Not legal advice: The week model below uses stated assumptions (messages arrive evenly across the week, a reply goes out the moment you open, arrivals even within each hour), not your traffic.
"How fast should we reply?" has a dozen popular answers (one hour, four hours, one business day) and almost no good sources behind them. Most figures in circulation trace back to vendor blogs repeating each other, and the few primary studies measure different things: what customers say they expect, what companies actually do, and what a dashboard reports when it ignores nights and weekends.
We read the primary sources we could reach, set aside the ones we could not trace, and added one calculation of our own: how long a message really waits in an Israeli work week. The short version is that the headline number matters less than the clock behind it. In one 2026 dataset of 26.5 million emails, the average reply (replied emails only) took 4 hours 17 minutes in work hours and 13 hours 20 minutes in real elapsed time.
Key takeaways
- Customers' stated expectations are tighter than company performance, but not absurd: in a March 2020 survey of 3,200+ consumers, 88% said a reply within an hour would meet their expectations, and only about 11% of US respondents expected 15 minutes or faster.
- The same reply can be "4 hours" or "13 hours" depending on the clock. In a Q3 2026 dataset of 26.5 million emails, the average was 4h 17m counting work hours only and 13h 20m in elapsed time. Unreplied emails are excluded from both.
- We modelled it. With Sunday to Thursday, 9:00 to 18:00 hours, 73% of the week is closed, and a message arriving at a random moment waits 14.5 hours for the doors to open. About 23% of messages wait more than 24 hours.
- Adding Friday 9:00 to 13:00 cuts the average wait by about a third (to 9.1 hours) and roughly halves the share waiting over 24 hours (to about 12%).
- WhatsApp's customer service window is 24 hours from the customer's last message, per Meta's documentation. Past it, free-form replies are not allowed and only approved templates can be sent, so the weekend gap is a policy problem as well as a service one.
- Speed is not the only lever. Widely cited research on customer effort links easy resolution, not delight, to repurchase intent, and platform data from one ecommerce helpdesk shows handoffs by contact form taking 36 hours to resolve with 42% abandonment versus 8 hours and 13% by chat.
How fast do businesses actually reply to customers?
Honest answer: slower than most owners assume, and the data is thinner than the blogs suggest. The most-repeated figure comes from a SuperOffice benchmark of roughly 1,000 companies: an average email reply time of 12 hours 10 minutes, with 62% of companies not answering at all. Coverage of it dates from 2018, so it is an old snapshot that some sites republish as current, and SuperOffice sells CRM software. Treat it as a rough direction rather than a current norm.
The cleanest recent number is not about support at all. EmailAnalytics, which sells inbox analytics, publishes quarterly averages from anonymized metadata of its customers' Gmail and Outlook accounts. For Q3 2026 (26.5 million emails, 20 industries) it reports 4 hours 17 minutes in work hours. Only emails that received a reply within 14 days count, and the sample is the vendor's own customers, so it describes responsive, email-tracking teams.
Ecommerce platform data shows a wide spread by sector. Gorgias, an ecommerce helpdesk, reports first response times from 1.6 hours (hardware) to 8.8 hours (apparel) across 14 verticals in a $10M GMV band, a 5.5x gap. It does not publish a sample size for each metric.
| Figure | Source | What it measures | Caveat |
|---|---|---|---|
| 12h 10m average email reply; 62% never answered | SuperOffice benchmark, ~1,000 companies | Company behaviour | Reported in 2018 coverage; vendor study |
| 4h 17m (work hours), 13h 20m (elapsed) | EmailAnalytics, Q3 2026, 26.5M emails | Replied emails in tracked mailboxes | Unreplied emails excluded; vendor's own customers |
| 1.6h (hardware) to 8.8h (apparel) first response | Gorgias Ecom Lab, 14 verticals, $10M GMV band | Ecommerce helpdesk tickets | No per-metric sample size; vendor data |
None of these is a random sample of businesses. Each comes from a vendor whose customers care about response time, so the real market average is probably slower, not faster.
How fast do customers expect a reply?
The best primary source for email is Jeff Toister's repeated consumer survey. The March 2020 round (3,200+ consumers across the US, UK and Canada, sponsored by SuperOffice) found that 88% said a reply within one hour would meet their expectations. Fewer wanted more than that: among US respondents only 11.3% expected a reply within 15 minutes, down from 14.9% in 2015 and 13.6% in 2018.
Read that carefully. It is a survey of stated expectation, not of behaviour, and "within an hour" means within an hour of asking, in the customer's mind, not within your opening hours.
88%
of 3,200+ consumers said an email reply within one hour would meet their expectations. About 11% of US respondents expected 15 minutes or faster.
Source: Toister Solutions consumer survey, March 2020 (sponsored by SuperOffice)
Other widely quoted "90% of customers expect an immediate response" figures trace to a 2018 HubSpot consumer survey and are repeated mostly by secondary sites. We could not reach the original wording, so we do not rely on them here.
Why does the number on your dashboard look better than the customer's experience?
A response-time metric has three silent choices built in: which hours count, which messages count, and when the clock stops. EmailAnalytics shows the first choice plainly. Its Q3 2026 average is 4 hours 17 minutes when the clock runs only from 8:00 to 17:00, and 13 hours 20 minutes in real elapsed time, three times as long, on the same emails.
Neither number includes messages nobody answered. SuperOffice's 62% no-reply figure is the other half of the picture: an average computed over replies alone describes the people you did get to, not the ones you did not.
Q3 2026, 26.5M emails analysed
How long does a customer really wait in an Israeli work week?
We built the simplest honest model. A week has 168 hours. If you answer Sunday to Thursday, 9:00 to 18:00, you are open 45 of them, which is 26.8%. Assume messages arrive evenly across all 168 hours and that your reply goes out the moment you open, with zero queue and zero lunch break. How long does a message wait just for the doors to open?
Averaged over every message, 14.5 hours. For a message that arrives while you are closed, 19.8 hours. And 23.2% of all messages (31.7% of those that arrive while you are closed) wait more than 24 hours, because anything sent after Thursday 18:00 meets a 63-hour gap until Sunday 9:00.
Opening Friday 9:00 to 13:00 changes the picture more than extending Sunday to Thursday by an hour. The longest gap shrinks from 63 to 44 hours, the average wait drops to 9.1 hours, and the share over 24 hours falls to 11.9%. Extending the weekday close to 19:00 on top of that barely moves the numbers (8.7 hours).
73%
of the hours in a week are outside Sunday to Thursday, 9:00 to 18:00. A message sent after Thursday 18:00 waits 63 hours for a first human reply.
Source: Our calculation (45 open hours of 168)
Real traffic is not uniform: most messages land during business hours, so the average wait in practice is shorter than the model. The point is the shape. The weekend gap, plus Shabbat and holiday eves, is where long waits come from, and it is fixable with an acknowledgement rather than more staff.
Why does the 24-hour WhatsApp window make the weekend gap worse?
On WhatsApp the business can send free-form replies only while a 24-hour customer service window is open. Meta's documentation says the window starts when a user messages or calls and resets to 24 hours each time they send another message. Outside it, only approved template messages can be sent.
That turns the model above into a practical rule. A customer who writes Thursday at 17:30 and gets a first human reply Sunday at 9:00 has waited about 63.5 hours, so the reply lands outside the window and has to go as a template. An automatic acknowledgement sent within minutes, while the window is open, is both the kinder reply and the cheaper one to design around.
Whether a given message is charged depends on category and market under Meta's per-message pricing, which has changed more than once. We cover those rules separately; here the point is only the 24-hour window.
Does a faster reply actually win or keep customers?
Less clearly than vendors suggest. The evidence that speed itself drives loyalty is mostly vendor surveys of expectation. The research most often cited on what does drive loyalty is CEB's customer-effort work by Dixon, Freeman and Toman, summarised in the July–August 2010 Harvard Business Review article "Stop Trying to Delight Your Customers" (a study of more than 75,000 customers). Its argument is that making a problem easy to resolve does more for loyalty than exceeding expectations. We do not quote its percentages, because we could not confirm them against the article text.
What the effort finding adds is that the first reply is only the start of the clock. Gorgias platform data on AI-touched tickets handed to a human shows how much the second half matters: AI-touched contact-form tickets handed to a human took 36 hours to resolve with 42% abandonment, email 32 hours with 30%, chat 8 hours with 13%. Gorgias also reports that about a third of tickets handed from automation to a human never received a human response. It is one vendor's ecommerce data, but the pattern (fast first touch, slow finish) is exactly what a first-reply dashboard hides.
| Channel (AI-touched tickets handed to a human) | Time to resolve (creation to close) | Abandonment |
|---|---|---|
| Contact form | 36 hours | 42% |
| 32 hours | 30% | |
| Chat | 8 hours | 13% |
Source: Gorgias Ecom Lab (vendor platform data, ecommerce stores; no sample size stated). Chat tickets are also simpler by nature, so do not read this as a clean channel comparison.
How do you set a response standard that survives the weekend?
- 1Pick two numbers, not one: a first-reply target (someone has seen this and owns it) and a resolution target (the problem is solved). Most small teams only measure the first.
- 2Choose the clock on purpose. Run it on your business hours and holidays, but publish the hours, and decide what happens to a message that arrives outside them.
- 3Automate the acknowledgement. Within a minute of a customer's message, send a short reply that says it was received, when a person will answer ("Sunday from 9:00") and what to do if it is urgent. It keeps the 24-hour WhatsApp window useful and sets an honest expectation.
- 4Split by urgency. A payment problem or a cancelled booking should have a tighter target than a general question, and the ones that breach should escalate to a named person.
- 5Measure the unanswered, not just the average. Count messages with no human reply at all after 24 hours, and the share that took more than one touch to resolve.
- 6Review the worst week, not the average week. The Thursday-evening and holiday-eve queues are where the customers who leave were lost.
| Message type | Reasonable first-reply target | Reasonable resolution target |
|---|---|---|
| Urgent (payment failed, booking changed) | Within 1 hour in business hours | Same business day |
| Standard question | Within 4 business hours | 1 business day |
| Complex or technical issue | Within 1 business day | 3 business days |
| Outside business hours | Automatic acknowledgement in under 1 minute | Next business day |
These are our suggested starting points, built from the stated-expectation data above, not benchmarks from a study. Tighten or loosen them from your own response-time history after a month.
How do you hold the standard in oToK?
In oToK's shared inbox you set a first-reply target and a follow-up target, and the clock respects your business hours and holidays. At-risk and overdue conversations are flagged in the list, a dedicated folder shows what needs attention now, escalation alerts go to the people you choose, and a breach can trigger an automation. Response-time analytics show where you stand per team member and over time.
Every new inquiry (a form, a Meta lead ad, a booking or a first WhatsApp message) opens an item with an owner and its own first-response clock. Customer tickets, with a help centre, routing rules, response targets, canned replies and auto-close, are available on Growth and up. An automation triggered by a customer's first message can send the immediate acknowledgement, and Shabbat and holiday observance holds automations over Shabbat and holidays and resumes them afterwards.
- Define your targets once, see every conversation against the clock
- Acknowledge automatically, answer personally
- Escalate breaches to a person, not a dashboard
Frequently asked questions
- What is a good customer support response time?
- For email, a reply within one hour of the customer's message in business hours is the standard 88% of surveyed consumers said would meet their expectations (Toister, March 2020, 3,200+ consumers). Within four business hours is realistic for most small teams, and an automatic acknowledgement closes the gap outside hours.
- What is the average email response time for businesses?
- It depends on the clock. EmailAnalytics' Q3 2026 data (a dataset of 26.5 million emails, replied emails only in the averages) shows 4 hours 17 minutes counting work hours and 13 hours 20 minutes in elapsed time. A widely cited SuperOffice benchmark of about 1,000 companies put the average at 12 hours 10 minutes, with 62% not answering at all, though coverage of it dates from 2018.
- Should response time count nights and weekends?
- Report both. A business-hours clock is fair for staffing, but the customer experiences elapsed time. In our Israeli week model with Sunday to Thursday hours, a message waits 14.5 hours on average for the doors to open, and about 23% wait more than a day.
- Does the WhatsApp 24-hour window affect support?
- Yes. Per Meta's documentation, free-form messages can be sent only within 24 hours of the customer's last message; after that only approved templates can be sent. A first reply that arrives after a weekend gap often falls outside the window, so send an automatic acknowledgement while it is open.
- Is first response time the metric that matters most?
- It is the easiest to measure, not necessarily the one that matters most. Research on customer effort points to easy resolution as the stronger driver of repurchase intent, and platform data shows resolution taking many times longer than the first reply. Track first reply, resolution time and the share of messages never answered.
