--- title: "Dispatch KPIs With Formulas and a Practical Scorecard" url: "https://www.upperinc.com/blog/dispatch-kpis/" date: "2026-10-13T11:00:18+00:00" modified: "2026-10-13T11:00:21+00:00" type: "Article" resource: "https://www.upperinc.com/blog/dispatch-kpis/" timestamp: "2026-10-13T11:00:21+00:00" author: name: "Riddhi Patel" categories: - "Dispatch" word_count: 2194 reading_time: "11 min read" summary: "Dispatch KPIs are measures that show whether work was assigned promptly, delivered as promised, and completed at an acceptable cost. Each one needs a defined population, a time period, and an event..." description: "Define 11 dispatch KPIs with clear formulas and denominators. Use a worked scorecard to keep late, incomplete, unattempted, and unassigned orders visible." keywords: "dispatch KPIs,dispatch metrics,dispatcher KPIs,dispatch performance metrics, Dispatch" language: "en" schema_type: "Article" related_posts: - title: "How to Calculate and Maximize Auto Dispatch ROI" url: "https://www.upperinc.com/blog/auto-dispatch-roi/" - title: "Second Trip vs Extra Vehicle for Delivery Overflow" url: "https://www.upperinc.com/blog/second-trip-vs-extra-vehicle/" - title: "How to Reduce Dispatcher Workload: 5 Practical Fixes" url: "https://www.upperinc.com/blog/reduce-dispatcher-workload/" --- # Dispatch KPIs With Formulas and a Practical Scorecard _Published: October 13, 2026_ _Author: Riddhi Patel_ ![Dispatch KPIs](https://www.upperinc.com/wp-content/uploads/2026/10/dispatch-kpis-heros-1024x585.png) Key takeaways - OTD measures timeliness; OTIF also requires the order to be delivered in full. - Use one clearly defined population, time period, and event timestamp per measure. - Dividing by completed deliveries instead of orders due hides every undelivered order. State which denominator you use. - Keep unassigned and unattempted work visible beside completion-based metrics. - Investigate changes before attributing them to software or driver performance. Dispatch KPIs are measures that show whether work was assigned promptly, delivered as promised, and completed at an acceptable cost. Each one needs a defined population, a time period, and an event timestamp. Without those three, two dispatchers can report different numbers for the same day. Start with a small scorecard tied to decisions your team can change. Keep the definitions fixed so an apparent improvement does not come from changing the denominator. [AI dispatch software](https://www.upperinc.com/ai-dispatcher/) can support assignment decisions, but a software label does not tell you how a metric is calculated. This guide defines 11 dispatch metrics with formulas and shows how to avoid a common reporting error: making performance look better by leaving unfinished work out. It covers delivery and field service dispatch, where the unit of work is an order or a job. Long-haul freight measures such as revenue per mile and deadhead percentage are not covered here. ## The 11 dispatch KPIs at a glance | **Dispatch KPI** | **Formula** | **Denominator population** | |---|---|---| | 1. On-time delivery rate (OTD) | On-time eligible orders ÷ eligible orders due × 100 | Orders due in the period, including undelivered | | 2. On-time in-full rate (OTIF) | Orders on time and in full ÷ eligible orders due × 100 | Orders due in the period, including undelivered | | 3. First-attempt delivery success | Orders completed on first attempt ÷ orders receiving a first attempt × 100 | Attempted orders only | | 4. Cost per completed delivery | Delivery operating costs ÷ deliveries completed | Completed deliveries | | 5. Driver time utilization | Productive hours ÷ available hours × 100 | Scheduled available hours | | 6. Travel time share | Driving hours ÷ paid hours × 100 | Paid hours | | 7. Initial dispatch response time | Total time to first meaningful response ÷ requests answered | Requests receiving a response | | 8. Time to confirmed assignment | Total time from ready-to-plan to confirmed assignment ÷ orders assigned | Assigned orders | | 9. Verified hard-rule match rate | Assignments with every required check verified ÷ reviewed assignments × 100 | Reviewed assignments | | 10. Changes after dispatch | Distinct orders reassigned or resequenced ÷ orders released × 100 | Released orders | | 11. Route finish variance | Actual finish time minus planned finish time, in minutes | Per route, reviewed as a distribution | ## Define the dispatch KPI scorecard before calculating it For each metric, record its owner, purpose, numerator, denominator, source fields, reporting period, and exclusions. Say whether you count orders, stops, routes, attempts, or drivers. One stop may contain several orders, so those units are not interchangeable. When a denominator is zero, report the metric as N/A rather than zero or 100 percent. Use a fixed cutoff for deciding which work belongs in the period. Retain the original customer commitment and approved changes. Quietly moving a late order into tomorrow’s population makes today’s report look better without fixing the miss. The definitions below are a practical measurement design. Adapt them to your service commitments, then document any change. They are not universal industry benchmarks or a claim that every dispatch product reports these exact formulas. ## Delivery reliability metrics ### 1. On-time delivery rate (OTD) OTD measures the share of eligible orders delivered within the agreed timing requirement. OTD = eligible orders delivered on time ÷ all eligible orders due in the period × 100. For this scorecard, an order due but not delivered remains in the denominator. Define whether the commitment applies to arrival, unloading start, or completion. Record approved cancellations and changes consistently. Note the denominator difference between reports. Upper’s [on-time delivery KPI guide](https://www.upperinc.com/blog/on-time-delivery-kpi/) uses total deliveries as the denominator and excludes canceled orders, which is a completion-based view. This scorecard uses orders due. Both are valid, but a report must say which one it uses. ### 2. On-time in-full rate (OTIF) OTIF adds a quantity check to the same eligible population. OTIF = eligible orders delivered both on time and in full ÷ all eligible orders due in the period × 100. Define “in full” at the order level, including how partial deliveries and substitutions count. Do not multiply separate OTD and in-full percentages: you need the count satisfying both conditions. Track damage separately unless your agreed definition explicitly includes it. ### 3. First-attempt delivery success This measures successful delivery on the first physical attempt. First-attempt success = orders completed on their first attempt ÷ orders receiving a first attempt in the period × 100. Orders never attempted are outside this denominator, so report their count beside it. Do not call this first-time fix rate: completing a delivery and resolving a service fault on the first visit are different outcomes. ## Cost and resource metrics ### 4. Cost per completed delivery Cost per completed delivery = delivery operating costs incurred in the period ÷ deliveries completed in the period. Define the cost pool, including labor, vehicle running costs, software, and any allocated overhead. Costs from failed attempts still belong in the relevant cost pool. For the difference between average cost and the cost of one extra stop, see the [cost to serve guide](https://www.upperinc.com/blog/cost-to-serve/). Keep units consistent and disclose any large job that starts in one period and finishes in the next. For cost per mile, use miles as the denominator and state which costs are included. ### 5. Driver time utilization For this scorecard, productive time includes delivery travel, loading, and customer service. Available time is scheduled time available for those activities after agreed exclusions such as breaks. Time utilization = recorded productive hours ÷ available hours × 100. This does not measure vehicle load utilization. Weight, volume, and pallet utilization need their own capacities and units. Check the per-driver distribution: a fleet average can hide an overloaded route. ### 6. Travel time share Travel time share = recorded driving hours ÷ recorded paid hours × 100. Use the same period for both values. A lower share can result from shorter trips, more customer service, longer waiting, or a different job mix. Read it alongside completed work and service outcomes before calling it an improvement. ## Assignment and recovery metrics ### 7. Initial dispatch response time Average response time = total elapsed time from request receipt to first meaningful dispatch response ÷ requests receiving that response. Define the response event. An automated receipt acknowledgment may not mean anyone has assessed the work. State whether the clock uses elapsed or staffed hours, segment urgent requests from advance bookings, and report unanswered requests separately. ### 8. Time to confirmed assignment Average assignment time = total elapsed time from an order becoming ready to plan to its first confirmed assignment ÷ orders assigned. Keep order creation and readiness timestamps separate so missing-data delays remain visible. Pair the average with the count and age of still-unassigned orders at the review cutoff, and record why each one is blocked. The [unassigned delivery orders guide](https://www.upperinc.com/blog/unassigned-delivery-orders/) covers the common blocking reasons. For a larger sample, a median and an upper percentile can expose a slow tail hidden by the average. ### 9. Verified hard-rule match rate Hard rules are the requirements an assignment must satisfy, such as required skills, equipment, territory, working hours, and vehicle capacity. [Automatic driver assignment](https://www.upperinc.com/blog/what-is-automatic-driver-assignment/) applies these rules before proposing a match, but the match still needs to be verified. Verified match rate = reviewed assignments with every required check verified as passed ÷ all reviewed assignments × 100. Unknown checks are not passes. Report them separately from known failures. State whether the review covers every assignment or a sample, and retain the relevant rule version. This is an operating quality check, not a certification that the system enforces every legal or physical requirement. ### 10. Changes after dispatch Changed-order share = distinct released orders reassigned or resequenced after dispatch ÷ all orders released in the period × 100. Count an order once in the numerator even if it moves repeatedly. Separately count total change events and their reasons to expose repeated rework. A high change share may reflect successful recovery, preventable planning errors, or volatile demand; the number alone cannot distinguish them. ### 11. Route finish variance For each route, calculate actual finish time minus the finish time in the plan released at dispatch, in minutes. Positive values mean the route finished later than that plan. Review the median, range, and individual outliers across comparable routes. Retain approved revisions separately to see whether recovery helped; the [route recovery guide](https://www.upperinc.com/blog/prevent-missed-delivery-windows/) explains how to record a revised plan without losing the original. Raw clock-time differences between drivers are misleading when shifts start at different times. Finish variance also does not establish that workloads were fair or caused staff turnover. ## A worked example of misleading denominators This illustrative daily dataset contains 100 eligible orders due today. It is not customer data or an Upper benchmark. Assume there were no excluded cancellations or approved changes to the due-date population. - 85 orders were delivered on time, including 82 delivered in full. - 10 orders were delivered late. - 3 orders were attempted but not delivered. - 2 orders were never attempted and remain unassigned. The groups total 100 orders. OTD is 85 ÷ 100 = 85%. OTIF is 82 ÷ 100 = 82%. Dividing the 85 on-time orders by only the 95 completed orders produces about 89.5%, but that is a different, completion-only measure that excludes the 5 undelivered orders. | **Measure** | **Numerator** | **Denominator** | **Result** | |---|---|---|---| | OTD (orders due) | 85 | 100 | 85.0% | | OTIF (orders due) | 82 | 100 | 82.0% | | On-time share of completed deliveries | 85 | 95 | 89.5% | | First-attempt success | 90 | 98 | 91.8% | | Cost per completed delivery | $1,900 | 95 | $20.00 | Suppose 98 orders received a first attempt and 90 completed on that first attempt. First-attempt success is 90 ÷ 98 = 91.8%, rounded to one decimal place. The other 5 completed orders required later attempts. The 2 unattempted orders still need a separate line. If delivery operating costs were an illustrative $1,900, cost per completed delivery would be $1,900 ÷ 95 = $20. The denominator here is intentionally completed deliveries; it should not be reused automatically for OTD. ## Choose dispatch KPIs that lead to a decision A small team does not need all 11 on day one. Start with OTD, first-attempt success, cost per completed delivery, and time to confirmed assignment, then add the rest as a specific problem appears. Match the problem you see to the measures to review together, then narrow them to your actual failure pattern: | **Current problem** | **Measures to review together** | **Decision to investigate** | |---|---|---| | Customers receive late or partial orders | OTD, OTIF, unattempted count | Receiving commitments and fulfillment readiness | | Dispatch is slow to place new work | Assignment time, open-order age, blocking reasons | Data cleanup, available drivers, approval workload | | Routes repeatedly finish late | Finish variance, changes after dispatch, time utilization | Service estimates, route changes, workload distribution | | Cost rises while output stays flat | Cost per delivery, failed attempts, travel share | Cost components and preventable repeat work | Compare similar work, dates, and operating conditions. A quieter period or shorter-distance customer mix can change results without any improvement in dispatch quality. Keep a dated change log before attributing an outcome to a new process. For the wider dispatching workflow these metrics sit inside, see the [fleet dispatching guide](https://www.upperinc.com/blog/fleet-dispatching/). ## What you can inspect in Upper Upper’s help center article on [available dashboard metrics](https://support.upperinc.com/docs/crew/analytics/available-dashboard-metrics) lists five KPI cards: Total Routes Completed, Total Stops Completed, On-Time Delivery Rate, Missed/Skipped Stops, and Average Stops per Route. The Driver Performance section adds Routes Completed by Driver, On-time Rate by Driver, Total Stops Completed by Driver, and Average Service Time per Stop. Those documented measures do not establish that all 11 formulas in this guide are native reports. The On-Time Delivery Rate card is defined as a percentage of on-time deliveries, and the article does not state its denominator, so reconcile it with your scorecard before comparing the two. Stop-level arrival and completion times in the Route Plan Detailed Report, and estimated versus actual route duration in the Route Plan Summary Report, give you the event data for route finish variance. The Driver Timesheet Report supplies shift hours for time utilization where shift tracking is enabled. Assignment-timing measures need exported event data. The [route management analytics page](https://upperinc.com/features/route-management-analytics) is a starting point for reviewing available reporting. Upper’s assignment recommendations show the reasoning behind each match, including skills, proximity, availability, and workload, and a confidence score. Upper describes the confidence score as a measure of match quality, not a probability of on-time delivery. It is not an observed delivery outcome. Evaluate assignments against verified requirements and actual results. ## Build a scorecard your team can explain Pick the decisions you need to improve, define their populations and clocks, and review exceptions alongside averages. A useful KPI lets the next dispatcher reproduce the number and identify the next action. Bring your current report and a few disputed orders to check how Upper’s records fit your measurement needs. [Book an Upper demo](https://calendly.com/upper/demo). ![](https://www.upperinc.com/wp-content/uploads/2026/05/txwfp9rgjemor38un3.svg)See it in action #### Build a dispatch scorecard your team can explain Bring your current report and a few disputed orders to check how Upper's records fit your measurement needs. [Book a Demo →](javascript:void(0)) ![Build a dispatch scorecard your team can explain](https://www.upperinc.com/wp-content/uploads/2026/05/svgviewer-output-1.svg) ## Frequently asked questions Dispatch KPIs are measures of how promptly work is assigned, how reliably it is delivered, and what it costs to complete. Common examples are on-time delivery rate, on-time in-full rate, first-attempt delivery success, cost per completed delivery, and time to confirmed assignment. Each needs a stated population, period, and timestamp to be comparable. Set the target from your commitments and a measured baseline for comparable work. There is no universal percentage that fits every delivery operation, and a target is meaningless without its denominator. In this guide’s due-order definition, an eligible order due but not delivered on time remains in the denominator. If another report uses completed deliveries only, label the difference and show the undelivered population separately. OTD counts an order as successful if it arrived within the agreed timing requirement. OTIF counts it only if it also arrived with the full ordered quantity. The same order population is used for both, so OTIF can never be higher than OTD. No. A justified change can improve customer outcomes. Compare execution with the version of the plan in force, retain the original baseline, and inspect why the route changed. --- _View the original post at: [https://www.upperinc.com/blog/dispatch-kpis/](https://www.upperinc.com/blog/dispatch-kpis/)_ _Served as markdown by [Third Audience](https://github.com/third-audience) v3.6.1.1_ _Generated: 2026-10-13 11:00:24 UTC_