InfiniSynapse
What is OTIF, and what is the formula?什么是 OTIF？计算公式是什么？
OTIF means on-time in-full: the share of eligible orders or deliveries that satisfy the agreed timing rule and completeness rule simultaneously. A delivery that is early or late outside the approved window fails “on time”; a delivery that is short of the agreed quantity fails “in full.” Passing only one condition earns no OTIF credit.
OTIF 是 On-Time In-Full（准时足量交付率）：在合格订单或交付中，同时满足约定时间规则和完整性规则的占比。早到或迟到且超出允许窗口的交付不算“准时”；低于约定数量的交付不算“足量”。只满足其中一项不能计入 OTIF 成功数。
The word AND is the important part. Do not average on-time delivery and fill rate. Do not assume that multiplying those percentages produces OTIF. The same order-level, delivery-level, line-level, case-level, or unit-level record must pass both tests. Which grain is appropriate depends on the commercial agreement and analytical purpose.
公式中最重要的是“且（AND）”。不要把准时交付率与满足率求平均，也不要假定两者相乘就等于 OTIF。同一条订单级、交付级、行项目级、箱级或件级记录必须同时通过两项测试。采用哪一种粒度，应由商业协议和分析目的决定。
Define the OTIF measurement contract before calculating计算 OTIF 前先定义指标口径
OTIF is simple arithmetic built on consequential business definitions. Two teams can calculate different rates from the same transactions and both be mathematically correct. Before publishing a score, write a metric contract that answers the following questions.
OTIF 的算术很简单，但背后是影响重大的业务定义。两个团队可能基于同一批交易计算出不同结果，而且数学上都没有错误。发布指标前，应书面回答以下口径问题。
| Decision口径决策 | Choices to document需要记录的选择 | Why it changes OTIF为何会改变 OTIF |
|---|---|---|
| Promise date承诺日期 | Customer request, original confirmation, latest confirmation, appointment, or contract date客户要求日、首次确认日、最新确认日、预约日或合同日期 | Repeatedly re-promising late orders can make performance look better without improving service.反复重设逾期订单的承诺日，可能在服务未改善时美化结果。 |
| Delivery event交付事件 | Ship, carrier handoff, arrival, check-in, unload, or goods receipt timestamp发运、承运商接货、到达、签到、卸货或收货过账时间 | A warehouse ship-time metric is not the same as customer delivery performance.仓库按时发运并不等于客户按时收到。 |
| On-time window准时窗口 | Exact date, not-late, ±1 day, appointment window, timezone, weekends, and holidays准确日期、不迟到、前后一天、预约窗口、时区、周末与节假日 | Early deliveries may create storage and receiving problems and may need to fail.过早到货可能造成存储和收货压力，因此也可能需要判定失败。 |
| In-full rule足量规则 | Exactly 100%, agreed tolerance, ordered versus confirmed quantity, substitutions, cancellations, and overages严格 100%、约定容差、订购量或确认量、替代品、取消与超量 | A 99% unit fill can still fail a strict whole-order rule.在严格的整单规则下，99% 数量满足率仍可能失败。 |
| Population and grain统计范围与粒度 | Orders due, orders shipped, deliveries received, order lines, cases, or units; exclusions and cancellations到期订单、已发运订单、已收货交付、订单行、箱或件；以及排除项与取消项 | Changing the denominator or grain changes the weight assigned to large and small orders.改变分母或粒度，会改变大订单和小订单的权重。 |
Version the contract and show the definition beside the dashboard. If a customer or supplier contract specifies a different rule, calculate that contractual OTIF separately rather than silently blending it with an internal service metric.
应对口径进行版本管理，并在仪表板旁展示定义。如果客户或供应商合同规定了不同规则，应单独计算合同 OTIF，而不是与内部服务指标静默混合。
Data required for a defensible OTIF calculation可核验 OTIF 计算所需的数据
At minimum, each measurement unit needs a stable identifier, original and current promised dates, an actual delivery event, ordered or confirmed quantity, delivered or accepted quantity, status, and the business dimensions used for diagnosis. Preserve timestamp history rather than overwriting promised dates. Normalize timezones and units of measure before comparing records.
每个统计单元至少需要稳定标识、原始与当前承诺日期、实际交付事件、订购量或确认量、交付量或验收量、状态，以及诊断所需的业务维度。应保留时间戳历史，不要覆盖原承诺日期；比较前还要统一时区和计量单位。
Order ID, line ID, customer request date, original confirmed date, latest confirmed date, ordered quantity, cancellations, and change reason.
订单 ID、行项目 ID、客户要求日、首次与最新确认日、订购数量、取消记录及变更原因。
Shipment ID, warehouse release, carrier events, destination arrival, proof of delivery, receipt posting, accepted quantity, and exception codes.
发运 ID、仓库放行、承运商事件、目的地到达、签收证明、收货过账、验收数量及异常代码。
Join one-to-many shipments and receipts carefully. A split delivery may create several event rows for one order. Decide whether the order passes only after the complete quantity arrives, whether an agreed split is acceptable, and which arrival timestamp closes the measurement.
处理一对多的发运与收货记录时必须谨慎。一个拆分交付可能对应多条事件记录。需要明确：是否必须等全部数量到齐后整单才通过、约定分批是否可以接受，以及用哪个到达时间关闭统计。
OTIF calculation example, step by stepOTIF 完整计算示例
Hypothetical example: a distributor has 1,000 eligible orders due in one month under an order-level rule. Nine hundred twenty arrive within the agreed window. Nine hundred fifty meet the in-full rule. Only 880 orders satisfy both conditions on the same record.
假设示例：某分销商一个月内有 1,000 个符合条件的到期订单，采用整单统计规则。其中 920 单在约定窗口内到达，950 单满足足量规则，但只有 880 单在同一订单记录上同时满足两项条件。
OTIF is 880 ÷ 1,000 × 100 = 88.0%. The 40-point gap between on-time orders and OTIF orders means 40 on-time orders failed the completeness test. The 70-point gap between in-full orders and OTIF orders means 70 complete orders failed the timing test. These groups may overlap only in the complement, so failure flags should be retained for accurate root-cause analysis.
OTIF = 880 ÷ 1,000 × 100 = 88.0%。准时订单与 OTIF 成功订单相差 40 单，说明有 40 个准时订单未通过足量测试；足量订单与 OTIF 成功订单相差 70 单，说明有 70 个足量订单未通过时间测试。失败原因在补集内可能重叠，因此必须保留失败标记，以便准确分析根因。
OTIF calculatorOTIF 准时足量交付率计算器
Enter counts from one population, period, and measurement grain. “Both” must not exceed the on-time count or the in-full count. This calculator reports component rates for diagnosis but calculates OTIF from the strict intersection.
请输入来自同一统计范围、期间和粒度的数量。“同时满足”不能超过准时数或足量数。计算器会显示两个分项指标用于诊断，但 OTIF 严格按交集计算。
For operational reporting, reproduce the result from row-level flags and reconcile the denominator to a controlled order or delivery report. A calculator validates arithmetic; it does not validate promise-date history, exclusions, delivery evidence, or the business definition.
在运营报表中，应根据明细级标记复算结果，并将分母与受控订单或交付报表核对。计算器只能验证算术，无法验证承诺日期历史、排除项、交付证据或业务定义。
OTIF vs on-time delivery, fill rate, and perfect order rateOTIF 与准时交付率、满足率和完美订单率的区别
| Metric指标 | Core question核心问题 | Important limitation主要局限 |
|---|---|---|
| OTIF | Did the same delivery meet both timing and completeness rules?同一次交付是否同时满足时间与完整性规则？ | Highly sensitive to promise date, window, grain, and denominator definitions.对承诺日期、窗口、粒度和分母定义高度敏感。 |
| On-time delivery准时交付率 | Did delivery happen within the agreed window?交付是否发生在约定窗口内？ | Can look strong when incomplete orders arrive on time.即使订单不完整，只要准时到达，指标也可能很好看。 |
| Fill rate / in-full rate满足率／足量率 | How much demand or how many records were fulfilled completely?有多少需求或记录被完整满足？ | May ignore whether the quantity arrived when needed; unit and order fill are different.可能忽略数量是否按时到达；件级和订单级满足率也不同。 |
| Perfect order rate完美订单率 | Did the order meet timing, completeness, accuracy, damage, and documentation rules?订单是否同时满足时间、完整性、准确性、无损与单据规则？ | Broader and usually lower than OTIF; requires more reliable event and quality data.覆盖范围更广，通常低于 OTIF，并需要更可靠的事件与质量数据。 |
Use OTIF as an outcome metric, then keep its components and supporting measures visible. Inventory availability, pick accuracy, dock-to-stock time, carrier transit variance, appointment compliance, and proof-of-delivery completeness explain different parts of the outcome. OTIF alone does not identify who caused a failure.
可把 OTIF 用作结果指标，同时保留分项与支持指标。库存可用性、拣选准确率、月台至入库时间、承运时效波动、预约遵守率和签收证明完整性，分别解释结果的不同部分。OTIF 本身不能判断失败责任方。
Diagnose OTIF failures without hiding the overlap在不隐藏重叠的情况下诊断 OTIF 失败
Create separate flags for on-time pass, in-full pass, OTIF pass, and every known exception. Then classify four mutually exclusive outcomes: both pass, time-only pass, full-only pass, and neither pass. This prevents double-counting when a delivery is both late and short.
应分别建立准时通过、足量通过、OTIF 通过和各类异常标记，再把结果划分为四个互斥类别：两项都通过、只通过时间、只通过足量、两项都失败。这样可以避免迟到且短装的交付被重复计数。
Production delay, allocation delay, late picking, missed cutoff, carrier pickup delay, transit variance, rejected appointment, receiving congestion, or timestamp latency.
生产延迟、分配延迟、拣选逾期、错过截单、承运商提货延迟、运输波动、预约被拒、收货拥堵或时间戳延迟。
Stockout, inaccurate inventory, short pick, damaged units, substitution rejection, packaging multiple mismatch, unapproved split, or receipt discrepancy.
缺货、库存不准、少拣、货损、替代品被拒、包装倍数不匹配、未批准拆分或收货差异。
Segment the rate and failure counts by customer, service tier, supplier, warehouse, carrier, lane, SKU, product family, promise lead time, and week. Always show the eligible count with the percentage: 50% from two orders is not equivalent evidence to 50% from two thousand. Avoid publishing supplier or carrier rankings until volume mix and controllability are reviewed.
可按客户、服务等级、供应商、仓库、承运商、线路、SKU、品类、承诺交期和周次拆解比率与失败数量。百分比旁必须显示合格样本数：2 单中的 50% 与 2,000 单中的 50% 证据强度完全不同。在审查业务量组合与可控性之前，不要直接发布供应商或承运商排名。
A reproducible OTIF analysis workflow可复现的 OTIF 分析工作流
- Approve the metric contract.批准指标口径。 Name the owner, business purpose, date source, delivery event, tolerance, completeness threshold, grain, denominator, exclusions, timezone, and version date.明确责任人、业务目的、日期来源、交付事件、容差、足量阈值、粒度、分母、排除项、时区和版本日期。
- Build a stable eligible population.建立稳定的合格统计范围。 Start from orders due in the period or another approved population. Freeze or snapshot it so late postings do not silently rewrite history.从期间到期订单或其他批准范围开始，并冻结或制作快照，避免延迟过账静默改写历史。
- Preserve promise and event history.保留承诺与事件历史。 Keep original, confirmed, and revised dates with reason codes. Reconcile carrier events and receipts instead of selecting whichever timestamp creates a pass.保留原始、确认与修改后的日期及原因代码；核对承运与收货事件，不要选择最容易判定通过的时间戳。
- Calculate row-level flags.计算明细级标记。 Evaluate on-time and in-full independently on the same grain, then set OTIF pass only where both flags are true. Retain failure reasons and data-quality flags.在同一粒度上分别判断准时与足量，只有两项均为真时才标记 OTIF 通过，并保留失败原因与数据质量标记。
- Reconcile and test.核对并测试。 Tie total records and quantities to controlled operational reports. Sample pass and fail records, inspect split deliveries, duplicates, cancellations, missing evidence, boundary dates, and timezone transitions.将记录数和数量与受控运营报表核对，并抽查通过与失败记录、拆分交付、重复项、取消项、证据缺失、边界日期和时区转换。
- Publish outcome and diagnosis together.同时发布结果与诊断。 Show OTIF, component rates, counts, trend, definition version, exception share, and the highest-volume actionable failure groups. Assign owners and verify whether actions change service, cost, and inventory.同时展示 OTIF、分项比率、数量、趋势、口径版本、异常占比和高业务量可行动失败组，并分配责任人，验证措施是否改变服务、成本与库存。
How to improve OTIF without gaming the metric如何在不操纵指标的情况下改善 OTIF
Start with the largest controllable failure family, not a generic target. If short shipments dominate, examine forecast error, allocation, safety stock policy, supplier reliability, inventory accuracy, and pick exceptions. If lateness dominates, examine promise logic, capacity, cutoff adherence, dock scheduling, carrier pickup, lane variability, and receiving constraints.
应从最大且可控的失败类别开始，而不是先设一个空泛目标。如果短装占主导，应检查预测误差、库存分配、安全库存政策、供应商可靠性、库存准确性和拣选异常；如果迟到占主导，则应检查承诺逻辑、产能、截单遵守、月台排程、承运商提货、线路波动和收货限制。
Guardrail: a higher OTIF is not automatically a better business outcome if it is achieved by extending every promise, holding excessive inventory, using costly emergency freight, or excluding difficult orders. Monitor lead time, inventory, expedite cost, cancellations, damage, margin, and customer complaints beside OTIF.
护栏：如果更高 OTIF 来自延长所有承诺、持有过量库存、频繁使用高成本紧急运输或排除困难订单，它不一定代表更好的商业结果。应同时监控交期、库存、加急成本、取消、货损、毛利和客户投诉。
Run changes as controlled operational experiments when possible. Define the affected population and expected mechanism, compare pre/post periods on the same metric version, inspect unintended shifts, and confirm persistence. A dashboard trend is evidence of change, not proof that one initiative caused it.
条件允许时，应把改善措施作为受控运营实验。明确受影响范围与预期机制，在同一指标版本下比较前后期间，检查意外变化并确认效果是否持续。仪表板趋势只能证明发生了变化，不能单独证明某项措施导致了变化。
Analyze OTIF across connected order and delivery data跨订单与交付数据分析 OTIF
Prepare order promises, shipment and receipt events, quantities, exception codes, and the approved OTIF definition. InfiniSynapse can be used as an analysis layer to explore connected data, compare segments, and investigate changes. It does not replace ERP, WMS, TMS, carrier execution, purchasing, or contractual approval workflows.
请准备订单承诺、发运与收货事件、数量、异常代码和已批准的 OTIF 定义。InfiniSynapse 可作为分析层，用于探索关联数据、比较分群并调查变化；它不替代 ERP、WMS、TMS、承运执行、采购或合同审批流程。
Open the InfiniSynapse analytics workspace打开 InfiniSynapse 在线分析工作区Frequently asked questions about OTIFOTIF 常见问题
What does OTIF mean?OTIF 是什么意思？
OTIF means on-time in-full. It is the percentage of eligible orders or deliveries that satisfy the agreed timing rule and the agreed completeness rule at the same time.
OTIF 即 On-Time In-Full，中文可称准时足量交付率，表示在合格订单或交付中，同时满足约定时间规则与完整性规则的占比。
What is the OTIF formula?OTIF 的计算公式是什么？
OTIF percentage equals the number of eligible orders delivered both on time and in full divided by total eligible orders, multiplied by 100. Apply both tests to the same measurement unit.
OTIF（%）等于同时准时且足量的合格订单数除以合格订单总数，再乘以 100。两项测试必须作用于同一个统计单元。
Can I multiply on-time delivery rate by fill rate?可以用准时交付率乘以满足率计算 OTIF 吗？
Not reliably. The two component failures may occur on different records and are not necessarily independent. Calculate OTIF directly from records where both flags are true.
不能可靠地这样计算。两项失败可能发生在不同记录上，也不一定相互独立。应直接统计同一记录上两项标记均为真的数量。
What is a good OTIF rate?什么水平的 OTIF 算好？
There is no universal rate. Use the contractual requirement where one exists; otherwise establish a like-for-like baseline, compare stable segments, and set an improvement target that accounts for industry, service promise, sample size, cost, inventory, and customer impact.
不存在适用于所有企业的统一水平。有合同要求时应以合同为准；否则应建立可比基线、比较稳定分群，并结合行业、服务承诺、样本量、成本、库存和客户影响制定改善目标。
Should OTIF use the original or revised promise date?OTIF 应使用原始承诺日还是修改后的承诺日？
The metric contract must specify this. Original customer-request or confirmed dates often preserve the planning experience, while an approved revision may be relevant to a contractual measure. Keep both dates and report separate views when the business questions differ.
必须由指标口径明确规定。原始客户要求日或确认日通常更能保留计划体验，而经批准的修改日期可能适用于合同指标。应保留两类日期，并在业务问题不同的情况下分别报告。
Sources, definition boundaries, and verification资料来源、定义边界与核验说明
The strict numerator-over-total formula was checked against the Manufacturers Alliance Foundation's Order Management KPIs reference. The need to define on-time and in-full criteria, including the confirmed delivery time and confirmed quantity, was checked against the first-party Kongsberg Automotive Supplier Logistics Manual.
严格的“成功数除以总数”公式参考了 Manufacturers Alliance Foundation 的订单管理 KPI 资料。关于准时与足量标准、确认交付时间和确认数量的定义要求，则参考了第一方的Kongsberg Automotive 供应商物流手册。
This page provides an analytical workflow, not contractual, accounting, legal, audit, procurement, or logistics execution advice. Definitions vary by trading agreement, ERP model, industry, and responsibility split. Approve the rule with commercial and operational owners before using OTIF for penalties or supplier decisions. For monitoring design after calculation, see the InfiniSynapse guide to a focused data dashboard. Inventory availability can also be investigated alongside the planned days inventory outstanding guide when both pages are deployed.
本页提供分析工作流，不构成合同、会计、法律、审计、采购或物流执行建议。定义会因交易协议、ERP 数据模型、行业和责任划分而异。在把 OTIF 用于罚则或供应商决策前，应由商业与运营责任人批准口径。完成计算后，可参考 InfiniSynapse 的聚焦型数据仪表板指南设计监控方式；两页部署后，也可结合存货周转天数指南调查库存可用性。