WhatsApp links: driving conversations from your website

A WhatsApp link opens a conversation with a specific number, optionally with a message already typed into the input field. It is an ordinary web link, so it works anywhere a link works: a website, an email signature, a printed QR code, a social profile, an advertisement.

The construction is simple enough that no tool is required. The address takes the form wa.me followed by the phone number, and an optional text parameter carries the pre-filled message. The number must be in full international format — country code included, no plus sign, no spaces, no dashes, no brackets. This is the single most common reason a link does not work, and it is worth checking character by character before assuming something more complicated is wrong.

The pre-filled text has to be URL-encoded, which is what the various link generators mostly do for you: spaces and punctuation become escape sequences. A generator is a convenience rather than a necessity, and understanding the shape of the link means being able to fix one by hand and to build variants without going back to a tool.

One thing the link is not: it is not a form of consent. Someone who clicks a link and sends a message has started a conversation, which is different from having opted in to receive marketing later.

The default instinct is to pre-fill a greeting. "Hello" is the most common pre-filled text on the internet, and it wastes the opportunity entirely.

A conversation opening with "Hello" tells the business nothing. Somebody has to ask what the person wants, wait for an answer, and only then begin being useful. That is two messages spent establishing something the context already implied — the person clicked a link on a specific page, so the business could have known what it was about.

A better pre-filled message states the thing being enquired about, phrased as something the customer would plausibly have typed themselves. On a page about a particular service, it should name that service. This lets the first reply be substantive, and it makes the customer feel understood rather than processed.

It should be short and editable. The customer can change the text before sending, and a long paragraph written in the business's voice reads as a script being put into their mouth — some people will delete it and type "hi" instead, which returns you to the original problem. One line, in plain language, phrased as a question the person actually has.

Because the pre-filled text can vary by link, it doubles as a source attribution. A different phrasing on each page tells the business where an enquiry came from without any tracking infrastructure at all.

Placement determines what kind of enquiry arrives, so it is worth thinking about intent rather than adding a floating button and hoping.

A product or service page produces the most useful conversations, because the visitor is already thinking about a specific thing and the pre-filled text can reflect it. This is the highest-value placement and the one most often neglected in favour of a site-wide button.

A contact page is expected and should be there, though the traffic is lower — people arrive with an intention already formed.

A site-wide floating button catches people wherever they are, which means the enquiries are less specific. It is worth having, with a generic but still useful pre-filled message, while recognising it will not perform like a page-specific link.

QR codes bridge physical and digital: on a shop sign, a receipt, packaging, a leaflet, a delivery note. They work well precisely because the alternative — typing a number into a phone — is enough friction to stop most people.

Email signatures and social profiles cost nothing and occasionally produce a conversation that would not otherwise have happened.

What is worth avoiding is putting the link everywhere with identical pre-filled text. That maximises clicks and minimises the information each one carries, which trades away the main advantage of the mechanism.

The link generates a conversation, and everything that determines whether it is worth anything happens afterwards.

Because the customer sends the first message, a customer service window opens for 24 hours, during which the business can reply freely in its own words rather than through an approved template. That is a genuinely favourable position, and it is also a deadline: let the window close and the same enquiry requires a template.

So the readiness questions matter as much as the link. Which number receives these conversations, who is watching it, during which hours, and what happens to an enquiry that arrives at midnight. A link on a busy website generates enquiries continuously, and one that nobody answers is worse than no link, because the customer has now been ignored by a business that invited them to make contact.

An honest holding message helps where immediate replies are not possible, provided it states when a reply will come and does not pretend to be a person.

It is also worth deciding what gets recorded. An enquiry that arrives, is answered and scrolls out of view is a conversation the business cannot follow up. Knowing which link produced it — which the varied pre-filled text tells you — and what happens next is the difference between a channel and a series of disconnected exchanges.

A wrongly formatted number is the most frequent failure. The number needs the country code and nothing else: no plus sign, no zeros prefixed for domestic dialling, no spaces or punctuation. An Indian number written the way it appears on a business card will not work.

Using a personal number is the next. A link on a public website sends conversations to whatever number it names, and if that is someone's personal handset the business has merged its customer service into a private phone, with no ability to hand over, cover absence or separate work from life. It is also difficult to undo later, because the number is by then printed on things.

Linking to a number that cannot receive business messaging is a subtler version of the same problem, and worth verifying before printing a QR code on packaging.

Assuming the link implies consent is a policy mistake rather than a technical one. A customer who messaged through a link has opened a conversation; they have not agreed to receive marketing afterwards. Adding them to a promotional list on that basis is sending marketing without consent.

Finally, a link never tested on a real phone. Test it on both a phone with WhatsApp installed and one without, and check that the pre-filled text renders correctly rather than showing escape sequences — an encoding error is invisible on the page and obvious to the customer.

Once links work, a few refinements are cheap and useful.

Vary the pre-filled text by placement so each enquiry identifies its own source. This is the closest thing to free attribution available on this channel, and it requires no more effort than writing a different sentence.

Make the button text specific. "Chat with us on WhatsApp" is clear and unremarkable; wording that names what the person will get — asking about availability, requesting a quote — sets the expectation before the click and produces better-qualified conversations.

For QR codes, remember the person is standing in front of something physical, so the pre-filled text can reference it: the product on the packaging, the service on the leaflet. That context is available and almost always wasted.

Set expectations near the link rather than after the click. A line stating business hours beside the button prevents the disappointment of an unanswered midnight message better than any automatic reply does.

And review the enquiries the links actually produce. If most are questions the website should have answered — price, hours, whether something is in stock — the link is functioning as a workaround for missing information, and answering those questions on the page reduces the messages while improving the site. The links then carry the conversations that genuinely need a person.