T&E spend is the total cost of employee travel and business expenses, covering airfare, lodging, ground transport, meals, conference fees, entertainment, and mileage. No standard governs the category edges, so four finance leaders will produce four different totals. Four metrics measure a program honestly, and they are worth only as much as your audit coverage.
Key takeaways
- Category definition comes first. One company counts meetings and events, unfiled corporate card spend, and employee software purchases inside T&E, and the next company leaves them out.
- Booked, committed, and reimbursed spend answer different questions. Only reimbursed spend reconciles to the general ledger, and unused tickets and uncredited refunds account for the gap between booked and reimbursed.
- Each of the four metrics improves in isolation for a reason finance would not welcome. Cycle time falls when you auto-approve more reports, and leakage falls when reviewers look less.
- Track your own trend lines over time and you learn something real. Cross-industry averages assembled from other companies' definitions teach you nothing.
Travel and expense (T&E) spend also covers the small out-of-pocket purchases employees make while working. That variation in the edges is why cross-company benchmarks rarely compare like with like, and why a program can look efficient on one measure while losing money on another.
Where the T&E spend category boundaries fall
Three boundaries account for most of the disagreement. Meetings and events are often contracted by procurement or marketing, so one company books a $400,000 user conference outside T&E and the next books it inside. Corporate card spend that never enters the expense system is the second gap, since corporate liability charges post whether or not anyone files a report. The third is employee purchases that behave like operating expense, including software subscriptions and mobile plans.
None of those choices is wrong in itself. They become a problem when someone compares a number built one way with a number built another way. Record which general ledger accounts, card programs, and vendor categories your T&E figure includes, and do it before any benchmarking exercise. The general ledger is the master record of accounts an accounting system posts to. Keep that written definition stable, because redefining the category changes a trend line more than most policy changes do.
Booked, committed, and reimbursed spend answer different questions
Every T&E spend figure depends on where in the cycle you take the measurement.
Booked spend is what the booking tool or travel agency recorded at purchase. It is the earliest signal available and travel managers use it in supplier negotiations. It also overstates final cost, because tickets get cancelled and refunds arrive later.
Committed spend is what the company owes and has not yet paid. Approved travel requests, non-refundable tickets, and posted card transactions awaiting a report all belong here. It tells a controller what is coming, and most teams do not track it at all.
Reimbursed spend is what has left the company, through payroll reimbursement or card settlement. It is the only figure that reconciles to the general ledger, and it lags the trip by weeks.
The gap between booked and reimbursed spend deserves review in its own right. Unused tickets, duplicate bookings, and uncredited refunds all account for part of that difference.
The four T&E spend metrics worth tracking together
Four measures tell you most of what matters. Each is set out by what it measures, how to calculate it, and how it misleads.
Cost per expense report
What it measures. This metric is the fully loaded internal cost of processing one report, covering submitter time, approver time, audit review, and system cost.
How to calculate it. Add the labor hours spent on submission, approval, audit, and correction over a period. Multiply by loaded hourly rates, add allocated system cost, then divide by the number of reports processed.
The reference point. GBTA research from 2015 put the processing cost of an expense report for a single night hotel stay at $58 and 20 minutes. The same research found that 19 percent of reports contain errors or missing information, adding $52 and 18 more minutes each to correct. That research is now more than a decade old and was never inflation-adjusted, so treat it as directional. It is a structural finding about where effort accumulates rather than a current price.
How it fails. The metric rewards fewer, larger reports. Batch a quarter of travel into one submission and the cost per report improves while every downstream review gets harder. It also ignores what the process caught, and a cheap process that reimburses bad claims is not efficient.
Leakage as a share of T&E spend
What it measures. Leakage is money paid out that policy or supporting evidence should have stopped. That includes duplicate submissions, out-of-policy spend reimbursed anyway, inflated claims, and unrecovered value-added tax (VAT), a consumption tax charged in many countries and often reclaimable on business travel.
How to calculate it. Divide the total value of confirmed recoverable and preventable findings by total T&E spend for the same period. Count only validated findings, not every flag raised.
The reference point. Our platform runs duplicate and unauthorized spend checks against every transaction rather than a sample. The recoverable total then comes to 2 to 5 percent of expense spend, a figure our analysis of 100 percent audit coverage sets out.
How it fails. Leakage measured by a sample-based program partly measures how hard the reviewers looked. Most enterprises review 10 to 20 percent of expense transactions, and finding little in that slice says nothing about the remainder. Detection also matures over time, so early numbers understate the real rate. Duplicate findings alone grow about 700 percent from month one to month twelve as the checks accumulate history.
Policy compliance rate
What it measures. This rate is the share of transactions or reports that conform to policy without an exception.
How to calculate it. Divide compliant transactions by total transactions reviewed. Break the result down by policy rule, department, and region. The aggregate figure says little on its own, and the breakdown is where decisions get made.
How it fails. A compliance rate measures the rule set as much as it measures behavior. Loosely written policies produce high compliance alongside high spend. Rules the system never checks, such as receipt authenticity or attendee lists, count as compliant by default. A rising compliance rate paired with flat cost per trip usually means the policy stopped asking hard questions.
Cycle time from submission to reimbursement
What it measures. Cycle time counts the elapsed days from report submission to payment, split into time in approval, time in audit, and time in payment.
How to calculate it. Take the median rather than the mean of days elapsed per stage. Report the 90th percentile alongside it, because the tail is what employees complain about and what makes employees work around policy.
How it fails. Cycle time is the easiest of the four to improve for the wrong reason. Auto-approve more reports and it shortens immediately. Pair it with the leakage rate and it stays honest. Read it alone and it rewards a team for looking less.
Audit coverage underpins all four. The share of transactions reviewed against evidence before payment determines what the other numbers are worth, so it belongs in the same report, with a statement of what each review pass checks in a transaction.
Why cross-company benchmarks mislead
Cross-company comparisons of T&E spend mislead more often than they inform.
The category definition differs, as described above. The card program model differs too, and it determines whether spend appears as a card transaction, an employee reimbursement, or both. Trip mix matters more than policy, because a company running four-day client trips will never match one running one-day site visits. Per diem programs pay a fixed daily allowance, actual-expense programs pay what was spent, and the two produce different distributions from identical behavior. Headcount denominators break as well, since spend per employee tracks the ratio of travelers to non-travelers rather than efficiency.
Few benchmarks capture time to detection at all. The Association of Certified Fraud Examiners (ACFE) studied 2,402 cases for Occupational Fraud 2026 and reported a median scheme length of 12 months before detection. A program that measures only annual totals has no view of that lag.
The workable alternative is internal benchmarking. Hold the category definition constant, track the four metrics by quarter, and segment them by department, region, and trip type.
How we approach measurement
We treat measurement and enforcement as the same problem. Our AI reads every line of every receipt on every report before reimbursement, and audits 100 percent of card transactions as purchases post. With complete coverage, leakage becomes a counted figure rather than an estimate, because the denominator is all spend rather than a sample.
The metrics change shape once coverage is complete. Cost per report falls as routine reports auto-approve and reviewers spend their hours on genuine exceptions. Cycle time and leakage stop trading against each other, since faster approval no longer means less scrutiny. AI Analytics turns findings into prescriptive insight by policy rule and department, and Team Insights puts the same view in front of managers. Customers running AI expense audit reach automation rates above 80 percent and reduce finance operating costs by up to 50 percent.
The bottom line
Write down your category definition first, because every later comparison depends on it. Then read the four metrics as a set, since each improves in isolation for reasons a finance team would not welcome. Place your validated leakage rate beside the share of transactions reviewed before payment, and decide whether the first number is small because the spend is clean or because the review was narrow.
Frequently asked questions
What is T&E spend?
T&E spend is the total travel and expense cost a company incurs for employee business activity, including airfare, lodging, ground transport, meals, conference fees, client entertainment, and mileage. Where meetings, events, and employee software purchases sit varies by company, so any T&E figure needs a stated category definition.
How much does it cost to process an expense report?
GBTA research from 2015 put the average at $58 and 20 minutes for a report covering a single night hotel stay. The same research found that 19 percent of reports contain errors, costing another $52 and 18 minutes to correct. Those figures are more than a decade old and not inflation-adjusted, so they are directional evidence about cost structure rather than a current price.
What is expense leakage?
Expense leakage is spend that policy or supporting evidence should have prevented and that was reimbursed anyway, including duplicate claims, out-of-policy purchases, and personal charges on corporate cards. Our platform data across enterprise programs puts recoverable leakage at 2 to 5 percent of total expense spend, a vendor-reported range rather than an independent measurement.
Why are cross-company T&E spend benchmarks unreliable?
Companies define the category differently, run different card program models, and have different trip profiles, so identical behavior produces different numbers. Per diem policies, actual-expense policies, and the ratio of travelers to non-travelers all change the figures independently of program performance.
Which T&E spend metrics should finance teams track together?
Leakage rate, policy compliance rate, cost per expense report, and submission-to-reimbursement cycle time work as one set, with audit coverage reported alongside them. Each improves in isolation for the wrong reason, so reading cycle time against leakage keeps the picture honest.