| Ticket ID | Requested At | First Response At | Channel | Priority |
|---|---|---|---|---|
| T-1001 | 2026-02-26 09:10 | 2026-02-26 09:34 | High | |
| T-1002 | 2026-02-26 10:05 | 2026-02-26 11:02 | Chat | Normal |
| T-1003 | 2026-02-26 14:18 | 2026-02-26 15:05 | Phone | Urgent |
| T-1004 | 2026-02-26 16:55 | 2026-02-27 09:20 | Low |
This example includes an overnight case to show how working-hours mode changes totals.
- Response Time (minutes) = minutes between requested_at and first_response_at.
- Working-hours adjusted response = sum of minutes within your workday window, optionally excluding weekends and holiday dates.
- Average = (sum of response minutes) ÷ (number of tickets).
- Median = middle value after sorting response minutes (average of two middles when even).
- Percentile (P) uses linear interpolation on sorted minutes.
- SLA Compliance = (tickets with response ≤ SLA threshold) ÷ (tickets total) × 100.
- Paste your ticket CSV into the dataset box. Include requested_at and first_response_at columns.
- Set your SLA threshold, then choose a percentile to monitor (for example, P90).
- Enable working-hours adjusted mode if your SLA is measured during business hours.
- Optionally exclude weekends and list holiday dates (one per line).
- Click Calculate to view results under the header, above the form.
- Use the download buttons in the results panel to export CSV or PDF.
First-response time is a leading indicator of workload balance. When the average rises while the median stays stable, a small set of tickets is absorbing capacity. Track the gap between median and P90 to quantify tail risk. For example, a median of 25 minutes with a P90 of 140 minutes suggests queues and handoffs are the main bottlenecks.
Percentiles convert noisy ticket streams into dependable targets. A P90 goal means 90% of tickets receive a first reply within the threshold. If you analyze 1,000 tickets per month, a P90 breach rate of 10% implies roughly 100 customers experience slow starts. Reduce that by adding on-call coverage during known spikes.
Calendar-time can overstate delays for after-hours requests. Working-hours adjustment normalizes performance when teams operate on fixed shifts. If a ticket arrives at 16:55 and receives a reply at 09:20 next day, calendar-time is 745 minutes, while a 09:00–17:00 workday yields 25 minutes. Use the same mode your SLA contract uses.
SLA compliance rate links process health to customer impact. If your SLA is 60 minutes and compliance is 78%, then 22% of tickets start late. Combine this with ticket volume to estimate staffing. For 200 daily tickets, 44 late starts per day may justify additional triage shifts or automation.
Segmenting by channel and priority reveals where response debt accumulates. Chat often shows lower medians but volatile peaks, while email can have stable medians and heavier tails. If urgent tickets average 18 minutes but normal tickets average 75 minutes, your routing rules may be starving standard work. Rebalance with dedicated queues and explicit time blocks.
Review metrics weekly with three questions: What changed in volume, what changed in tail latency, and what changed in SLA rate? Pick one operational experiment, such as earlier handoff to specialists or a 15-minute triage buffer. Compare last week’s P90 and compliance after at least 300 tickets to avoid small-sample noise.
What is “first response time” in this calculator?
It is the minutes between requested_at and first_response_at for each ticket, measured as calendar-time or business-time depending on your selected mode.
Why should I track P90 instead of only the average?
Averages hide slow outliers. P90 focuses on the customer experience during busy periods and highlights queueing and handoff delays.
When should I enable working-hours adjusted mode?
Use it when your team operates in defined shifts and your SLA excludes nights, weekends, or holidays. It creates comparable results across weekdays.
How do holidays affect the calculation?
In working-hours mode, listed holiday dates are treated like non-working days. Any minutes falling on those dates are excluded from response totals.
What does “SLA met” mean for each ticket?
The ticket is marked “Yes” when its response minutes are less than or equal to the SLA threshold you set in the form.
What CSV format should I use for dates and times?
Use ISO-like values such as 2026-02-26 09:10 or 2026-02-26T09:10. Keep requested_at earlier than first_response_at for every row.