Average resolution time, its formula and benchmarks

To work it out, add up how long each solved ticket took, from the customer’s first message to the fix, and divide by the number of tickets solved. If the 20 tickets you solved last week took 240 hours between them, your average resolution time was 12 hours.

Helpdesks, the apps where each customer email becomes a ticket, name it differently. Zendesk keeps a first and a full resolution time, while Intercom and HubSpot call it time to close. Jira Service Management sets it up as time to resolution, a deadline for each ticket. In IT it’s MTTR, mean time to resolve, though Atlassian notes that the R can also stand for repair, recovery or respond.

For a founder answering support alone, it’s how long customers wait on you. Deacon is an AI customer support agent that answers your customers from your own docs in seconds, at any hour, so the questions it answers never wait on you at all. It has a free plan.

How to calculate average resolution time

Every tool agrees on the division. They disagree on when the clock stops, whether it pauses while you wait on the customer, and which hours count. One ticket shows how much that matters.

A customer of a scheduling app emails at 10:00 on Monday because their calendar has stopped syncing. At 12:00 you ask which calendar they use and set the ticket to pending, meaning you’re waiting on them. They answer at 14:00 on Tuesday, and you fix it and mark it solved at 15:00. At 9:00 on Wednesday they write back that it has stopped again, and you solve it for good at 11:00.

The 49 hours hold 26 hours of waiting on the customer and 18 hours when the ticket sat solved before it broke again. Zendesk’s guide to support metrics warns that total resolution time includes stages when nobody is working on the ticket, such as a pending status, and its recipe subtracts pending and on-hold time. Even then, the 18 solved hours stay in.

Office hours make the number smaller than the customer’s wait. Help Scout’s article on time to resolution reminds you that “a customer still has to wait the full length of time!” Gorgias counts every hour, nights and weekends included, while Help Scout’s email report averages from a conversation being created to it being resolved and can switch to office hours.

Reopened tickets split the tools too. Freshdesk’s report leaves out a ticket that was resolved in the period and reopened later. Gorgias counts to the last close, however many times the ticket was closed before.

Why the median often beats the average

One old ticket can swamp a month. Say eight tickets took 1, 2, 2, 3, 4, 5, 6 and 9 hours, and a ninth waited three weeks on a bug fix. The average is almost 60 hours. The median, the middle ticket once they’re sorted, is 4 hours, and it would still be 4 hours if the ninth had taken a year.

Gorgias shows the median “to minimize the impact of extreme values”, and Zendesk’s recipe suggests it for the same reason. Intercom’s reporting help explains how to switch a chart from the median to the average. Freshdesk and Help Scout show the average. Check which one a figure is before you compare it with anyone else’s, including the benchmarks below.

What is a good average resolution time?

For tickets, the best measured figure is a median of about 6 hours, and for chat it’s about 10 minutes. The glossaries that rank for this search disagree by a factor of 200 or more, and neither cites anything. Medallia’s says the general benchmark is around five to seven minutes, and Plecto’s says the cross-industry average appears to be 24 hours. Zendesk’s 2012 benchmark left full resolution time out altogether, because it “was all over the map”.

These are the published figures we found with a named sample, read on 23 September 2026. Both reports come from Freshworks measuring its own customers, and no study from outside the helpdesk vendors turned up.

The two reports can’t be set side by side, since the 2022 figures count business hours only and the 2025 report doesn’t say which hours it counts. If you run a software company, your sector was the slowest of any in 2022.

What a chatbot does to resolution time

Adding a chatbot may not move your helpdesk’s resolution time at all, because the questions it settles either never become tickets or are left out of the figure. In Freshdesk’s 2019 benchmark, companies using bots had the same 25-hour average resolution time as companies without them, though their first response time was 6 hours shorter.

Intercom’s time to close counts only conversations a teammate closed, so a conversation Fin closes on its own isn’t in it. A second figure, time to close including automation, counts a close by anyone, Fin, a bot or a workflow included.

Counted on its own, chat with a bot is quick. Freshdesk’s 2023 report put resolution time for simpler questions over chat with chatbots at 10 to 27 minutes, across around 3,000 Freshchat accounts, against 32 hours for Freshdesk tickets. So judge a chatbot by how many questions stop reaching you, not by your ticket resolution time.

Average resolution time vs MTTR, first response time and handle time

Each of these is a length of time, so they get mixed up, but each measures a different stretch. Here’s the calendar ticket again.

MetricNet’s column on MTTR counts business hours, so an incident reported at 16:00 on a Friday and fixed at 16:00 on Monday took 8 business hours, not 72 clock hours. Atlassian’s version also counts the work to make sure the failure won’t happen again.

The resolution rate counts something else again, the share of problems that get solved at all.

Cutting resolution time with Deacon

The quickest resolution is an answer the customer doesn’t have to wait for. Deacon answers from your own docs in seconds, at any hour and in the customer’s language, so a question asked at 2 in the morning doesn’t sit until you wake up. When your docs don’t cover it, Deacon says it doesn’t know instead of guessing, and asks for the visitor’s email.

Those questions reach you without a helpdesk. When a visitor leaves their email or asks for a person, the owners on your team get an email about it. You reply from the Deacon dashboard. If the visitor still has the chat open, your reply appears there under your name, and if they’ve left, it goes to the address they gave in the chat. Analytics › Widget counts everyone who left their email and is still waiting for your reply, however long ago they asked.

A question that waited on you once needn’t wait again. Every question your docs didn’t cover has an Answer this button. Write the answer once, press Save & check, and Deacon shows you the reply it now gives and how many other open questions your answer covers. Your answer becomes part of what Deacon knows, so the next person who asks can get it in seconds instead of waiting for you, and the more you answer, the less Deacon needs you.

It also shows you what’s behind the slow ones. Topics group similar questions, count how many people asked and label most groups as a bug report, a how-to, a feature ask and so on. A bug report that keeps coming back is a fix to ship, and a feature ask tells you what customers want you to build next.

Deacon doesn’t mark conversations solved, so it has no resolution time to give you. Human handoff shows how a conversation reaches you and goes back to Deacon, and you can try it all on the free plan, which needs no card.