The question I get asked most often about a website, once it has been live a while, is whether it is working. It is a fair question and a difficult one, because the number that is easiest to look at — how many people visited — is close to useless on its own.
Two hundred visits is either very good or slightly worrying depending on who those people were and what they did next, and the visitor count cannot tell you which.
Decide what the website is for first
This sounds like a preamble and it is actually the whole job. A website can be doing any of several different things, and they are measured differently:
- Generating enquiries from people who did not know you existed.
- Confirming credibility for people who already have your name from a recommendation, a van or a sign, and are checking you are real before they ring.
- Answering questions so you spend less time on the phone repeating your hours, your area and your prices.
- Taking bookings or payments directly.
Most small business websites are mainly doing the second one, and that is the hardest to see in any analytics package, because success looks like somebody reading two pages and then phoning you. If you judge that website by enquiry form submissions, you will conclude it is failing while it quietly does its job.
Write down, in one sentence, what you want the website to do. Every number below is only meaningful once that sentence exists, and a lot of disappointment with websites comes from measuring one against a sentence somebody else would have written.
The four figures worth watching
1. Enquiries, counted by route. Form submissions, phone taps and email taps, kept separately. Taps on a phone number or email link can be recorded as events in your analytics — it is a small piece of setup and it is usually the difference between seeing a third of your enquiries and seeing most of them. Count actual enquiries too, including the ones that arrive by phone, even if that means a tally on a notepad.
2. Where people arrive from. Broadly: search engines, direct (typed or bookmarked), referrals from other websites, and social. The mix tells you which of the jobs above your website is really doing. Heavy direct traffic usually means people already know your name — that is the credibility job, and it means the website should be optimised for reassurance rather than discovery.
3. Which pages people land on. Not which pages are most visited overall, which will always be the homepage. Which page did they arrive on from outside? That tells you what people are actually looking for, and it is frequently not what the website is built around.
4. What you appear for in search. Google Search Console is free, takes a few minutes to connect, and is the only place you can see the actual searches that showed your website, how often you were shown, and how often you were clicked. If you look at one tool, look at this one. A page shown a thousand times and clicked twenty is telling you something specific about its title and description.
The numbers that mislead
Bounce rate. Traditionally, the share of people who looked at one page and left. On a small business website that is often the sign of a job done well: they landed on your contact page, got the number, and rang you. A high bounce rate on a contact page is a success being recorded as a failure.
Time on page. Measured by comparing one page view to the next, so the last page of any visit has no reliable figure at all. It also cannot tell engrossed from confused.
Visitor numbers month to month. This is the one that causes the most unnecessary worry. A website getting a couple of hundred visits a month will swing by twenty or thirty per cent for no reason worth investigating — one shared post, one quiet week, one bank holiday. Compare quarters, or compare the same month last year. A month-on-month drop of fifteen per cent on small numbers is noise, and chasing it costs more than it is worth.
Set a reminder to look at this quarterly rather than weekly. On small numbers, checking more often produces more reactions and no more information.
The check that beats all of it
Ask people how they found you. One question, asked on every enquiry, written down somewhere consistent.
Analytics cannot see a recommendation in a pub, a van on a roundabout or somebody remembering your name from three years ago, and those are frequently the largest sources of work for a local business. Six months of that tally is worth more than any dashboard, and it is the only method that measures the whole picture rather than the digital slice of it.
It also tends to correct the instinct to spend more on the channel that is easiest to measure, which is not necessarily the one producing the work.
A note on consent
Analytics cookies are not strictly necessary for a website to function, which under the UK rules means you need consent before setting them, and a banner that only offers "accept" is not consent. This is worth getting right rather than copying from another website — the requirement is real, and it is also simpler than it looks. Some analytics tools are designed to work without cookies at all, which sidesteps the question and is worth considering if your needs are modest.
What good looks like after a year
You know roughly how many enquiries the website produces a month and by which route. You know the three searches that bring most people in. You know which page they land on. And you have a tally, in your own handwriting or a spreadsheet, of how people say they found you.
That is four facts. They are enough to decide where the next bit of effort goes, which is the only reason to measure anything in the first place.
If the answer turns out to be that people cannot find you in the first place, the survey I ran of local websites is a reasonable place to start — it is in the write-up here, with the dataset. And a monthly look at the technical side is what SEO Management covers.