Reimbursement processing time

Reimbursement processing time is the average number of days between an employee submitting an expense claim and the money landing in their account. If a claim is submitted on the first of the month and paid on the eighth, its processing time is 7 days. Averaged across every claim in a period, it tells you how long, in practice, your people wait to get their own money back.

The metric matters for two reasons that pull in the same direction. It is a clean measure of how efficiently the finance process runs, because a slow time almost always means manual approval chains, batch payment runs, or claims sitting in someone inbox. It is also a measure of employee experience, because being out of pocket for weeks is a real cost to an individual that quietly erodes trust and pushes people to avoid the expense process altogether.

Use the median rather than the mean as your headline number. A handful of disputed or queried claims can take weeks and drag the average far above what most people actually experience. The median tells you the typical wait, which is the number employees feel and the one worth managing to.

The calculation divides the total elapsed days across all reimbursements by the number of claims paid in the period. The value of the metric comes less from the headline average and more from breaking the elapsed time into its stages, because that is what tells you where the delay actually lives. The inputs below are what you need to settle first.

1. 1

   Total days from submission to payment

   The sum of elapsed calendar days for every claim, counted from the date the employee filed it to the date the funds were released. Decide whether to use calendar or business days and hold it consistent.

2. 2

   Number of claims paid

   The count of expense claims actually reimbursed during the period. Use claims paid rather than claims submitted, so the denominator matches the claims whose elapsed time you summed.

3. 3

   Stage timestamps

   The dates a claim moves between submission, approval, finance review, and payment. Without these you have an average but no diagnosis. They are what let you split the total wait into the stages that cause it.

4. 4

   Measurement window

   The period over which you average, typically a month or quarter. A consistent window keeps the trend comparable and stops a single large batch run from distorting the picture.

A worked example. In one month 60 claims are reimbursed with a combined 540 elapsed days, giving an average processing time of 9 days. Splitting by stage shows the claims spent on average 5 days waiting for manager approval, 1 day in finance review, and 3 days waiting for the next payment run. The headline of 9 days is useful, but the stage split is what tells you the fix is faster approvals and more frequent payment runs, not more finance headcount.

A metric tree decomposes processing time into the stages a claim passes through, then traces each stage down to the specific cause of delay. This turns a single average into a map of exactly where claims are waiting.

The first level splits the total time into the lifecycle of a claim: time to approval, time in finance review, and time to payment. Each stage decomposes further. Time to approval depends on whether the right approver is clear, whether they have a deadline, and whether the claim was complete when submitted. Time in review depends on receipt and policy checks. Time to payment depends on how often payment runs happen. When the average rises, the tree tells you whether claims are stuck waiting for approvers, failing policy checks, or simply waiting for the next batch payment.

This is the gap between a dashboard and a decision. A dashboard says reimbursements take 9 days. The tree shows that 5 of those days are claims sitting unapproved in managers inboxes, which is an approval-routing problem with a clear fix, not a finance capacity problem.

Processing time benchmarks vary with company size and how automated the expense process is, but the bands below give a practical sense of where an organisation sits. They assume the median wait from submission to payment, measured in business days, which is the number employees actually experience.

Watch the trend and the stage split, not just the level. A time creeping from 5 days to 9 over a quarter signals a process degrading, often because approvers are slipping. The benchmark says whether you are broadly competitive, the decomposition says which stage to fix to move the number.

Improving processing time means attacking the stage that owns most of the wait, which the metric tree identifies, rather than asking finance to work faster. Each stage has a concrete fix.

The decomposition decides the intervention. If the wait sits in approval, a deadline and escalation beat anything else. If it sits in payment, more frequent runs are the fix. Adding finance headcount to solve a problem that lives in slow approvers spends money on the wrong stage.

KPI Tree lets you model this by connecting processing time to the stages and owners behind it. Each stage of the claim, approval, review, and payment, carries RACI ownership, so a slow approval stage has a named accountable owner rather than being nobody fault. When the metric drifts past a threshold, it pushes to the owner of the stage that is dragging, so the slowdown is surfaced and acted on rather than discovered at the end of the quarter. The verified impact loop then checks whether a change, such as adding a payment run, actually moved the time, so you keep the fixes that work.

1. 1

   Starting the clock at approval

   Measuring from approval to payment hides the longest wait, which is the time a claim sits before anyone approves it. Always measure from submission, because that is the wait the employee feels.

2. 2

   Reporting the mean, not the median

   A few disputed claims taking weeks pull the average far above the typical experience. The median is the number most employees actually live, and the one to manage to.

3. 3

   Tracking the total without the stages

   An average with no stage split tells you the process is slow but not where. Without timestamps for each stage, you cannot tell whether to fix approvals, review, or payment.

4. 4

   Excluding queried claims entirely

   Dropping disputed claims from the count makes the number look better while ignoring the worst experiences. Keep them in, but report the median so they do not distort the headline.

5. 5

   Treating it as a finance-only metric

   Most of the delay lives in approvals owned by line managers, not finance. Holding only finance accountable for a number that managers largely control leaves the real bottleneck untouched.