First response time: what it means, how to calculate it, and honest benchmarks by channel
First response time is the customer service number most people quote and fewest people calculate the same way. This explainer gives you a plain definition, the formula, a worked example that shows how much the reporting choice changes the answer, and benchmarks by channel with a note on how trustworthy each one is.
Quick answer
First response time (FRT) is how long a customer waits between sending a message and getting the first reply from your business. Calculate it per conversation as first reply time minus message time, then report the median, ideally in business hours. Customers set the bar high: in HubSpot research, 60% defined an immediate reply as 10 minutes or less.
What is first response time?
First response time is the time between a customer's first message and the first reply from a person or system acting for your business. It measures how long someone waits before they know they have been heard, not how long it takes to solve their problem.
It sits alongside two related numbers that are often confused with it:
| Metric | What it measures | Question it answers |
|---|---|---|
| First response time | Wait until the first reply in a conversation | How fast do we pick up? |
| Average response time | Wait for every reply in the conversation, averaged | Do we keep up once a chat is going? |
| Resolution time | Time from first message until the issue is closed | How long does the whole problem take? |
For sales enquiries, first response time matters most, because a slow first reply sends the buyer to a competitor before the conversation starts. For support, resolution time matters just as much. Our customer service KPIs guide shows how the three fit on one dashboard.
How do you calculate first response time?
For one conversation: FRT = time of first reply minus time of first customer message. For a period: median FRT = the middle value of all conversations' FRTs, or mean FRT = sum of all FRTs divided by number of conversations.
The steps below are the method we suggest to owners who have no help desk software, and they work for WhatsApp, email and website chat alike.
- Pick a sample. Take every new conversation from one normal week, or at least 30 if you have more. Leave out spam and messages you started yourself.
- Record two timestamps. For each conversation, note when the customer's first message arrived and when your first reply was sent. Use the message times shown in the chat.
- Subtract. First reply time minus first message time gives the first response time for that conversation, in minutes.
- Decide on the clock. Calculate both calendar time and business-hours time, where the clock only runs while you are open. Report whichever you promise customers, and label it.
- Report the median and the slowest 10%. Sort the results and take the middle value as your headline figure, then look at the slowest tenth of conversations, because that is where customers are lost.
Worked example: why the median and the clock matter
Six hypothetical conversations at a business open 9am to 6pm. Watch what one late-night message does to the average.
| Conversation | Message received | First reply | FRT, calendar | FRT, business hours |
|---|---|---|---|---|
| 1 | 10:02 | 10:05 | 3 min | 3 min |
| 2 | 11:40 | 12:10 | 30 min | 30 min |
| 3 | 13:00 | 13:08 | 8 min | 8 min |
| 4 | 14:15 | 14:17 | 2 min | 2 min |
| 5 | 16:50 | 17:30 | 40 min | 40 min |
| 6 | 21:30 | 09:20 next day | 710 min | 20 min |
| Mean | 132 min | 17 min | ||
| Median | 19 min | 14 min |
The same week can honestly be reported as 14 minutes or 2 hours 12 minutes. Neither is wrong; they answer different questions. The business-hours median tells you how quick the team is when it is working. The calendar mean tells you what the customer who messaged at 9:30pm experienced, and that customer does not care about your opening hours. Report the business-hours median as your headline, and track the calendar figure for after-hours messages separately, because that is the gap after-hours cover is meant to close.
Does an automated reply count as a first response?
An acknowledgement that only says "we will get back to you" should not count. A reply that actually answers the question should, whether a person or an AI sent it.
If you use automation, track two numbers so you cannot fool yourself:
- First useful response time: time until the customer got an answer, a booking, or a specific next step.
- First human response time: for conversations that needed a person, time from the handover until the person replied.
A team that installs an away message and sees FRT drop to zero seconds has changed the measurement, not the service. In our own product we look at the second number closely: an AI agent that answers in seconds but leaves handed-over chats waiting four hours has moved the delay, not removed it.
What is a good first response time by channel?
There is no single official benchmark. The figures that get quoted come from different studies, years and methods, so here is what each one actually measured.
| Channel or context | Figure | Source and what it measured | How much weight to give it |
|---|---|---|---|
| Any service question (expectation) | 60% define "immediate" as 10 minutes or less | HubSpot research, cited on HubSpot's blog | A survey of what customers say they want; a vendor's own research |
| Web form leads (what firms do) | Average 42 hours; 37% replied within an hour; 23% never | Harvard Business Review audit of 2,241 US companies, 2011 | Strong method, older data |
| Web leads (effect of speed) | Odds of qualifying a lead drop 21-fold from 5 to 30 minutes | Lead Response Management Study, 2007 | Widely cited, industry-run study, old |
| Customer service email (what firms do) | Average 12 hours 10 minutes; 62% did not respond | SuperOffice benchmark of 1,000 companies | Vendor's own test of other firms |
| WhatsApp and live chat | No independent benchmark we trust | Our working target: a useful reply within 5 minutes while open | Our opinion, not a statistic |
| Phone | Answered live, or not at all | Callers rarely wait; a missed call is a missed enquiry | Use answer rate instead of FRT |
The practical takeaway is that most businesses are slower than customers want, so beating the average is not hard. For new enquiries, aim for minutes, not hours. For support email, same business day is a reasonable promise for a small team. Our guide to measuring WhatsApp lead response time has a sampling method built for WhatsApp chats specifically.
How do you bring first response time down?
In rough order of cost, from free to paid:
- Turn on notifications for the right person. Many slow replies are simply unseen messages on a shared phone.
- Write quick replies for your ten most common questions so a reply takes seconds, not a composed paragraph.
- Give each hour an owner. "Whoever is free" means nobody during busy periods.
- Publish prices, hours and policies where customers look, so fewer people need to ask.
- Automate the routine answers and hand everything else to a person with a clear promise of when they will reply.
Steps one to four are covered in how to reply to customers faster without hiring.
Where ReplyKit fits
ReplyKit is an AI agent for WhatsApp and website chat, so it attacks the first useful response directly: it answers from your own content within seconds, offers and books real open slots, and captures lead details. When it does not know an answer it says so and flags the chat, and you are notified when a chat needs a human, which is where your first human response time starts.
It does not replace measurement. It has no phone channel, and it does not produce a response-time report for you, so use the sampling method above on your inbox before and after. WhatsApp connects through Quick Connect, a QR-code pairing that is unofficial and not sanctioned by Meta; use a dedicated number. Plans start at $29 a month with AI replies included; try it free for 7 days with no credit card and compare your median before and after.
Frequently asked questions
What is first response time?
First response time is how long a customer waits between sending their first message and receiving the first reply from your business. It measures how quickly you acknowledge or answer someone, not how long it takes to solve their problem.
How do you calculate first response time?
For each conversation, subtract the time of the customer's first message from the time of your first reply. For a period, take the median of those values, and say whether you counted calendar hours or only business hours.
What is a good first response time?
For new enquiries on chat or WhatsApp, aim for a useful reply within minutes; HubSpot research found 60% of customers define an immediate response as 10 minutes or less. For support email, replying the same business day is a reasonable target for a small team.
Is first response time a KPI?
Yes. It is one of the most common customer service KPIs because it is easy to measure and closely linked to whether new leads convert. It should be read alongside resolution time and customer satisfaction, because a fast first reply that solves nothing is not good service.
What is the difference between first response time and average response time?
First response time only measures the wait for the first reply in a conversation. Average response time measures every reply, so it shows whether you keep up once a conversation is underway.
Does an automated reply count toward first response time?
Only if it is useful. A message that only acknowledges receipt should not count, while an automated reply that answers the question or books an appointment should; track first human response time separately for chats that need a person.
Cut your first useful response to seconds
ReplyKit answers WhatsApp and website messages from your own content and hands the rest to you. 7-day free trial, no credit card.
Sources and further reading
- HubSpot, Customer loyalty trends · 90% rate an immediate response as important or very important; 60% define immediate as 10 minutes or less
- Harvard Business Review, The Short Life of Online Sales Leads (2011) · 2,241 companies audited; average 42 hours; 37% within an hour; 23% never responded
- Lead Response Management Study (2007) · 21-fold drop in the odds of qualifying a lead between 5 and 30 minutes
- SuperOffice, Customer service benchmark report · 1,000 companies; average response 12 hours 10 minutes; 62% did not respond to email
ReplyKit is an independent product and is not affiliated with or endorsed by Meta or WhatsApp. Benchmarks are quoted from the sources listed as read in October 2026. This article is general guidance.