A bad website can kill your business. A good one won't save it.

Two businesses, one trade, two results

Principles are arguable, cases are observable. So let's start with two situations I run into regularly, with comparable businesses and results that aren't comparable at all.

Dupont and Snow, heating engineers

Two heating firms, same area, same skills, same diligence. Both have a website.

Dupont has a more or less decent site. On a desktop it holds up. On a phone the blocks overlap and you hunt for information. And to request a quote, one option only: the contact details. A number, an email address. Call us.

Snow has a very simple page. Genuinely simple: what they do, a few completed jobs, reviews from their customers. Nothing spectacular. But across the whole page one invitation keeps coming back: arrange a visit to your home.

And that invitation is bookable. The visitor opens a calendar, picks a thirty-minute slot that suits them, leaves their details, confirms. Done.

What Snow understood, and Dupont doesn't

The difference isn't technical. It's human.

Requesting a quote from Dupont means picking up the phone, possibly reaching voicemail, or writing an email explaining your situation. It's slow, it commits you, and it has to happen when the company answers. Clearing that hurdle takes motivation, or an outright emergency.

But the person looking for a heating engineer isn't choosing yet. They're browsing. Three tabs open, comparing, and from where they sit everyone looks the same.

On Snow's site they meet something else: “I can book an appointment, they come to me, and I pick the slot.” Thirty seconds. No call to make. No situation to explain.

And here's the point most businesses miss: the appointment is booked before qualification, not after.

No, not every booking will be a good one. Someone will book for a problem outside your trade. That's fine, and one call settles it: “you've booked a visit, I'd like to ask you two or three questions before I come out.” You qualify, you do your job as a salesperson, and if it doesn't fit your business you cancel.

But in the meantime, the person has committed. They're expecting you. They're no longer in the three tabs.

That's exactly what having one door and one objective on a page means, rather than a list of contact methods left to the visitor's initiative.

Restaurant Rouge and Restaurant Bleu

Second case, closer to pure experience. Two restaurants, both with online booking and payment.

Restaurant Rouge runs a booking system from another era. It more or less works, but it feels amateur. Wobbly steps, a dated interface, the odd error. Nothing dramatic on the face of it.

Except the person booking has to pay. And at that point an interface detail turns into a question of trust: am I giving my card to this thing?

Restaurant Bleu offers a modern system, fast, without a hitch, with the payment methods everyone already uses: Apple Pay, Google Pay. Three seconds, no card to dig out of a wallet, a fingerprint and it's booked.

Bleu fills more tables. It isn't even debatable, and it isn't a matter of taste.

You don't control everything, but you control this

Here's the real lesson from both cases.

Your customers don't compare your booking flow to the restaurant across the road. They compare it to Amazon and Airbnb, because that's where they buy the rest of the time. Their reference point is set by platforms that spend fortunes making payment invisible.

You can't do anything about that. But you can remove every objection that depends on you:

• trust in your system, solved with payment methods the person already knows;
• friction, solved by cutting the journey into short steps;
• doubt about your seriousness, solved with visible reviews and completed work;
• the effort required, solved by offering the easiest possible action.

Once those objections are gone, only one decision is left: the date, the time, the menu. And that one is already yours.

At equal skill, it isn't the best tradesperson who gets the appointment. It's the one where booking took the least effort.

What drives people away before they read a word

Now the detail. These points look obvious, which is precisely why they're never checked.

Loading time

A slow site loses people before it has said anything. Not because it's frowned upon, because it's irritating. You live it several times a day yourself: a page spinning, a thumb moving back up, gone.

This isn't a purist's concern. It's the first filter, and it applies before your offer, before your photos, before your reputation. A stuttering interface, animations that drag, a module that takes four seconds to open: every time, a share of your visitors disappears without a trace.

Where you're supposed to click

Second filter: can people tell, immediately, what to do.

Plenty of sites make you search. The button sits at the bottom, or hides in a menu, or drowns among six other links of equal weight. The visitor won't dig. They promised nobody anything, and nobody will make the effort for you.

The good news is that none of this is a secret art. There are people whose job it is, and most of what they know is now documented and available. Usually it's enough to apply simple rules and honestly put yourself in the position of someone arriving fresh.

The phone

This is where the overwhelming majority of visits happen, and where most sites fall apart. I wrote a whole article about it, because the subject comes up in nearly every review I do.

The mechanism is always the same: the site was designed, presented and approved on a large screen. Nobody ever opened the version visitors will actually see.

A visitor isn't judging your site. They're trying to do something, and they give up when it resists.

What makes people want to, or not

The visitor has cleared the first filters and is still there. Now the question is whether what they see makes them want to go further.

Design, that false friend

It's tempting to assume a more elaborate site produces more results. It rarely does, and I made it a subject in its own right.

Design is subjective. What pleases a company director tires their customer. What impresses a peer helps nobody book. Past a certain threshold of tidiness, adding aesthetics adds nothing to the result.

What matters is elsewhere: simple, clear, effective, and above all functional. If it works, it succeeded.

That said, there is a floor. A site that looks fifteen years old casts doubt on everything else, your seriousness included. It isn't fair, but it's how it goes, and it's often one of the genuine signals that a rebuild is due.

Language

An overlooked point with an immediate effect. If your business receives foreign visitors, a shop in a tourist town, accommodation, a restaurant, a French-only site turns away part of your demand.

And often the part that spends most. Someone travelling, on holiday, on a work trip, decides quickly and spends more. If they can't read your page, they open the next one.

The awkward part is that this loss is completely invisible. Nobody will write to tell you they didn't understand.

Visuals, and the photographer excuse

Some trades are judged by eye. A patisserie, a restaurant, a salon, a tradesperson who installs things. There, the image isn't decoration, it's the argument.

And yet, how many sites have been waiting two years for “the photo shoot” that was never scheduled, showing stock images everyone recognises in the meantime.

So let's be clear: you don't need a photographer, and you don't need a shoot.

• A recent phone takes very decent photos, often better than you'd expect.
• Daylight by a window is enough, no equipment required.
• Light retouching, contrast and cropping, handles the rest in two minutes.
• If you know how to use it, AI can clean a photo up, provided it doesn't show.
• A real photo of your work beats a perfect image that isn't yours.

Take twenty photos of your products on your worktop one morning and you'll have enough for a year. That's exactly the gap between a digital storefront kept current and a site left as it was.

A note in passing: my own site contains no photographs at all. That's a deliberate choice, consistent with what I sell. It doesn't apply to a patisserie.

The videos nobody hears

A video on a site is often a good idea. A talking video without subtitles, never.

Picture the real situation: the person is on a train, in a waiting room, sitting beside someone, with no headphones. They press play out of curiosity, they hear nothing, they watch a mouth moving. They stop after three seconds.

You paid for a production, you featured it prominently, and it conveys nothing to the majority of the people who play it. Subtitles cost a few minutes.

The most expensive thing isn't what your site fails to show. It's what it shows that nobody can use.

Standing where your visitor stands

Everything above follows from one reflex, and it costs nothing.

Where exactly did they come from?

A visitor doesn't arrive from nowhere. They come from a specific place, in a specific frame of mind, and that changes what they expect on arrival:

• from your Google listing, looking for a local professional, right now;
• from an Instagram post, in discovery mode, with no intent to buy;
• from an ad, holding a precise promise the page has to deliver;
• from a recommendation, with trust already granted and high expectations.

Four situations, four expectations. Sending all four to the same homepage is what turns an acquisition budget into water poured into a sieve. And it's also why knowing where your traffic comes from matters more than having more of it.

On which phone, exactly?

The most widespread mistake fits in one sentence: a site always works perfectly on the device belonging to whoever signed it off.

You test on your phone: recent, large screen, solid connection. All fine. Now:

• open it on a phone that's four or five years old;
• open it on a small screen, where layouts break;
• open it on average mobile data, not on your home wifi;
• go all the way through, to confirmation, not just the homepage;
• have someone who doesn't know your trade do it, and stay quiet while they try.

That last one is the most instructive and the most uncomfortable. Watching someone spend twenty seconds hunting for a button you find obvious is worth every report.

First impressions never replay

And here's why all of this matters more than people think.

Someone who had a bad experience on your site doesn't come back. They didn't judge you consciously, they simply retained an impression: this doesn't work, these people aren't on top of things. They found their answer elsewhere within five minutes.

And you have no way of getting them back. You can't write to them to say it's fixed. You don't even know they came.

That's what makes this loss so peculiar: it's permanent, silent, and it appears in none of your figures. Unless you decided to measure it.

You will never see the people your site drives away. Which is exactly what makes the problem so easy to ignore.

Look further than your competitors

You can't judge your own site on your own, you know it too well. You need points of comparison, and they aren't where people usually look for them.

Your competitors, including three hundred kilometres away

It sounds banal, almost too simple, and yet very few businesses do it seriously.

Open Google, type what your customer would type, and go through the sites that come up. Not skimming: go all the way, as if you meant to buy. You'll see immediately who has thought about it and who has merely put a site online.

Then do the same in another town, another region. Someone has probably already solved your problem without being your competitor, so they have no reason to hide it. Watching what others do isn't an admission of weakness, it's reconnaissance.

Your real competitors are the platforms

Here's the point I find most useful, and the least practised.

If you have a booking system, your benchmark isn't the restaurant next door. It's Uber Eats, it's Airbnb, it's the last app your customer used yesterday evening. They set the standard, and they spent money on it you'll never have.

So use them. Open those journeys and observe precisely:

• how many steps before confirming, and what each one asks for;
• when they ask for your details, and when they manage without;
• what they display to reassure, and exactly where they place it;
• what they don't ask for, which is often more instructive than the rest.

You don't have their means. You can copy their logic, and a tool you control is precisely what lets you adapt it to your trade instead of accepting whatever a subscription lets you configure.

And then, count

All of the above stays guesswork until you count. Instinct is a good starting point and a very poor finishing one.

How many visitors arrive, how many start an enquiry, how many finish it. Three numbers are enough to begin, and they'll tell you what to fix first instead of rebuilding everything.

It's also what will stop you pouring more ad budget into a journey that loses people halfway.

Your customers' standards weren't set by your industry. They were set by the apps they use every day.

What a good site won't do

I've spent a while on what a bad site costs you. It would be dishonest to stop there.

The reverse isn't true

I've spent this article demonstrating that a bad site is expensive. Now for the other half, even though it doesn't suit me.

A good site won't save a business whose offer doesn't hold.

If your price is out of line, if your service disappoints, if nobody is looking for what you sell, if you don't call people back: no website on earth will make up for it. You'll simply have more visitors witnessing the problem.

A site amplifies what exists. It doesn't replace it. That's why I sometimes turn projects down: rebuilding the shopfront of a business with a deeper problem means taking someone's money to delay their real question. That would go against everything I stand for.

The asymmetry is there, and it's cruel: a bad site can sink a healthy business, while a good site won't rescue an ailing one.

Where to pick this up, concretely

If you want to know where you stand, do this in order. It takes an hour:

• open your site on your phone and go all the way through an enquiry, timing yourself;
• count how many clicks and how many fields separate arrival from submission;
• check that the main action is visible without scrolling;
• replace at least one generic image with a real photo of your work;
• have someone from outside run the journey, without helping them;
• write down the three numbers you're missing to know what the site produces.

You don't need to rebuild everything. In most cases I see, two or three targeted fixes change more than a full redesign, and for a fraction of the budget.

Your site will never tell you it's costing you customers. That's for you to go and check.

If you'd like an outside opinion on yours, write to me. I'll go through it the way one of your customers would, and tell you straight what's blocking, what's a quick fix, and what isn't worth the trouble.