Change your mind as often as you like

Ask someone why their website still looks like it did in 2019 and you rarely hear that they like it. You hear that it works, more or less, and that the last time somebody touched it something went wrong, and that they would rather have a dated site than a broken one.

That is a completely rational position. It is also the reason most small business sites quietly get worse every year while nobody is allowed to touch them.

Two things remove the risk that position is protecting against.

Nothing you try is live until you say so

The site you are working on and the site your customers see are not the same site.

Everything you change lands in a preview that only you and the people you have invited can look at. Your live site carries on exactly as it was, serving visitors the version it has always served, until you press publish. There is no moment where a half finished experiment is accidentally facing the public.

This is worth sitting with, because it changes what an afternoon of messing about costs. It costs nothing. You can rebuild your homepage three different ways on a Tuesday, hate all three, and no customer will ever know it happened.

Every round of changes leaves a saved version behind

The second thing is that going backwards is easy.

Every round of changes saves a version. There is no button to remember and nothing to name, which matters because backup systems that require discipline are backup systems that do not exist. When you want an earlier state, you ask for it in the same conversation you use for everything else. Something like "put it back the way it was this morning" is enough. You get a short list of your recent versions, you pick one, and the site returns to it.

Two things worth knowing about how far that goes.

It is a list of your recent versions, not an archive of every day since you started. It is built for "yesterday was better", which is what people actually need, rather than for browsing your site's entire history.

And going back does not cost you anything else. The enquiries people have sent you, and the content you have typed into your own forms, sit outside the version history. Restoring an older look never wipes them.

So use it differently

Once you believe both of those, the sensible way to work changes.

The cautious approach, where you make one small change, look at it nervously, and stop, is a habit built for a world where mistakes were permanent. It produces sites that are slightly less bad every eighteen months.

Try the bigger thing instead. Ask for the homepage to lead with your work rather than your words, and look at it. Turn the whole site darker and see whether it suits your trade. Ask for the page you have been vaguely thinking about for a year. Rewrite your opening paragraph three different ways and compare them properly, on the actual page, instead of imagining them.

Most of those will be worse. That is fine, and it is the point. You find out in a minute instead of a month, and the one in four that is better is a real improvement you would never have got to by being careful.

The small print, honestly

Two things this does not do, so you know where the edges are.

It is not the same as undo on each word. It works on the whole site rather than on individual sentences, so for a small mistake it is usually faster to just say what was wrong and have it corrected than to restore a version.

And it is a list of recent versions, as above. If you want to keep a particular state around for a long time, publishing it is the thing that fixes it in place.

The point

A website that is safe to change is a website that gets better. A website that is frightening to change is a website you have already stopped maintaining, whether or not you have admitted it yet.

You can try all of this without publishing anything, because building and previewing are free and your live site is not involved. Build free.