A support report says replies are faster this week. The inbox still contains old enquiries, and a customer has contacted the team twice about the same unresolved issue. The report may be correct, but it is answering only part of the question.
A useful support dashboard connects speed, unresolved work and the quality of the answer. Before setting targets, agree what each measure includes, which clock it uses and what decision the result should help a manager make.
Write a definition before choosing a target
Record the event that starts the timer, the event that stops it, the channels included and any exclusions. Check your helpdesk’s documentation rather than assuming that similarly named measures behave identically.
For example, Zendesk defines first reply time around ticket creation and the first public agent response, with additional channel-specific details. That is different from a self-defined measure of when a customer first received any acknowledgement. An automated receipt and a useful human answer should not be treated as interchangeable simply because both are messages.
Be explicit about the clock
Calendar time and scheduled business time answer different questions. Calendar time reflects elapsed waiting. A business-hours measure counts the time included by the configured working schedule. Zendesk’s schedule documentation explains how its reports use business hours and holidays.
Here is a simplified example using a fictional Monday-to-Friday, 9 a.m. to 5 p.m. schedule with no intervening holiday. An enquiry arrives Friday at 4:45 p.m. and receives its first reply Monday at 9:15 a.m. The calendar interval is 64 hours and 30 minutes. The scheduled working interval is 30 minutes: 15 minutes on Friday and 15 on Monday. Label the measure so nobody mistakes one result for the other.
Use a small set of complementary measures
1. First-response time
This describes how long enquiries wait for the defined first response. Review it by channel and workload. A phone queue, live conversation and email inbox may have different expectations. A single combined average can hide where customers are actually waiting.
2. Age of unresolved work
Look at what is still open, including the oldest items and the reason they remain unresolved. Separate work waiting on the team from work awaiting a customer, supplier or approval. Suggested age bands should fit the operation; they are not universal industry benchmarks.
3. Resolution time and returning issues
Agree whether your report follows the first resolution or a later one after a case reopens. Review returning issues alongside completion speed. A quick closure is less informative if the customer immediately needs to restart the conversation.
4. Quality from a defined sample
Review a manageable sample of interactions for accuracy, completeness, appropriate tone, approved decisions and useful notes. Write the review criteria first. If you review only five interactions, say five; do not present a small sample as the quality score for every conversation handled.
5. Promised follow-ups completed
Track commitments with a due time and owner. A simple internal rate is completed-on-time commitments divided by commitments due in the period, multiplied by 100. Define how cancellations and revised deadlines are handled so the result cannot be improved merely by moving due dates.
Look beyond the average
Suppose four fictional enquiries receive replies after 2, 3, 4 and 51 minutes. Their average is 15 minutes, while the median is 3.5 minutes. Both calculations are correct. Neither, on its own, explains the enquiry that waited 51 minutes.
Show the number of observations and examine unusually long waits. Where a percentile is used, record the calculation method and keep it consistent. Avoid comparing tiny groups as if they provide stable evidence about individual performance.
Turn the report into a practical review
Start a review with three questions: What changed? Which cases explain the change? What action is within our control? The answer may involve a missing instruction, an approval delay, a surge in one enquiry type or an unrealistic mix of responsibilities.
Choose one improvement and an owner, then review comparable work in the next period. If the issue is inconsistent handling, write or improve the procedure for that task. Keep the reporting definitions available so changes to a dashboard do not silently change the meaning of its results.
The aim is a report that helps someone make a better operational decision. A small set of clearly defined measures, supported by actual case review, is a practical place to begin.
About the publisher
Written by Corporate Hours. Find more practical business operations resources at corporatehour.com. Questions or feedback: info@corporatehour.com.
Original article: https://corporatehour.com/resources/customer-support-metrics/
