Skip to content

Guides

First response time: how to measure it, set a target and improve it

First response time is how long a customer waits between writing in and hearing from a person. It is the number customers feel most, and the easiest one to measure badly.

Last reviewed

What counts as a first response

A first response is the first reply written for that customer by a person on your team, or by an AI that actually answered the question. An automatic “we got your message” email doesn't count, because it doesn't help the customer. Track it separately if you like, since it still tells them the message arrived.

Measure from the moment the message arrives to the moment the reply is sent, not from when an agent opened the ticket.

Count business hours, not clock hours

If you are open 9 to 5 on weekdays, an email that arrives on Friday at 4:55pm and is answered on Monday at 9:30am took 35 business minutes, not 64 hours. Counting clock hours makes every weekend look like a failure, so teams stop trusting the number.

Use the SLA due time calculator to work out when a ticket is due given your hours, and use the same hours for reporting.

Use the median and a high percentile, not the average

A few tickets that waited two days drag an average far above what most customers experience. Report the median (half were answered faster) and a high percentile such as the 90th (nine in ten were answered within this time). The first tells you what is typical; the second tells you how bad the worst experiences are.

Set a target you can meet almost every time

Look at your last month first. If the median is 3 business hours and the 90th percentile is 9, a 4 hour target will be missed one ticket in five or more, and the team will learn to ignore it. Pick a target you meet on nearly every ticket, publish it, and tighten it once you are comfortably inside it.

Set different targets by channel and priority: customers expect chat answered in minutes and email in hours, and an urgent problem faster than a question. The ticket priority matrix gives a starting point for each level.

What actually brings it down

Most gains come from removing waiting time rather than typing faster.

  • Make sure someone sees every new ticket. Unassigned tickets that nobody owns wait longest. A shared queue with an owner for the morning check fixes more than any tool.
  • Answer the repeat questions with saved replies. If the same answer is sent every week, write it once as a macro.
  • Put the answers customers need in a help center, and link to it. Questions that are answered before the ticket is written never wait at all.
  • Let an AI answer the routine ones. Tickets it finishes have a first response measured in seconds, and your team sees fewer.
  • Acknowledge after hours. A reply that says when you will answer does not stop the clock on your target, but it stops the customer writing a second email.

Do not optimize it alone

A fast but wrong first reply creates a second ticket. Read first response time next to the share of tickets reopened, CSAT, and how long tickets take to resolve. If the first falls while the others rise, the team is replying quickly to get a number, not to help.

Questions

What is a good first response time?

It depends on the channel and what customers expect from you. A target you can meet on almost every ticket is better than an ambitious one you miss. Check your current median and 90th percentile, then set the target just outside them and tighten it over time.

Does an auto-reply count as a first response?

Usually not. Most teams count only a reply that a person (or an AI answering the question) wrote. An acknowledgement email is worth sending, but it is a separate measure.

Should weekends count?

Not if you are closed. Count business hours and skip holidays, so a ticket that arrives Friday evening is not late on Saturday morning.

Written by Flatdesk, a help desk for small support teams at one flat price per agent. Try it free for 14 days, no card.