Corporate Ground Transportation KPIs: Travel Manager Metrics Complete Guide

A useful KPI is a reproducible question, not a polished dashboard label. “On time” has no stable meaning until the program defines which trips qualify, which timestamp counts, how early or late is treated, and what happens when data is absent.

Use this guide to build a measurement dictionary for the corporate travel program. For supplier requirements and evidence requests, use the ground transportation RFP guide. The booking confirmation remains the source for actual service terms and any reporting deliverable.

Start with a trip-level event record

Assign every booked movement a stable trip ID. Keep planned, observed, and financial fields separate. Do not overwrite the original scheduled pickup after a change; retain the version, change time, requester, approver, and revised value.

Write a measurement contract for every KPI

For each metric, record its business question, numerator, denominator, inclusion and exclusion rules, timestamp definitions, source fields, reporting period, owner, refresh date, target source, and missing-data treatment. Freeze those rules for the reporting period. If a definition changes, show the break in the trend rather than silently restating history.

Report numerator and denominator beside the percentage. A 100% result based on two trips should not look equivalent to a 99% result based on hundreds. Show excluded and unknown counts as well.

Core service metrics and defensible denominators

Vehicle-on-location punctuality

Formula: eligible pickups where vehicle-on-location meets the written threshold ÷ eligible pickups with valid scheduled and vehicle-on-location timestamps.

Define the pickup boundary: geofence, venue loading area, terminal holding area, or exact curb. A vehicle staged elsewhere may be ready but is not necessarily at the passenger meeting point. Choose whether early arrival, a changed pickup, and authority-directed staging qualify before the month begins.

Passenger connection rate

Formula: eligible pickups with a confirmed passenger connection ÷ pickups where a connection was expected.

Use a defined event such as direct contact, greeter handoff, or passenger boarding. Do not infer connection from flight landing or vehicle arrival. Track time from passenger-ready to connection separately when both timestamps are reliable.

Trip completion rate

Formula: completed eligible trips ÷ trips expected to operate after the program's cancellation cutoff.

Keep timely traveler cancellations, late traveler cancellations, provider cancellations, duplicates, test records, no-passenger events, no-vehicle events, and unresolved trips in separate buckets. Separate status reporting does not automatically remove a trip from the denominator: provider cancellations after the eligibility cutoff remain eligible in this method. Completion alone does not show whether a trip was punctual or met the requested handoff.

Exception rate

Formula: eligible trips with at least one defined exception ÷ eligible operated or attempted trips.

Count affected trips for the rate, then also count individual incidents. One trip may contain several events. Preserve an “unknown or under review” reason so incomplete investigations do not become a convenient provider, traveler, venue, weather, or airline attribution.

Data completeness

Formula: populated required field opportunities ÷ all required field opportunities.

A required field opportunity is one field that should exist for one eligible trip. Fields that legitimately do not apply should be coded “not applicable,” not left blank. Audit timestamp order, time zone, duplicates, impossible durations, and status conflicts in addition to null values.

Cost and invoice metrics

Cost per completed trip

Formula: eligible invoiced spend ÷ completed trips in the matched population.

State whether spend includes tolls, parking, taxes, wait, overtime, cancellation charges, gratuity, credits, and foreign exchange. Compare like service types, markets, time periods, and booking conditions. Cost per mile is often misleading when airport access, hourly reservations, waiting, parking, tolls, or minimums drive the price.

Invoice accuracy

Trip-level formula: invoiced trips with no validated billing error ÷ invoiced trips reviewed. A line-level version can use accurate line items ÷ reviewed line items.

Publish which version is used. A duplicate trip, wrong quoted rate, unapproved change, missing credit, incorrect toll, or mismatched cancellation treatment can be coded separately. Disputes under review belong in a pending category until resolved.

Quote-to-invoice variance

Formula: invoiced amount minus the comparable confirmed quote, shown as an amount and, where the quote is nonzero, a percentage of the quote.

Reconcile approved itinerary changes before calling a variance an error. Keep expected additions, approved changes, disputed items, and confirmed errors separate.

Accessibility and traveler-experience measures

Accessible-request confirmation

Formula: trips with an accessibility request confirmed by the defined cutoff ÷ trips containing an accessibility request.

The record should describe the requested assistance, actual confirmed vehicle configuration, companions, mobility device and handoff needs with the traveler's consent. Measure fulfillment separately from confirmation. Airport or airline assistance and ground transportation are distinct handoffs.

Traveler feedback response and issue closure

A satisfaction average needs its scale, question, invitation count, response count, collection window, and handling of missing responses. Never treat nonresponse as satisfaction. For issue closure, measure cases resolved under the organization's written definition and show open-age groups alongside the rate.

Illustrative calculations

The figures below are examples for testing a workbook. They are not Detailed Drivers results, targets, industry benchmarks, or promises.

• Punctuality example: 96 pickups have valid required timestamps; 91 meet the chosen threshold. 91 ÷ 96 = 94.8%. Four missing-timestamp trips appear separately and do not disappear from the report.
• Completion example: 120 bookings include 10 timely traveler cancellations and two duplicates, leaving 108 trips expected to operate at the eligibility cutoff. Three of those trips are later canceled by the provider and stay in the denominator. If 102 trips are completed, the result is 102 ÷ 108 = 94.4%. Report reasons for all six uncompleted trips, including the three provider cancellations.
• Invoice example: 80 invoiced trips are reviewed and three contain validated billing errors. Trip-level invoice accuracy is 77 ÷ 80 = 96.3%. Pending disputes are shown separately until resolved.
• Completeness example: 50 eligible trips require eight fields, creating 400 field opportunities. If 388 are populated or validly marked not applicable, completeness is 388 ÷ 400 = 97.0%.

Segment carefully and protect the record

Break results down by market, service type, airport versus non-airport, planned versus changed movement, supplier, and booking lead-time band only when the sample supports a useful comparison. Small groups can expose traveler behavior or create unstable percentages. Suppress or combine them under the organization's privacy rules.

The Federal Trade Commission's business cybersecurity guidance recommends knowing what personal information is kept, retaining only what is needed, protecting it, disposing of it securely, and planning for incidents. Apply those principles to passenger names, phones, itineraries, accessibility details, and location timestamps. Limit access and publish aggregated reports where trip-level identity is unnecessary.

Set targets from evidence, not borrowed benchmarks

Start with an agreed definition and a clean baseline. Then set a target from business consequences, traveler needs, contractual commitments, market conditions, and sample size. Record who approved it and when it changes. A copied percentage can reward the wrong behavior when its denominator or threshold differs.

Use medians and percentiles for elapsed times because an average can hide a long tail. Publish the count, p50, p90, and maximum only where the data is reliable and privacy rules allow it. A percentile describes the observed distribution; it does not explain cause.

Build a monthly review that leads to action

1. Validate the population, duplicates, status codes, time zones, missing fields, and invoice period.
2. Publish counts, denominators, definitions, exclusions, unknowns, and prior-period comparability.
3. Review the largest exception categories using trip-level evidence.
4. Assign a corrective action, owner, due date, and verification method.
5. Close the action only when the next evidence check supports closure.

A dashboard is a decision aid, not proof of cause. Pair trends with incident records, itinerary changes, venue or authority instructions, flight or rail events, and traveler feedback before assigning responsibility.

Keep service policies in the source data

Detailed Drivers airport pickups include 45 minutes of complimentary waiting for domestic flights and 60 minutes for international flights, starting at actual landing, with flight tracking. Non-airport pickups include 15 minutes. Overtime is billed for the actual additional minutes at the applicable per-minute rate for your selected vehicle.

The standard hourly minimum is 3 hours; a longer event minimum applies only when booking-confirmed. Sedan and SUV bookings receive a full refund when cancelled at least 24 hours before the scheduled pickup time. Sprinter bookings receive a full refund when cancelled at least 72 hours before the scheduled pickup time. Cancellations inside the applicable notice period are charged in full. Quoted all-in pricing includes gratuity, so an additional tip is optional. Store the confirmed terms used for each trip so later measurement does not apply a different rule.

Review the airport transfer, hourly service, and rate information pages for service context. Submit the actual itinerary through the reservation request. Related articles are available in the blog.

Frequently asked questions