When should you rebuild your website?.

"Our site is three years old, we should rebuild it." I hear that sentence every week. It's always said in the same tone, the tone of the obvious, the way you'd say a car needs new tyres. And every time, I ask the same question back, the one that makes everyone uncomfortable: why?

There's usually a silence. Then the real answers come: because it looks old. Because a competitor rebuilt theirs. Because it's been three years, that's why. None of those answers mention customers, money or results. And that's exactly the problem.

You don't rebuild a moving train

There's an image that sums up how I see a website, and it comes from a film: a train circling the planet endlessly, which must never stop. Everything is repaired in motion. You swap an engine part while it moves, you fix a carriage while people are inside, you adjust, replace, improve, without ever cutting the machine.

A website that works is that train. It runs while you sleep, it receives visitors, it turns some of them into customers. You don't stop it for six months to rebuild it, because during those six months you lose everything it was producing, and because you'll have thrown out what worked along with what didn't.

A stopped website is a different story. A train going nowhere, carrying nobody, rusting on a siding, you can dismantle without hesitation: you're destroying nothing, since it produced nothing. But then don't just repaint it. It needs an engine.

The whole decision is there. Not in the age of the site, not in its style, not in the current fashion: in knowing whether your machine is running or stopped. And plenty of owners simply don't know.

The only starting question

Before talking design, technology or budget, there's one question, and everything flows from it: does your website bring you anything?

Not "is it beautiful". Not "is it modern". Not "does it have traffic". Does it bring you customers, quotes, bookings, orders, appointments. How many, per month. And at what cost.

If you can't answer that with a number, then the redesign debate is meaningless: you're deciding whether to dismantle a machine without knowing if it produces. It's as absurd as that. And yet it's the situation of most companies I meet. Until you have a way to measure what happens on your site, the only thing to do is install the instruments. Not rebuild anything.

And while we're at it: look at everything else the same way. The ad budget, the videos, the community manager running your social accounts. Does all of it bring you business? How much? Plenty of companies spend to make noise around a website they've never evaluated. Watering the field without ever checking whether anything grows.

Deciding to rebuild a site you measure nothing on is throwing away a machine without knowing whether it was producing.

Case 1: the machine is running

Your website brings you customers. Few or many, but it brings some. In that case my answer is simple: don't rebuild it. Open it up, look inside, and find where it leaks.

Because a website that works always loses people along the way. Always. And that's where the money is, not in a new homepage. The questions are concrete:

• Where do people stop? There's almost always one page where you lose half your visitors. Find it before touching anything else.
• Does the form actually work? I don't just mean the technical bug. I mean the twelve required fields nobody wants to fill in.
• And the booking, order or quote funnel? Every step is a door, and every door loses people. Test them one by one, like a customer.
• Do you understand what to do within three seconds? A site with no clear call to action leaves the visitor to decide alone. They decide to leave.
• Does it work on a phone? Really, outside, on 4G, one-handed. That's where your customers are.

And those fixes, the details nobody finds glorious, are exactly what pays the most. Two serious numbers to convince you.

The first comes from a study Deloitte ran for Google across 37 major brands and more than 30 million sessions: improving mobile load speed by 0.1 seconds was enough to lift retail conversions by 8.4%, and above all to increase by 20.6% the number of visitors reaching the contact page (Milliseconds Make Millions). One tenth of a second. Not a redesign. One tenth of a second.

The second comes from the Baymard Institute, which has compiled cart abandonment studies for years: on average, 70% of people who fill a cart never reach payment, and the reasons are disarmingly mundane. Fees appearing too late, a forced account, a funnel that's too long (Baymard Institute). These aren't design problems. They're badly placed doors.

That's why I'm so wary of redesigns: in a running site, the biggest gains are almost always hidden in details nobody will ever see in a screenshot.

Case 2: the machine is stopped

The other case is simpler, and more common than people admit. Your website brings you nothing. It exists, it's online, it cost money, and it produces nothing. It isn't a machine: it's a leaflet someone put on the internet.

So yes, in that case, rebuild everything. Without hesitation, since there's nothing to preserve. But beware the trap, because this is where the bad decision gets made: nine times out of ten, you'll rebuild the same leaflet, prettier. New colours, new photos, new typography, an animation, and six months later you'll have exactly the same result. Which is to say, none.

Rebuilding a website that brings nothing isn't decorating. It's construction. You don't start from the design, you start from the goal: what must this site produce, for whom, and by which path. You map the customer's journey before you draw the page. You decide what you won't include. You install measurement from day one. And the design comes last, to serve all of it.

A website that brings nothing doesn't have an appearance problem. It has an intent problem. Nobody ever decided what it was for.

The fresh-paint trap

Take a hair salon. The owner repaints the walls, changes the lighting, buys new chairs. That's fine: it's fresher, it's nicer, customers notice. I'd never say it's pointless, because the effect of new is real, and it counts.

But let's be honest about the return: nobody crosses town for a repainted wall. The same budget invested in a real experience, a machine that massages the scalp during the wash, a rethought welcome, a service the salon across the street doesn't offer, produces entirely different effects. In one case, you refresh. In the other, you give people a reason to come.

Your website is the same. The cosmetic redesign is the repainted wall. It pleases the owner, it pleases the board, it makes a nice LinkedIn post. It changes nothing about your revenue. What changes revenue is the journey you rebuild behind it: the message, the steps, the proof, the way you carry someone to the point of purchase.

So if you want to repaint because your site embarrasses you: go ahead, it's your right, and a company's pride isn't nothing. But don't kid yourself, and above all don't call it an investment. It's a comfort expense. That's not the same accounting line, and it's not the same job.

Repainting a wall brings nobody in. Rebuilding a website because it "looks old" is repainting a wall and hoping for revenue.

Your website is a full-time salesperson

Here's how I suggest you look at it, and I think it changes everything.

Your website is a salesperson. It works twenty-four hours a day, seven days a week, without a salary, without holidays, without complaining. It greets every prospect, explains what you do, answers objections, and tries to move them to action. It's probably the colleague who talks to the most customers in your company.

Now ask yourself: how do you treat a salesperson who isn't selling enough? You don't fire them every three years on principle, and you don't replace them because their suit looks dated. You look at their numbers. You listen in on their meetings. You look for the moment they lose the customer. You train them, you fix their pitch, you give them better tools. And if there's really nothing to salvage, then you change.

Your website deserves exactly the same treatment, no more, no less. Yet in most companies that salesperson never gets any feedback, any training, any correction. You let them repeat the same failed pitch for three years, then one day you replace them entirely with a brand new one, who you won't tell anything either. It's absurd, and it's the norm.

A frozen website is a dead website

We need to kill a stubborn idea: the site you deliver, put online, and never touch again. It isn't a business card. It isn't a plaque by the door. It's a working tool, and a working tool needs maintenance.

A living site is one where things happen: pages added because a question keeps coming up with customers, text rewritten because it doesn't convince, articles that bring people in, offers that change with your seasons, proof stacking up as projects finish. A company moves constantly: so does its site, or else it lies about what the company has become.

And to those who ask me for the right rhythm, I answer: the one you can hold. You can improve your site every month if your budget allows. You can do it every quarter. What matters isn't the frequency, it's that it's never finished. A small change every month, based on what you observe, will always produce more than a big redesign every three years decided on a hunch.

You can spot the sites nobody touches anymore instantly. The prices are wrong. The team shown left long ago. The latest news is four years old. That's not just dated, it's a signal to your customers: this company isn't looking after anything. The web's graveyard is full of them, and they're not ugly sites. They're abandoned ones.

So, when should you rebuild?

After all that, here's my answer, as sharp as I can make it.

The five good reasons to start over:

1. Your site brings you nothing, and was never designed to. There are no foundations: they need laying.
2. Your offer, your audience or your positioning has changed. The site speaks to someone who is no longer your customer.
3. The technical base is dead: nothing can be changed without breaking everything, nobody knows how to touch it, or you don't even have access to it.
4. It's unusable on a phone, and patching it would cost more than rebuilding it properly.
5. It's impossible to measure anything on it, so it's impossible to improve. You've been repairing blind for too long.

The four bad reasons, which I hear every week:

• "It looks old." Age isn't a criterion. Performance is.
• "A competitor rebuilt theirs." You don't know whether theirs works. Neither do they, probably.
• "It's been three years." Three years isn't an expiry date, it's an agency habit.
• "We changed our logo." Then change the logo. Not the foundations.

And if you take away only one thing: "when should I rebuild my site?" is the question of an owner looking at their façade. The right question, the one a business owner asks, is: what do I fix this month so it brings me more?

That's not the same question. And it doesn't cost the same price.