Chat response time: the gap between 30 seconds and 30 minutes

When a visitor opens the chat on your site, they aren't looking for a chat. They're making a decision, and one question is standing in the way: will this fit me, will it ship tomorrow, can I return it.

The moment they ask is the moment their intent is highest. An answer in thirty seconds makes the chat part of the sale. An answer in thirty minutes arrives in a tab they've already left.

That's why one metric outranks the rest in live chat: first response time.

What first response time actually is

The gap between the customer's first message and the first reply written by a human. The automated greeting doesn't count. "Hi, how can we help?" is closer to a doorbell than to an answer.

The distinction matters, because a measurement that counts the greeting will show you a perfect dashboard while the customer is still waiting.

You can't shorten what you don't measure

Most teams judge response time by feel: "we're usually quick." Usually quick hides being very slow at particular hours. Lunch, after 6pm, and weekends: the bad numbers always pile up in the same three places.

You don't need elaborate tooling to start. Looking at when conversations get picked up is enough to show the pattern: which hours back up, and which channel waits longer.

Cutting it without hiring

1. Remove the cost of switching channels

If an agent watches WhatsApp on a phone and site chat in a browser, part of their day goes to switching windows. Worse: to know whether anything is waiting, they have to check both.

With both channels in one list, that cost disappears. Waiting work is on one screen, oldest at the top.

2. Put repeat answers behind a shortcut

Write down every answer you send for a day. Odds are the top five cover half the volume: shipping times, refund policy, size chart, payment options, opening hours.

Saved as canned responses, /shipping is the whole interaction. The time you get back goes beyond typing time, because the agent also stops re-deciding how to phrase it.

3. Make ownership explicit

"Everyone's watching it" means "nobody's watching it" in practice. When claiming a conversation puts a name on screen, both failure modes end: two people jumping on one message, and nobody jumping at all.

4. Put context in front of the agent

Seeing which page the visitor is on, which city they're in, and what you discussed last time turns the first reply from "hello, how can I help" into an actual answer. Telling someone standing on a product page "we have that in a 42" collapses a three-message exchange into one.

5. Don't look available when you aren't

This is the expensive one. If the widget says "Online", the visitor expects an immediate answer; when nobody is there, the experience is worse than if the chat had never existed.

Outside staffed hours, letting the chat fall back to a contact form is far more honest. The visitor leaves an email, their expectation is set correctly, and your answer reaches them. In Lunachat, presence is computed from real connection state: when the last agent closes their tab, the widget goes offline on its own.

What target is realistic

It varies by industry, but a practical frame:

• Site chat: the visitor is at the screen. Aim for under a minute.
• WhatsApp: expectations are looser, but under 15 minutes during working hours is a good bar.
• Offline form: first thing the next working day.

The number matters less than its consistency. A team that averages two minutes but occasionally leaves someone for two hours leaves a worse impression than a team that averages five and never slips.

Getting started

The first step is usually organisational rather than technical: make waiting work visible in one place. Lunachat's inbox keeps web chat and WhatsApp in the same list, with canned responses and ownership on the same screen.