A website is never finished.

Your website is a salesperson

First, an image has to be broken, because it's the one doing all the damage. As long as it holds, nothing that follows makes sense.

It isn't a business card

Most people file their website mentally in the same drawer as their brochure or their business card. Something you have made, sign off, hand out, and forget.

That's understandable, and it's wrong. A business card does nothing. It waits in a pocket. It speaks to no one, offers nothing, adapts to nobody.

Your website works. Every day, weekends included, and it receives more people than any of your salespeople will meet in a career. It welcomes, explains, offers, directs. And in the overwhelming majority of cases it does all of that on a phone screen, in conditions you have never tested.

It isn't a document. It's a member of your team.

A salesperson gets trained

Take any salesperson, front desk or field. On their first day they know nothing. Six months later they present differently, they've worked out which questions keep coming up, they've found the phrases that land.

How? Through learning, in all its forms:

• their own experience, the meetings that go well and the ones that don't;
• other people's experience, what a colleague tried that worked;
• what you pass on to them: your process, your arguments, your limits;
• and their own attempts, half of which lead nowhere.

Nobody finds this remarkable. No company says "our salesperson is trained, we're done, nothing more to change". A company that never adjusts anything doesn't grow, it survives.

So why accept exactly the opposite for the one team member who meets the most people?

Beware the false meaning of "a living site"

Precision matters here, because the phrase gets thrown around, often by people selling maintenance contracts.

A living site is not a site whose text gets rewritten every quarter. Rephrasing a paragraph, swapping a photo, adding a news item: none of that produces anything. Frankly, nobody cares.

A living site is a site you optimise. Considerable difference. You don't change things to change them, you change them because you observed something and you're trying to fix it.

Which assumes you observed something. And therefore that you measure.

A business card waits in a pocket. A salesperson learns. Your site is the second, and you treat it like the first.

A salesperson is judged on numbers

Second piece of the reasoning, and without it the first is worthless. You can't improve what you never look at.

"I get the feeling it works better"

In plenty of companies, sales performance is judged on feel. "I get the feeling it goes better when I lead with the price." "People seem more engaged since we started showing the workshop."

That isn't ridiculous. In contact trades, a good salesperson's instinct is worth a great deal, and some companies have structured all of it very well. It works.

But on a website you don't have that option. You're not in front of anyone, you can't read a face, you can't hear a hesitation. Feel doesn't exist. All that remains is what you measure.

And the good news is that everything can be measured: how many people arrive, through which door, what they click, where they drop out, how many come back.

No, it won't be perfect

Let's defuse the objection now, because it serves as an excuse for a lot of people.

Between consent, blockers and browser protections, some of your visitors escape measurement. The numbers are incomplete, and they'll keep getting more so.

That changes nothing about the method. You don't need an exact count, you need trends. Knowing that seven people in ten always abandon at the same step is enough to decide what to fix. Anyone waiting for perfect data never decides anything.

What that looks like in practice

The method isn't mysterious, and I've detailed it elsewhere: your site is the only salesperson you can measure without them objecting. In summary:

• Write the site's objective in one sentence. One.
• Break the path to it into visible steps.
• Make sure each step leaves a trace.
• Look at where it breaks, and form a hypothesis.
• Fix one thing, then wait and see.

It's exactly the work you'd do with a salesperson, except that here the conclusions aren't up for debate. Without that cycle, your site goes back to being what it should never have been: a business card online.

You don't need exact numbers. You need to know where people stop, and that is always visible.

Without volume there is nothing to judge

Third piece, and the one that's always forgotten. You install the tools, you wait, and nothing comes. The tool isn't the problem.

A salesperson who meets ten people a month

Back to the comparison, it's a telling one.

You hire a salesperson and they meet ten people over the month. Two sign. What can you conclude? Absolutely nothing. Are they good? Does their pitch work? Should the argument change, or the prospect list? Impossible to say on ten meetings.

That's exactly the situation of a site with no traffic. The measurement tool is installed, everything is properly tracked, and it says nothing. Not because it's badly configured, but because it has nothing to analyse.

Add to that something always underestimated: a visit is not a meeting. In person you pick up body language, questions, silences. On a site you pick up a click or the absence of one. The signal is vastly weaker, so you need vastly more of it.

Having visitors isn't enough

Here's the most important point in this article, and the least understood.

There's the traffic you endure and the traffic you decide. They're not the same, and they can't be analysed the same way at all.

Traffic that arrives on its own is outside your control. You don't know why those people came, what they were after, in what frame of mind. Very different intentions blend into the same figure. For a clear trend to emerge, you need considerable volume.

A campaign is the opposite. You choose who comes, with which message, onto which page. The result is readable almost immediately, because you set the conditions of the experiment.

It's the difference between a salesperson at the front desk taking whoever walks in, and one who chooses their meetings. The first works, and even learns things. The second learns ten times faster because they know who they're talking to.

In other words, an advertising budget doesn't only buy visits. It buys understanding. And that's often the most durable thing it returns.

The circle that sets in

With those three pieces in place, the mechanism appears, and it turns both ways.

The right way: controlled traffic produces data, the data reveals where things break, the fix improves the journey, and a better journey improves what the traffic returns. Then round again, slightly better each time.

The wrong way: no traffic, so no data. No data, so no reason to touch anything. No changes, so no progress. No progress, so no appetite for investing in traffic.

The site is then perfectly static. It works, it's online, and nothing happens. It's the most expensive situation there is, precisely because it's invisible.

A site with no traffic isn't a site to improve. It's a site where you have no way of knowing what to improve.

The chessboard

All that's left is why I'm telling you this in one piece rather than three separate articles. Because these three parts can't be worked on separately.

Every piece has a role, only the combination counts

Digital resembles a chessboard. You have pieces: the site, the journey, the offer, the campaigns, the posts, the reviews, the measurement. Each has its own move and none wins the game alone.

A beginner looks at their pieces one at a time. They move the bishop because they can, the rook because it's free. A player who improves looks at the board: what this piece opens up, what it exposes, what it sets up two moves ahead.

That's exactly why an isolated action is almost impossible to evaluate. You rebuild a page, but traffic hasn't moved, or you launched a campaign at the same time. You'll never know what produced what.

Avoid the ten-step recipes

A direct corollary, and it needs saying plainly: be wary of guides promising a method that works every time.

There isn't one. What works for your neighbour may fail for you, for perfectly valid reasons: your sector, your area, your seasonality, your reputation, your customers. Step-by-step lists are reassuring because they suggest a marked path. There isn't one of those either.

And I hold myself to the same rule: don't take what I write here as absolute truth. It's what I observe, often, and it remains to be verified in your case.

The good news: bad moves can be recovered

Digital has one enormous advantage, and nobody ever mentions it.

You'll get things wrong. You'll try things that lead nowhere, run a campaign that doesn't take, rebuild a page that converts worse. That's normal, it happens to everyone, myself included.

And it doesn't matter, on one condition: that you saw what happened. A bad move you've identified can be reversed the following week. Unlike a badly chosen premises or a badly bought machine, a digital decision can be undone.

It's precisely mastery of the whole that makes that possible. When you can see the board, a mistake is just one more piece of information.

What I am not saying

An honest clarification, because this is my trade and I'd rather state it myself than let you guess it.

I'm not saying nothing can be done without ongoing support. Of course it can. Plenty of things get sorted alone, with no budget, using common sense: a clearer page, a shorter form, a visible action on mobile.

What I am saying is that some questions can't be handled in isolation. Knowing where you're losing money, choosing between three possible fixes, understanding why one thing worked and another didn't: that requires seeing the whole board, not one square.

A website is never finished, because a salesperson is never trained once and for all.

If you'd like to look at your board together, write to me. We identify the piece that's blocking, and it's usually just one.