Traffic is up, leads are not: where they actually get lost

A familiar conversation: “We invested in promotion, traffic doubled, and we get the same number of enquiries.” The first instinct is to doubt the traffic quality. Sometimes that's right, but far more often something else is happening: the leads exist, they just don't arrive. The form posts into nowhere, the email lands in spam, the call is never counted, the analytics goal measures the wrong event, and the salesperson replies the next day. Each of these breaks quietly on its own and kills the result collectively. Below is the path from landing on the site to a conversation with a salesperson, broken into steps, with where it usually tears and how to check it in an evening without commissioning a full audit.

Step one: verify the form works at all

It sounds insultingly basic, and it is the most common finding. Forms break silently: after a template update, a change of mailbox, a captcha being added, or a move to new hosting. On the surface everything is fine — fields fill, the button clicks, a thank-you appears. The email goes nowhere.

Each form has to be tested separately, not just the one on the homepage. A typical site has more than you think: homepage, service pages, product cards, a pop-up, the footer, the contact page, a call-back form. I regularly find sites where half the forms point at a mailbox nobody has had access to for years.

Then there's where the email lands. Automated site mail goes to spam readily, especially when it's sent from an address unrelated to the domain. The check is simple: submit a request yourself, wait for the mail, and see which folder it sits in and how long it took. If it took ten minutes, that's already a problem: the customer has written to a competitor by then.

Step two: the calls nobody counts

This is where the most painful part of the statistics disappears. On a phone, a person taps the number — that event can be captured and is usually counted as a call. But if the number is dialled by hand from a desktop, or copied into a messenger, there is no event at all and that customer never appears in any report.

The practical consequence: if a meaningful share of your enquiries arrive by voice, site reports systematically understate the result and decisions get made on a partial picture. The fix is call tracking — a dedicated number for the site — and until it exists, any conversation about “site conversion rate” stays approximate.

The same applies to messengers. A “message us” button hands the person to an app, after which they're outside analytics. If that's your main channel, count button clicks and reconcile them with what actually arrived in the chat — otherwise the analytics picture and the business picture diverge, and both look convincing.

Step three: goals that measure the wrong thing

A classic: the goal is set on visiting a thank-you page, and there is no thank-you page — the form shows an inline message. The goal never fires, the report shows zero conversions, and everyone hunts for a traffic problem. The inverse also happens: the goal fires on a button click rather than a successful submission, and the report shows attractive numbers behind which there is not one actual enquiry.

The check takes five minutes: submit a request and see whether the conversion registers. With several forms, do it for each — I've seen sites where the goal was attached to one form out of forty and the owner sincerely believed the site didn't work.

Another trap is the counter missing from some pages. After a new section is added, or part of the catalogue moves to a different template, it's easy to miss. Part of the customer journey then becomes invisible, and analytics confidently reports that people leave at a point where they are in fact still going.

Step four: the path to the form

Once the technical layer is sound, the substantive problems start. The most common is that people can't tell where to go. The catalogue is organised around the company's internal logic rather than the buyer's; a service page carries no price and no answer to “roughly what does this cost”; reaching the enquiry takes four clicks.

Second is the first screen. If the first seconds don't convey who this company is, what it does and for whom, nobody goes further. That hurts most on mobile, where the first screen is small and half of sites still fill it with a large image and no text.

Third is trust. No prices, no photographs of real work, no clear contacts or address, no answers to the obvious questions — lead times, delivery, guarantees, how to pay. Nobody writes “I don't trust you”; they close the tab, and in analytics it looks like an ordinary bounce.

Step five: what happens after the enquiry

The enquiry reaches a salesperson — and more is lost here than people assume. Response speed decides almost everything: the difference between replying in five minutes and replying the next day is the difference between a deal and a polite “we've already found someone”. For a site you are paying to promote, a next-day reaction devalues the entire budget.

Second, where the enquiry lands. If it arrives in a shared mailbox checked “when there's time”, some of it is physically lost. A CRM is needed here not for elegant funnels but for one plain thing: every enquiry becomes a deal with an owner and a status, and losing it silently stops being possible.

Third, what happens to those who didn't buy immediately. In most sectors the decision isn't made on the first visit. If a customer vanishes for good after the first conversation, you pay for that traffic twice. A simple sequence of follow-ups to people who left contact details usually pays back faster than any work on traffic.

The order to fix things in

Order matters, because the cost of these steps differs by an order of magnitude. First the technical check: forms, email delivery, goals, counters, call tracking. It takes a day or two, costs almost nothing, and regularly recovers enquiries you already paid for with traffic.

Then response speed and routing: who replies, how fast, where the enquiry lands. That's organisational work requiring no site changes, and it usually yields more than any design edit.

Only then the substantive layer: structure, first screen, prices, trust, follow-ups. It's slower and costlier, so starting there before the first two levels are closed means paying for the hard thing while the cheap thing is undone. And only after all of that does scaling traffic make sense: while the path leaks, more visitors simply means more loss.