What Is a Greeting Message? Best Practices for Live Chat

A passive greeting appears when a visitor opens the widget themselves, while a proactive one appears uninvited and has to justify the interruption.

The passive greeting

Someone who opened the widget has already decided they want help, so the greeting only needs to orient them rather than persuade them to engage.

This is where scope and options matter most, since the visitor is deciding what to ask rather than whether to ask anything.

The proactive greeting

A proactive message interrupts, which means it has to be relevant enough to justify the interruption or it will be dismissed reflexively along with everything that follows.

Relevance comes from referencing what the visitor is looking at, which is why the same proactive message across every page performs poorly everywhere.

Different wording for each

Using one message for both is a common configuration shortcut and it produces a passive greeting that feels like an advertisement or a proactive one that feels pointless.

Writing them separately takes ten minutes and is among the higher-return adjustments available.

The most effective greetings differ by page, by availability and by whether the visitor has been here before.

By page

A pricing page, a support article and a checkout each attract different intent, and a greeting matching that intent converts considerably better than a site-wide default.

Configuring three or four page-group greetings captures most of the benefit without becoming a maintenance burden.

By availability

The greeting should change when nobody is available, stating plainly what can still be handled and when a person will respond.

A greeting promising immediate help outside coverage hours is the most avoidable expectation failure in any chat deployment.

By visitor history

A returning visitor with an open conversation should be greeted with that context rather than an introduction, since being reintroduced signals the system has forgotten them.

This requires the widget to recognise the visitor, which is one of the simplest and most noticeable personalisation wins available.

The frequent failures are excessive length, false urgency, pretending to be human and asking a question the visitor cannot usefully answer.

Too long to read

A greeting running to several sentences is skipped entirely, particularly on mobile where it can occupy most of the visible widget.

One or two short sentences plus options is the practical ceiling, and anything beyond that reduces the chance the message is read at all.

Manufactured urgency

Countdown language, limited availability claims and pressure tactics in a greeting damage trust before any conversation has started.

They also perform badly, since visitors have learned to dismiss anything that looks like a promotional interruption.

Asking an unanswerable question

Greetings asking what brought the visitor here today or how their day is going produce no useful information and delay the conversation by a turn.

If a question is asked, it should be one whose answer routes the conversation, such as which of several areas the visitor needs.

Mobile is where most greetings are seen and where most are badly formatted, since they were written and reviewed on a desktop.

Space is the constraint

A greeting occupying half a phone screen feels intrusive in a way the same message on a desktop does not, and dismissal rates reflect that difference.

Reviewing every greeting on an actual phone before publishing catches problems that no amount of desktop review will.

Tappable options over typing

Quick reply buttons matter more on mobile than anywhere else, since typing a question on a phone is enough friction to end the interaction.

Three options is usually the practical maximum before the list itself becomes a scrolling problem.

Timing differs on mobile

Proactive triggers tuned for desktop dwell time frequently fire too early on mobile, where visitors scroll faster and read differently.

Setting separate timing for mobile is a small configuration change that noticeably reduces dismissal rates.

Greetings are the easiest element to test and among the highest impact, since a small wording change affects every visitor who engages.

What to measure

Reply rate is the primary measure for both greeting types, with dismissal rate mattering particularly for proactive messages where an unwanted interruption has a cost.

Reading these alongside resolution rate prevents optimising for engagement that produces conversations the system cannot handle.

Testing one thing at a time

Changing the wording, the options and the timing simultaneously makes the result uninterpretable, which is how most greeting tests waste their traffic.

Greetings receive enough impressions that sequential tests reach a conclusion quickly even on modest traffic.

Testing by page group

A greeting that wins on a pricing page frequently loses on a support article, so results should be evaluated per page group rather than site-wide.

Applying a site-wide winner across pages with different intent typically undoes the gain it produced where it was tested.

Watching what visitors ask first

The questions arriving immediately after a greeting show whether the stated scope matched what visitors actually wanted, which no engagement metric reveals.

A mismatch here usually means the greeting is promising the wrong things rather than being poorly written.

Chatboq supports separate passive and proactive greetings, page-group targeting, availability awareness and quick reply options.

Context-aware by default

Greetings vary by page group and change automatically when nobody is available, so the message always reflects what the visitor can actually expect.

Returning visitors with an open conversation are greeted with that context rather than reintroduced to the assistant.

Disclosure handled once

Identity disclosure is part of the greeting configuration rather than something each variant has to remember to include, which prevents it being omitted.

The same applies when a person joins later, since that transition is announced automatically.

Reporting by greeting variant

Reply and dismissal rates are reported per greeting and per page group, which makes testing straightforward rather than requiring separate analytics work.

Resolution rate is reported alongside, so a greeting driving engagement the system cannot serve is visible rather than celebrated.