asks whether the promises that mattered most were met, whether the original commitment remained stable, how severe the misses were, and where leadership attention should go next.
The measure should never be used to excuse low-value failures or to make performance appear better than it is. Its purpose is prioritization and truth. By preserving raw results alongside weighted reliability, the organization can see both the breadth of execution and the depth of consequence.
The most mature supply chains will use ONRI not as another number to chase, but as a way to improve promise design, allocation decisions, recovery playbooks, and cross-functional learning.
Reliability should reflect the promises that matter most, not merely the number of transactions completed.
That is the discipline ONRI is meant to create.
Disclaimer
Industry situations in this chapter are composite illustrations unless explicitly attributed to a public source. They are not claims about any particular company, plant, vendor, or incident. External standards, research, and public case studies should be verified before publication. Implementations must be validated against local safety, quality, cybersecurity, regulatory, contractual, labour, privacy, and data-governance requirements. AI recommendations and autonomous actions should remain within clearly defined human authority, operational controls, and tested recovery procedures.
#SupplyChain #OTIF #SupplyChainMetrics #CustomerService #OrderFulfillment #Logistics #SupplyChainAnalytics #DecisionIntelligence #OperationsManagement #SupplyChainLeadership
Takeaways
Excerpt | Practical point / context |
|---|---|
“If a supply-chain dashboard counts both misses as one failure, it is technically consistent and commercially incomplete.” | Average OTIF can hide differences in customer and business consequence. |
“OTIF is useful. It is not sufficient.” | Raw service performance should be complemented, not discarded. |
“Are we reliably protecting the customer commitments that matter most to the business?” | The central question ONRI is designed to answer. |
“Order value should therefore be one input, not the definition of criticality.” | Revenue alone cannot determine the importance of a promise. |
“The customer may experience it differently. The original commitment was not kept.” | Revised dates should not erase the history of the original promise. |
“Reliability answers whether the promise was met. Severity answers what happened when it was not.” | Separate delivery status from the consequence of failure. |
“A metric that does not change attention, allocation, promise design, or improvement work is only reporting.” | Measures should lead to decisions and action. |
“The aim is not to create a perfect score. It is to create a better question and a more honest conversation.” | ONRI is a management lens, not a false-precision ranking. |
“Reliability should reflect the promises that matter most—not merely the number of transactions completed.” | The article’s closing principle. |
Further reading
- Why AI Should Optimize Decisions, Not Dashboards
DATTS
ONRI is designed to help leaders decide which customer commitments deserve attention and protection, not merely to report a percentage. This related article reinforces the decision-centric principle that metrics and AI should support a named choice, owner, constraint, and outcome.
- The Rise of Decision Products
DATTS
The article turns ONRI into an operating practice by linking the score to recovery, prioritization, and promise decisions. The decision-products article explains how to package this evidence, context, alternatives, workflow, and learning into a maintained capability.
- From Data Lakes to Decision Lakes
DATTS
ONRI depends on preserving original promises, date changes, interventions, cost, customer consequence, and outcome. The decision-lake article is relevant because it describes how to retain that context for audit, learning, and better future reliability decisions.
- The Boardroom Meets the Control Room
DATTS
ONRI makes customer-critical service exposure visible, but leaders still need to connect that exposure to planning, production, quality, allocation, and logistics actions. This related article provides the executive-to-control-room perspective needed to turn ONRI findings into coordinated action.
- A Roadmap to the Autonomous Manufacturing Enterprise
DATTS
The autonomy roadmap is relevant because it explains how decision authority should expand only when evidence, safeguards, escalation, and recovery readiness are sufficient.
