Your site is your best salesperson. Nobody reads its numbers.

You measure your salespeople, not your site

The obvious point everyone repeats and nobody applies

You can only improve what you measure.

Everyone knows that sentence and nobody disputes it. A company with a sales team looks at their results. How many meetings, how many quotes, how many signatures, at what average value. It's so widely accepted it doesn't even need defending.

Now that same company has a website. That site works every day, weekends included, and receives more people than any salesperson will meet in a career.

And nobody reads its numbers.

The only seller you can measure without them objecting

There's an irony here I find rather enjoyable.

Measuring a salesperson is delicate. Nobody enjoys being compared, ranked, reduced to a table. You move carefully, you explain, you handle it.

Your site will never say a word. You can test it any way you like, compare two versions, change a button and measure the effect, then start again tomorrow. It won't take offence, it won't resign, it won't feel surveilled.

It's the easiest member of the team to evaluate. It's the only one nobody evaluates.

A site has an objective, and objectives get counted

Every site has one main job. For a brochure site it's almost always getting an enquiry or a quote request. For an online shop it's selling. For a booking system it's filling a slot.

That job can be counted. How many enquiries this month, how many last month, how many closed. If you can't answer those three questions, you don't know what your site earns you, and you'll never be able to judge what's worth investing in it.

Every marketing action is an investment, so it gets tracked

You pay a videographer several thousand euros for a film. Fine, often a good call. But it's an investment, not a decorative expense.

An investment gets tracked. Where was it seen, what did it send to your site, what happened next. Same for a campaign, a print insert, a trade show, a flyer.

Everything has a cost, so everything deserves a look at what it produced. Otherwise you're not doing marketing, you're spending and hoping. And that's exactly how you end up discussing an advertising budget without being able to say what the last one returned.

Making things traceable, even as everything closes

No, it isn't perfect any more. And that's fine.

Let's be honest about where things stand. Between mandatory consent, blockers, the gradual disappearance of third-party cookies and browser-level protections, a share of your visitors escapes measurement entirely.

Plenty of companies draw the wrong conclusion: since it's imperfect, why look at anything.

That's an accountant's reasoning, not a decision-maker's. You don't need exact numbers, you need trends. Knowing that your quote page converted better last month, or that seven visitors in ten abandon at the same step, is more than enough to decide what to fix. Anyone waiting for perfect data never decides.

One operation, one page, one link

The best way to know where people come from isn't technical, it's organisational: give every operation its own front door.

A concrete case. You buy space with a local outlet, or you sponsor an event. The reflex is to give them your website address. From that moment everything blends together and you'll never know what it produced.

Do the opposite. Build a page dedicated to that operation, with its own address, its own offer, its own form. You give that link, and only that link, to the partner.

• The page is only reachable through that medium, so everything landing there came from it.
• You need no sophisticated tracking to know it.
• The page's message repeats the promise made in the medium, so it converts better.
• When the operation ends, you know what it cost and what it returned.

That's exactly what a dedicated landing page is for, and it's the only reliable way to measure an offline medium.

From the “.tv” of TV ads to today's QR code

None of this is new. It has simply changed shape.

In the early days of the web, advertisers running television or radio spots bought a dedicated subdomain. You'd hear “visit mycompany dot t v”, repeated twice to be sure. People typed that address, landed on a page built only for them, and the advertiser could watch the traffic spike at the exact hour of the spot.

It wasn't perfect either. It was infinitely better than nothing.

Today the trick works less well: search engines pull everyone back to the main domain. But the principle holds, and it has an obvious successor. The QR code.

They're everywhere now, on menus, posters, business cards, vans. Most of them point dumbly at the homepage. Point them at a page of their own and every medium becomes measurable at no extra effort.

A site is designed to be measured

This is the heart of it, and it's what almost nobody does. Measurement isn't bolted on afterwards. It's decided at the moment you draw the journey.

A single page leaves you blind

You often hear that everything should sit on one page. A complete form, every field visible, no intermediate step.

The problem is that this format teaches you nothing. The visitor arrives, leaves, and you don't know why. Put off by the length? Stuck on a field? Cooled by a question that felt intrusive? Impossible to say. You have two facts: they came, they didn't send.

And a long list of fields is intimidating at a glance. You get both drawbacks at once.

Splitting into steps: the double benefit

Take a real example: equipment hire for events. The form asks for a date, a type of equipment, then contact details.

Instead of showing it all at once, you split it:

• Step 1: the date. One question, no commitment.
• Step 2: the equipment. You move forward, the person is already inside the journey.
• Step 3: contact details. Asked once the enquiry exists.

The first benefit is psychological: three short questions get crossed without thinking, where the same thing shown as a block discourages. What drives people away isn't the number of steps, it's the feeling that this will take ten minutes.

The second benefit is the one that matters here: you see exactly where people drop out. How many arrive, how many clear step 1, how many reach step 3. You're no longer measuring a result, you're measuring a journey.

What you do once you can see the drop-off

Once you can see which step breaks, the work becomes concrete. You ask simple questions and you try things:

• Is information missing at that exact moment, blocking the decision?
• Does the question come too early in the journey?
• Can two steps be merged, or should one be split?
• Is that field genuinely necessary, or was it added just in case?
• Should the order be reversed so commitment is asked for later?

These corrections look trivial. Their effects aren't. Removing one field or moving one question can take a step from where it was to five times better. That isn't an exaggeration, it's what you observe as soon as you look.

Which brings back a rule I repeat often enough: nobody will make the effort for you. Every friction you leave in place is paid for by someone, and that someone is you.

A tool you control makes you master of the data

This is an argument I use with clients, and it isn't a sales pitch, it's practical.

When I build a bespoke booking system, I know everything that happens inside it without relying on an external service and without needing to follow anyone. How many people reach the page, how many open the module, how many clear each step, where it jams. I don't need to identify the person, I need to understand the journey.

That step-by-step precision is something most subscription products don't give you. They're built for the majority, which is entirely legitimate, but they show you what they've decided to show you. That's the whole debate between the standard tool and the tool that belongs to you.

With a tool you control, you spot a redundant click and remove it the same day. And gaining 20% on one step doesn't mean 20% more clicks: it can mean twice as many enquiries at the end of the chain.

The arithmetic almost nobody does

From €100 to €20 per enquiry

Let's use simple numbers, because this is where everything is decided.

• You spend on advertising and each quote request costs you €100.
• You work on the journey, nothing else. Same budget, same channel, same targeting.
• The cost per enquiry drops to €30, sometimes €20.
• You now get three to five times more enquiries for the same money.

So no, you won't do five times more jobs. You have neither the time nor the people, and some of those enquiries will go nowhere.

But twice as many jobs, with no budget increase, no change of agency, no new strategy: that's entirely achievable. And twice as many jobs is a different accounting exercise altogether.

The detail that saves a company

I'm using the word “saves” deliberately.

A company that divides its acquisition cost by three doesn't make a small improvement. It changes trajectory. It goes from a business that struggles to one that fills its schedule, on the same budget with the same people. In a tight market, that's the difference between holding on and closing.

It's arithmetic to the point of seeming unreal. And that's precisely what makes it maddening: these are details only large companies used to know, because they had whole teams paid to watch them.

Today the tools and the methods are available to anyone. Access is no longer the problem. The problem is that nobody knows it exists, or nobody believes it.

What e-commerce has known for twenty years

None of this came out of nowhere. It comes from online retail, where it isn't an option but a condition of survival.

A serious online retailer measures every step of the funnel, tests constantly, and pays for the privilege. Tools costing hundreds a month, developers on daily rates north of €800, brought in for several days to optimise a single point of the journey.

That isn't excess, it's arithmetic. Half a point of conversion rate can mean 25% more annual revenue. At that level the spend justifies itself.

And this is where the contrast gets uncomfortable. A local business will hesitate to put €100 a month into a tool that would show it where it loses customers, while an online retailer of comparable size pays for several without a second thought. The first has no less to gain. It simply doesn't know the question exists.

An online shop that doesn't measure doesn't survive

I'll put it bluntly because it's verifiable.

Among the online shops that disappear, an overwhelming majority had no serious tracking. No measured funnel, no known conversion rate, no testing. They pushed traffic into a system whose leaks they couldn't see, until the budget ran out.

A brochure site doesn't die that fast, because it doesn't carry the revenue on its own. Which is exactly why it's left to drift for years without anyone ever opening it.

Having data isn't enough, you have to read it

Apples and oranges: context changes everything

Here's the mistake that cancels out all the previous work: comparing numbers that can't be compared.

A page fed by Google Ads and the same page fed by Meta or TikTok will never share a conversion rate. Not because one is better, but because the intent isn't the same. On one side, someone actively looking for a solution. On the other, someone scrolling a feed who stopped out of curiosity.

Both numbers are accurate and they don't mean the same thing. Putting them side by side to make a decision is comparing apples and oranges, and guarantees a wrong call.

A number can only be read with its context: where the traffic came from, over what period, on which version of the page, in what kind of campaign. That context is the difference between data and information.

What the numbers tell you that you'd never otherwise see

Here's something I see regularly.

A visually elaborate site, animations everywhere, hover effects, elements that reveal themselves when the cursor passes over them. On the designer's screen it looks great.

Except hover doesn't exist on mobile. There's no cursor. Every piece of information hidden behind those effects is simply invisible to the majority of your visitors. Nobody notices, because nobody tests, and because people keep looking at their site on a desktop.

Another very concrete example. You advertise on Google for local plumbing. People are actively searching for a plumber nearby, intent is as high as it gets. If out of a hundred people reaching your quote page only one starts the form, the market isn't the problem. Something is wrong with that page, and it needs finding.

Without data you'll see neither. You'll conclude that “advertising doesn't work”, and cut the budget that was working.

The healthy obsession: your own number from last month

Don't go looking for a benchmark conversion rate online. Averages blend sectors, channels and maturities that have nothing to do with you, and you'll come away either falsely reassured or needlessly discouraged.

The only useful benchmark is your own number from last month. You should know how many visitors request a quote, and you should want that number to go up. It's a healthy obsession, and it costs nothing.

One last precaution, the one always forgotten: don't change three things at once. You wouldn't know which produced the effect. And accept that volume comes before conclusions, exactly as it does for an advertising campaign. On twenty visits, all you're observing is chance.

Where to start

None of this requires a large-company budget. It requires deciding that the subject exists.

• Write your site's main objective in one sentence. One.
• Break the path to it into visible steps, and make sure each step leaves a trace.
• Give every operation its own page and its own link, online and offline alike.
• Look at those numbers once a month, always the same ones, always in the same place.
• Fix one thing at a time, and wait for enough volume before concluding anything.

You don't need to understand everything to begin. You just need to stop working blindfolded.

Your site is the only salesperson you can measure, correct and improve without ever bruising them. It would be a shame to keep ignoring it.

If you want to know what yours actually produces, write to me. We'll look together at where it loses people, and I'll tell you straight what's easy to fix and what needs real work.