Your Website Gets Traffic and Nobody Calls: Where the Leads Leak

We wrote recently about the site nobody can find. This is the other call, and it comes in just as often. The site is found. Analytics shows people arriving. The phone still doesn't ring, and the owner has now been told by three separate people that what they need is a redesign.

Sometimes they do. Usually they don't, because "traffic but no leads" isn't one fault. It's about six, they're fixed in six different places, and four of them cost nothing. One of them isn't a website problem at all, and it's the one we find most often.

Start with the arithmetic, because a fair number of these calls end there.

Do you have a conversion problem at all?

Published conversion rates for local service sites sit somewhere between one and three percent of sessions, with a well-built page catching a high-intent search reaching five or six. The sources disagree enough that we treat those as orders of magnitude rather than targets. Put your own traffic against them.

Two things fall out of that table, and both are worth sitting with.

At sixty sessions a month, a month with zero enquiries is unremarkable. It isn't evidence of anything. You don't have a conversion problem, you have a traffic problem, and every hour spent rewriting the home page is an hour spent on the wrong one.

And a conversion fix is only measurable when there's enough traffic to measure it. Going from one percent to three on sixty sessions is one extra enquiry, which you can't tell apart from a good week. On a thousand sessions it's twenty, which you can feel. Below a few hundred sessions a month, work on getting found. Above it, everything below applies.

First, check that nobody is calling

This is the first thing we look at on any audit, because when it's the answer, everything else is wasted work. It's the answer more often than anyone expects.

Fill in your own form. Not from the desktop you built the site on, and not on the office wifi. From your phone, on cell data, the way a stranger does it. Use an address you can actually open.

Then go and find where it landed. The junk folder. A forwarding rule pointing at somebody who left in March. Or the modern version: the form sends through your host's mail server, your domain has no SPF or DKIM records, and Gmail drops the message without bouncing it. That one fails silently in both directions. The visitor sees "thanks, we'll be in touch" and believes it.

Count your missed calls. Most business lines will tell you how many came in and how many were answered. A missed call at eleven on a Tuesday from a number you don't recognise is a lead, and an owner losing twenty a month usually has no idea, because a missed call leaves no trace anywhere they're looking.

Listen to your own voicemail greeting. If the mailbox is full, that's your answer. If the greeting doesn't say when you'll call back, most people won't leave one.

Where a visit leaks

Leak one: the page answers a different question than the one that brought them

Somebody searches "emergency plumber Mississauga" at ten at night. They land on a home page whose first screen is a photograph of a van and the words "Quality Service Since 2004". Nothing there says you come out at night, how fast, or roughly what it costs. They're back on Google in four seconds, and your analytics records a visit.

The fix is the first hundred words, and it's free. What you do, where you do it, what happens next. The rest of the page can stay exactly as it is.

It gets worse when one page is carrying every search. If people reach your home page from five different queries, that page can't answer all five, and trying makes it answer none. Open Search Console, take the three queries that matter commercially, and make sure each has a page that opens by answering it.

Leak two: nothing on the screen makes a stranger believe you

A visitor who has never heard of you is doing one thing on your site, and it isn't reading your mission statement. They're deciding whether you're real.

What settles that is boring. A review count and rating they can go and check. The name and face of the person who'll turn up. A street address. Photographs of work you actually did, rather than the stock photo everyone in your trade is using. We publish our own Google rating with all five reviews printed underneath it, and five is not many. We publish it anyway, because a number somebody can verify beats an adjective they can't. Getting those reviews is its own job, and it's the highest-return unpaid work a local business has.

Leak three: the ask isn't where the decision happens

The decision to call doesn't happen at the bottom of the page. It happens wherever the visitor becomes convinced, and you don't get to pick where that is.

So the ask goes everywhere. A tappable phone number in the header of every page. Another after the section that does the persuading, usually pricing or the work itself. On a phone, for a trade or a clinic, a call button that stays on screen. One action, repeated, in the same words each time.

The failure we see most is a long and genuinely good service page with a single contact form under three thousand words. Everyone who was convinced at word eight hundred left at word nine hundred.

Leak four: the form asks a stranger for more than they'll give

Every required field is a place to stop. The trade is between what you'd like to know and how many people finish.

Our own contact form asks four things: name, email, phone, and what you're looking to build. Our quote form asks six, adding budget range and enquiry type, and that difference is deliberate. The quote form is for somebody who has already decided to ask for a quote and expects a question about money. Putting "budget range" in front of a first-time visitor who was still deciding whether you're real costs you the visitor.

So: no "how did you hear about us" on first contact. No company name for a residential trade. No dropdown of twelve services where a sentence would do. If a field exists to help you qualify leads faster, ask it on the phone call you just earned.

One exception. A field that is the thing they came to do isn't friction. A preferred date and time on a clinic booking form is the point of the form.

Leak five: it's slow, or it's awkward in one hand

Google's threshold for Largest Contentful Paint is 2.5 seconds, measured at the 75th percentile of real loads and split between mobile and desktop. That last part is the part people miss. Your site is fast on your laptop on office wifi. The number that counts is the slowest quarter of your visitors, most of them on a phone, some of them on three bars in a parking lot.

On a designed site the usual culprit isn't images, it's the scroll animations. Sections that fade and rise as they come into view look considered, and they also mean the largest thing on the screen is deliberately not painted until a script decides it's time. We've had that argument with ourselves about our own site and lost it.

The rest is mechanical. Tap targets big enough for a thumb. Text readable without zooming. A menu that opens on the first tap. Numbers that dial when touched.

Leak six: you answer too late, and this is the big one

Harvard Business Review published a study in 2011 that audited 2,241 US companies by submitting test enquiries and timing the first response. Firms making contact within an hour were around seven times more likely to qualify the lead than those who took longer, and more than sixty times more likely than the ones who took a day or more. Twenty-three percent never replied at all.

It's a fifteen-year-old study of B2B sales teams and the specific multiples aren't your multiples. The mechanism hasn't aged, because it was never about sales technique. Someone with a leaking tap or a sore tooth contacts three businesses inside the same ten minutes. Whoever calls back first is talking to a person who still has the problem. Everyone after that is calling somebody who has already booked.

This one is free, and it's why we'd rather fix your response time than your home page if we only get one of them. An auto-reply isn't a reply, but an honest one carrying a real phone number and a real window will buy you the hour.

An hour of checks, in this order

1. Open analytics and write down last month's sessions. Under a couple of hundred, stop here and go and read the indexing article instead.
2. Submit your own contact form from your phone, on cell data. Find where it landed.
3. Check the junk folder and the forwarding rules on the address it goes to.
4. Count last month's missed calls. Listen to the voicemail greeting.
5. Open your busiest page on a phone and time how long until the main heading is readable.
6. Read its first hundred words. Do they say what you do, where, and what happens next?
7. Count the required fields on your form. Delete the ones that exist for your convenience.
8. Write down how long it took you to answer the last five enquiries. That number is usually the whole article.

When it's none of the above

Sometimes the traffic is real, the site is fine, the replies are fast, and it still produces nothing. Two causes left.

The traffic is the wrong traffic. You rank for something informational, people arrive wanting an answer rather than a supplier, and they get it and leave. That shows up as decent sessions against queries starting with "how to" or "what is", and it's a targeting problem rather than a design one.

Or the offer is the problem. Your price, your availability, your area. A website can't fix being the most expensive option with the longest wait, and it shouldn't be asked to. That's a conversation about the business, and it's a more useful one than a redesign.

If you'd rather someone else did the reading

The checks above are meant to be enough on their own, and most owners can work through them in an afternoon. Where we come in is when the numbers say one thing and the site appears to say another, or when the fix is structural: a page that has to be built, a form that has to be rewired, a speed problem that turns out to be the theme it was built on.

That's website design and development work, and it starts with the same hour of checks. If you'd like company while you do them, book a call and have analytics open before it starts.