How to tell whether your website is working

Somebody sends you a monthly report. It says you had 412 visitors, up eleven per cent, and your bounce rate improved. You look at it for a while, feel vaguely that this is good news, and then get on with your day, because you have no idea what any of it means for whether you'll have work in March.

The reason that report is useless isn't that the numbers are wrong. It's that visits aren't what you want. You want enquiries, and the relationship between the two is much looser than anybody selling you a report would like to admit.

Why visits are the wrong headline

A site can double its visitors and produce fewer jobs. It happens regularly, and usually because the extra visitors came for something they weren't going to buy — the wrong town, the wrong service, a question that had nothing to do with hiring anybody.

Visits also fluctuate for reasons nothing to do with you. A quiet August, a cold snap, a competitor's advert, a change somewhere in Google. Reading meaning into a month-on-month wiggle is a good way to make bad decisions.

So park the visitor count. There are three questions that actually tell you whether the site is working, and you can answer all of them in about ten minutes.

Question one: how many people got in touch, and how

Count the enquiries. Not analytics — actual enquiries, from your phone, your inbox, your messages.

Then split them by route: rang, filled in the form, emailed, messaged, walked in, came through your Google listing. You'll need a month or two of counting before the picture's clear, and it's worth the small nuisance because it settles arguments no dashboard can.

Most owners are surprised twice. First by how few enquiries come from the website form compared with the phone. Second by how many people rang after reading the site, which means the site did the work and the phone got the credit. Ask new customers where they found you and write the answer down — that single habit is worth more than any analytics setup.

Question two: does your site answer the questions people ring to ask?

This is the best test there is and it doesn't involve a computer.

Keep a note by the phone for two weeks. Every time somebody asks you something before they'll commit, write it down. Do you cover Sale. What's the smallest job you'll take. How far ahead are you booked. Do you charge for a quote. Is there VAT on top.

At the end of the fortnight, look at the list and then look at your site. Every question that comes up repeatedly and isn't answered on the page is a hole, and every hole is costing you the people who wouldn't ring.

The ones who rang and asked are the minority. They're the evidence; they aren't the loss. The loss is everybody who had the same question, found nothing, and went elsewhere without ever appearing in any statistic you'll ever see.

Question three: can a stranger use it on a phone in a hurry?

Hand your phone to somebody who doesn't know your business. A relative, a neighbour, anyone. Ask them to find out three things: what you do, whether you'd come to their street, and roughly what a typical job costs.

Watch without helping. Time it. If it takes more than about twenty seconds, or if they can't find the third one at all, you've learned more than a year of reports would have told you.

Then ask them what they'd have done next if they actually needed the work doing. The answer is often "I'd have rung someone else first", and the reason why is usually one specific missing sentence.

Two numbers worth watching after all

If you want something you can track over time, there are two that mean something.

Enquiries per month, counted by hand, plotted over a year. Noisy month to month, meaningful over seasons. It tells you whether the changes you make do anything.

Where your name is searched. In your Google listing there's a section showing how people found you, including whether they searched your business name or searched for what you do. That second group is new customers finding you cold, and it's the number most worth growing.

Neither needs any software installed. Both are more honest than a visitor count.

What "working" actually means

Here's a way to think about it that makes the whole thing simpler. Your site works if somebody who needs what you do can, in under a minute, find out that you do it, that you come to them, roughly what it costs, and how to reach you — and then feels comfortable getting in touch.

That's the standard. Not traffic, not rankings, not how modern it looks. Every one of those four steps is a place somebody drops out, and the drop-outs are silent.

The fourth one is where most sites leak, because comfort comes from having your specific question answered, and a page can only hold so many. People have two dozen questions and you've predicted six of them.

That's the gap we built hailera around: you write your answers once, check them yourself, and publish them as an agent, so your site and the AI apps people now ask can answer the awkward specific ones directly instead of leaving somebody to guess. But the phone-note exercise above will show you the six that matter most, and those you can just add to the page for nothing — so start there.

Once a quarter, half an hour

Count the enquiries. Read the phone notes. Get one stranger to use the site on a phone. That's the whole review, and it will tell you more than any monthly report.

Then change one thing, and leave it a quarter before you judge it. The most common mistake after a review like this is changing six things at once and never learning which of them mattered.

Today: put a piece of paper by the phone. For the next two weeks, write down every question somebody asks before they'll commit. That list is your next set of website changes.

Written by the team at hailera — we help small businesses answer customer questions without being there.