The email address is the smallest piece of copy on a website and the only one a stranger will type out by hand. It is worth ten minutes of thought, which is about nine more than it usually gets.
October 2026
Every address says something before a single message arrives. It tells the reader what kind of thing is on the other end: a person, a department, a machine, or nothing at all. Four of them turn up on almost every website, and they do not say the same thing.
info@ is the oldest address on the web and the emptiest. Information is not a promise: it does not say who reads it, whether anyone answers, or how long that takes. It is a counter with nobody standing behind it, and everyone has written to one and heard nothing back.
An address is a promise about who will answer. Most sites make one without noticing.
contact@ is a form field that became a human being. It is correct, it is cold, and it reads like the label above the box rather than like an invitation.
noreply@ says outright that you are not to answer. Whatever else a business puts on its site about caring for its customers, this one line contradicts it.
hello@ is short, it can be said out loud over the phone without spelling, it survives in a sentence in almost any language, and it sounds like somebody picking up. That is the whole argument, and it is enough.
The unromantic reason to drop info@
Tone aside, the generic address is also the most heavily spammed one you can own. A harvester that finds a domain does not need to find an address at all: it guesses. It tries info, sales, admin, office, contact, and it has been trying them since the nineties. Thirty years of guessing have put those names on every list that was ever sold or leaked.
An address nobody guesses starts clean. It will not stay clean forever — nothing does — but the difference between a fresh mailbox and an inherited one is the difference between a filter that works and a folder you stop reading.
One address, not five
Small organisations like to publish a list: info@ for general questions, sales@ for prices, bookings@ for dates, and a personal address for whoever built the site. It looks organised. What it actually does is hand the visitor a decision they are not equipped to make, and every wrong guess lands in a mailbox somebody checks on Fridays.
One address, answered by a person, beats a switchboard that leads to three empty rooms. Sorting is your job, not the visitor’s. If the work later justifies a second address, add it then — and keep answering the first.
Putting it on the page without feeding the harvesters
Now the practical half. An address written out in plain HTML is collected within days of the page going live. The usual answer is a contact form, and a contact form is worse than it looks: the sender keeps no copy of what they wrote, cannot attach the drawing, cannot reply to their own thread, and has no way of telling whether the message arrived. Forms also fail silently, which is the only kind of failure nobody notices.
The second answer is to assemble the address with JavaScript. That hides it from the harvester and from everyone else too: from a reader with scripts blocked, from a screen reader that reads the page before the script has run, and from the search engines and AI assistants that are increasingly how people find a small business at all. Hiding your address from the things that recommend you is a poor trade.
What I use instead is built into WordPress and is older than most plugins: the address is written into the page as character codes. A visitor sees, selects, copies and clicks a perfectly ordinary address. A screen reader reads it out normally. A search engine indexes it. A crude scraper reading the source finds no pattern that looks like an email at all. It stops the cheap harvesters, not the clever ones — no trick does — which is why it is only half the job. The other half is at the mail host: the records that prove your mail is yours, and a filter that has something to work with.
The rule, in one line
One address per site, hello@ your own domain, written so that people and machines can read it and harvesters cannot, answered by somebody who says when they cannot help. Ten minutes at the start of a project, and it keeps paying out for years — which is roughly the definition of a small decision worth making well.
Read the originals
- RFC 2142: Mailbox Names for Common Services, Roles and Functions — IETF, 1997. Why postmaster@ and abuse@ are obligations and info@ is only a habit.
- Jared M. Spool, Brainsparks — UIE archive, on what happens to people who cannot find a way to ask.