Customer service response time is how long a customer waits for a reply, from the customer’s first message to the first reply that addresses it, autoresponders excluded. Survey-based standards put email at about one hour and social messages at 24 hours or sooner.
For an online store, the clock that matters most is the one on tickets with a deadline, such as a cancellation or address change before the parcel ships. If you also keep a set of reply templates, this guide shows which numbers to hold them to.
What is customer service response time?
Customer service response time is the time between a customer’s message and your team’s reply. It is usually reported as first response time, and sometimes as average reply time across every reply in the conversation. It describes how long the customer waits, not how long your team works on the problem.
Medallia defines average reply time as the measure of “the time it takes for a business or customer service team to respond to customer inquiries, messages, or requests”, with the formula Total Reply Time divided by Total Number of Replies.
Handle time is a different number. In Medallia’s words, handle time covers the response time plus the time spent actively working on the customer’s issue within the interaction, so it measures effort, not the wait alone.
An automatic “we got your message” does not stop the clock. HappyFox defines first response time as the time elapsed between receiving an issue and sending the first response, and says automatic responses do not count. For what an acknowledgement is for, see the acknowledgement template.
First response time vs resolution time: what is the difference?
First response time measures how quickly someone replies; resolution time measures how quickly the problem is gone. A fast first reply that only says “we are looking into it” improves the first number and does nothing for the second. Customers feel the second one.
| Metric | Clock starts | Clock stops | What it tells you | How it can mislead |
|---|---|---|---|---|
| First response time | Customer’s first message | First public reply that addresses it | How quickly the queue is picked up | A quick holding reply improves it without helping the customer |
| Next response time | Customer’s follow-up message | Your next reply | Whether conversations stall after the first reply | Hidden when only the first reply is tracked |
| Resolution time | Customer’s first message | The problem is solved | How long the customer lives with the issue | One ticket left over a weekend distorts an average |
| Average reply time | Each customer message | Each reply | The pace across the whole conversation | Blends quick replies with slow ones |
| Handle time | Customer’s message is received | The interaction ends | The response time plus the active work on the issue | Blends the wait with the effort, so it is not a measure of how long the customer waited |
Customers care about the outcome, not just a reply. In HubSpot’s survey, 67% of consumers expect their support ticket to be resolved within 3 hours. SuperOffice found that only 20% of companies answered questions in full on the first reply.
Support teams know response time matters: HubSpot reports CSAT and retention at 31% each as the top CX metrics, followed by response time at 29%.
What is a good response time for each channel?
Published standards put email at about one hour (15 minutes is world-class), Facebook messages at about one hour, X (Twitter) DMs at 15 minutes and social generally at 24 hours or sooner. Observed reality is slower, and many companies never reply to email at all.
| Channel | What customers expect (source) | What companies actually do (source) |
|---|---|---|
| ”Nearly a third of customers expect businesses to respond to emails in one hour or less”: US 27.7%, UK 33.4%, Canada 38.3%; recommended standard one hour (Toister, March 2020, 3,200+ people); in Toister’s 2018 survey, one hour standard and 15 minutes world-class | 62% of 1,000 companies did not respond; 12 hours 10 minutes average among those that did (SuperOffice email benchmark) | |
| Live chat | No verified survey. Our advice: reply while the shopper is still on the page | 2 minutes 40 seconds average wait; 21% of chat requests not answered (SuperOffice chat study, 1,000 websites, chats started during support hours) |
| Facebook messages | Standard of one hour, 15 minutes world-class for Facebook (Toister, 2018) | No reliable public benchmark |
| X (Twitter) DMs | Target 15 minutes (Toister, 2018) | No reliable public benchmark |
| Social in general | Most consumers want a reply within 24 hours or sooner (Sprout Social Index) | No reliable public benchmark |
How old is this data? The Toister surveys date from 2018 and 2020. The SuperOffice email benchmark page was last updated on 16 March 2023, its chat study on 26 January 2021, and the Sprout Social article is from 13 June 2024. Treat all of it as directional: your own median per ticket type beats any industry average.
Phone is a different queue with its own metrics, so this page covers text channels only.
How do you measure customer service response time?
Export the ticket created time and the first public reply time, take the median per ticket type and channel, and decide once whether you count business hours or calendar hours. Then report the same way every month so the trend means something.
Use the median, not the mean, so one ticket left over a weekend does not distort the picture. Our post on ROI metrics defines first response time the same way.
Business hours and calendar hours tell different stories. An email sent at 10pm and answered at 8am is minutes in business hours and ten hours for the customer. Report both if you can; the customer only experiences calendar hours.
Exclude autoresponders and internal notes, and count the first public reply that addresses the question. Segment by channel and by ticket type (WISMO, returns, damaged items, cancellations), because one blended average hides the slow queue.
If you use Zendesk and want targets and breach alerts on these numbers, see Zendesk SLAs. To measure email at all, it has to land in one queue first: see email tickets.
What response time should an online store aim for?
Set the target by the deadline the ticket carries, not one number for everything. Anchor email to the one-hour standard, chat to replying while the shopper is still there and social to the 24-hour standard (tighter on Facebook and X if you can), and put the tickets that expire first.
| Ticket type | Why the clock matters | Which metric to watch |
|---|---|---|
| Cancellation or address change | The order may ship before you answer | First response and resolution, both before dispatch |
| Pre-purchase product question on chat | The shopper may leave | First response time |
| WISMO | High volume, low urgency per ticket | Resolution in one reply, with the tracking status |
| Damaged item | The customer is already unhappy | Resolution time |
| Refund | The promise is an outcome, not a reply | Resolution time |
The numbers in this guide are the cited standards above. Anything finer, such as a target in minutes for each ticket type, should come from your own medians.
How do you cut response time without lowering quality?
Remove the work between reading a ticket and sending a complete answer: looking up the order, checking the policy and typing the reply. Then answer the repetitive ticket types without a queue. A faster holding message is not a faster answer.
Answer completely the first time. Put order status, tracking and the policy in the first reply so there is no second round. Most companies in the SuperOffice test did not.
Use reply templates for stable facts. The response template and email replies guides give you the wording.
Put deadline tickets first. A cancellation before dispatch should never wait behind a general question.
Cover evenings and weekends. That is when the queue builds and calendar-hour response time grows.
Offer self-service for WISMO. Tracking questions are the easiest to answer in one reply.
Let an AI agent draft or answer inside your help desk. More on that below, and in our agent comparison. For broader reply-quality tips, see better responses.
Where does Engaige fit?
Engaige (that’s us) isn’t a reporting or SLA tool: it’s the e-commerce-native AI support agent that works inside the helpdesk your conversations land in, with Engaige AI configuring and improving it. At Otrium, 60% of 120,000 annual tickets are resolved end to end, HelloPrint improved time-to-first-response by 90%, and the ticket types you switch to autonomous answer with no queue.
You go live in Agent Assist in hours. The agent drafts replies and suggested actions for your team inside your helpdesk, from your store data, your storefront and any docs you add, so the lookup work behind a slow reply disappears. Nothing closes without a human, and every reply your team sends becomes training data.
Speed comes in two steps. First the draft removes the minutes a person spends finding the order and checking the policy. Then you flip ticket types to autonomous one at a time, most brands in about a week, and those answer with no queue at all. Engaige is text-first: email, chat, WhatsApp, SMS and social DMs, no voice AI. The full loop is in the templates hub.
HelloPrint shows what that does to the first number. It improved time-to-first-response by 90%. Automation now covers 80% of its tickets, and the team went from 100 to 28 agents. Read the HelloPrint case.
Engaige proved to be invaluable. Their hands-on support during the implementation phase resulted in significant improvements to our automated resolution rate and CSAT.
Speed matters less than resolving on the first reply. Engaige reads the order, checks your policy and acts. Policies control how it reasons, what process it follows and when it may act, and every answer and action shows why, with the source.
Engaige works alongside your help desk. It does not set or manage SLAs. Keep your helpdesk. Replace the AI. You can see it in a demo.
Frequently asked questions
What is customer service response time?
It is how long a customer waits for a reply. Usually it is reported as first response time, from the customer’s first message to the first reply that addresses it, and sometimes as average reply time across the whole conversation.
What is the difference between first response time and resolution time?
First response time is how quickly someone replies; resolution time is how quickly the problem is solved. A holding message improves only the first. An automatic acknowledgement does not stop the first-response clock.
What is a good customer service response time for email?
Toister’s recommended standard is one hour, with 15 minutes as world-class. SuperOffice found that 62% of 1,000 companies did not reply to email at all, and those that did took 12 hours 10 minutes on average.
How fast should an online store reply on live chat and social media?
On chat, reply while the shopper is still on the page. For social, Toister suggests one hour for Facebook messages and 15 minutes for X DMs, and most consumers want a reply within 24 hours or sooner.
Should I measure response time in business hours or calendar hours?
Both if you can. The customer only experiences calendar hours. Use the median per channel and ticket type so one forgotten ticket does not distort the result.
How do you reduce response time without lowering quality?
Answer completely in the first reply, put deadline tickets first, and cover evenings and weekends. Removing the lookup and policy-checking work between reading and replying helps more than a faster holding message.
Can an AI agent reduce first response time?
Yes. Engaige’s Agent Assist goes live in hours and drafts replies inside your helpdesk, and autonomous ticket types answer without a queue. HelloPrint improved time-to-first-response by 90%.
Does Engaige set or manage SLAs?
No. Engaige works alongside your help desk and its SLA settings. It does not set targets.
