Resolved is not the same as closed
Plenty of tools auto-close conversations with no activity for a few days.
Those closures land in the average and improve it, even though the customer left without an answer. If your resolution time drops in the same month your reopen rate climbs, this is almost certainly why.
Why it matters
It is the metric that best reflects what each case costs.
A high resolution time almost never comes from slow writing; it comes from waits, handoffs between people, or missing information the customer had to supply twice. Every one of those stretches is time somebody pays for.
How it is calculated
The mean and median time between opening and resolution.
It is worth segmenting by type. Mixing "I forgot my password" with "your webhook has been failing for two days" produces a number that decides nothing.
Subtracting the wait on the customer
If you asked for a screenshot and it took two days to arrive, those two days say nothing about your team.
Almost every tool can pause the clock while the ball is in the other court, and without that pause the number is useless for management.
Where the time hides
In three places, nearly always.
Looking up information before replying, the handoff to another person who starts from zero, and the back-and-forth to obtain a fact that could have been looked up. All three shrink with access to the data, not with more people.
The formula, with an example
The sum of the time between opening and resolution of each conversation, divided by the conversations resolved in the period.
The figures in the example are illustrative.
- Example. 150 resolved conversations and 1,200 hours in total give a mean of 8 hours.
- With the wait subtracted. If 450 of those hours were spent waiting on the customer, the time that depends on the team is 5 hours.
