The report gets opened at the start of the month, somebody checks whether the number is bigger than last time, and it gets closed again. It is either bigger or smaller, that is where it ends, and next month goes the same way.
Almost every website has statistics on it. The question somebody is trying to answer with them is usually missing. The fault is not in the tool.
The visit count is the emptiest number there is
The visit count is the first thing every report shows, and the only one most owners read out loud. It moves almost entirely for reasons that have nothing to do with your business.
One share on Facebook. One robot that read your site all night. One piece of news in your field. Two weeks when the whole country was on holiday. A monthly number going up and down is usually not news, it is noise.
The visit count answers the question of how many people came. Your question is whether more work came in. The second does not follow from the first.
The number that actually counts
For a small website one figure matters more than all the others: how many enquiries you get per hundred visits.
It needs no new tool and no configuration. Take the month's visit count from your statistics. Count that month's enquiries by hand: emails, saved form submissions, calls. Divide the enquiries by the visits and multiply by a hundred. Two minutes of work.
That number has two useful properties. It does not move with your traffic, so a quiet summer cannot spoil it. And it can be improved without buying a single extra visitor.
If traffic grows and that number falls, you got the wrong visitors. If traffic is flat and that number rises, the site is doing something right.
Bounce rate does not say what you think it says
The most misread number on any website. Bounce rate does not measure boredom or disappointment. It counts the visits during which the person did nothing the measurement can see: opened no second page and clicked no measurable button.
Somebody who read your contact page, memorised the number and called you is a bounce in the statistics. They were your best visitor that day.
That is why a high bounce rate on a contact page, a price page or an opening-hours page is fine: the person got what they came for. It is worth worrying about when the bounce rate is high on a page whose job was to send the reader onwards.
"Direct" does not mean direct
Traffic sources are the most useful part of the report, but only if you read them honestly.
"Direct" is supposed to mean that somebody typed your address into a browser themselves. In practice it collects everything whose origin the measurement failed to recognise: clicks from email, clicks from Messenger and WhatsApp, links opened from a PDF, and QR codes.
So a large direct share does not mean your brand is strong. It usually means part of your business happens in channels the statistics cannot see. The only way to find out is to ask.
The most read page and the page that brings enquiries are not the same page
Every report can show which page was viewed the most. Almost always it is the homepage or some article that Google happens to like.
In our experience the most read page is the one that satisfies curiosity, and the enquiry arrives through some unremarkable service page with several times fewer visitors.
The practical conclusion: the top of the page-views table does not tell you which page to improve. What tells you is the question "where did you find us", asked of a customer who is already on the phone.
Comparing with last month is almost always noise
By default every report compares this month with the previous one. On a small site that is the most misleading view available.
August cannot be compared with March, because your customers are on holiday, at work or planning a renovation at different times of year. Most small companies have a rhythm across the year that they forget the moment they look at a chart.
Compare the same month with the same month last year. Then the season is the same on both sides, and what remains is what actually changed.
The honest part: most of your numbers are too small
If a couple of hundred people visit your site in a month, most of the statistics tell you nothing at all. Eight enquiries instead of last month's five is not a trend, it is chance. A percentage worked out from thirty visits is not a percentage.
In that situation the better source of data is close at hand and free. Call your last three customers and ask two things: where they found you, and what made them get in touch. Three honest answers tell you more than a month of watching a chart.
We say the same when somebody wants to buy an analytics setup from us: at numbers that small there is no point ordering one from us. A setup brings in no visitors, and a tidier report does not turn chance into a trend. If there is money to spend, it does more on one decent service page, or on keeping the site in order. But the honest answer here is often that you need to buy nothing.
Ten minutes a month
No more than that is needed. Do it in the first week of the month, in five steps.
- Write down the visit count. Write it down, do not interpret it. In a year you will have a row that says something.
- Count the enquiries by hand. Emails, saved form submissions, calls. Write that number next to it.
- Work out enquiries per hundred visits. This is the only number worth following month after month.
- Look for one change among the sources. Whether a source disappeared or a new one appeared. One question, not a whole table.
- Compare with the same month last year. Not with last month.
And at least as important, what to ignore: bounce rate, average time on page, the spikes of individual days, and, as a rule of thumb, any figure worked out from fewer than thirty visits.
If you want to know whether the site is technically sound, our free search engine audit checks that side from the outside. And if your enquiry number is low and you cannot tell why, ask us: we will also tell you when the answer is that your numbers are too small to measure.
Frequently asked questions
For a small website the one that matters is how many enquiries you get per hundred visits. It needs no new tool: take the month's visit count from your statistics, count that month's enquiries by hand (emails, saved form submissions, calls), divide the enquiries by the visits and multiply by a hundred. That number does not move with your traffic, and it can be improved without buying a single extra visitor. The honest limit: if a couple of hundred people visit in a month, most figures are within chance, and the better source is your last three customers, asked where they found you.
Bounce rate does not measure boredom or disappointment. It counts the visits during which the person did nothing the measurement can see: opened no second page and clicked no measurable button. So somebody who read your contact page, memorised the number and called you counts as a bounce, even though they were your best visitor. A high bounce rate on a contact page, a price page or an opening-hours page is fine: the person got what they came for. It is worth worrying about when the bounce rate is high on a page whose job was to send the reader onwards.
In theory it means somebody typed your address into a browser themselves. In practice direct collects everything whose origin the measurement failed to recognise: clicks from email, clicks from Messenger and WhatsApp, links opened from a PDF, and QR codes. So a large direct share is not evidence of a strong brand. It usually means part of your business happens in channels the statistics cannot see, and the only way to find out is to ask the person where they came from.
On a small site that is the most misleading view there is, even though every report offers it by default. August cannot be compared with March, because customers are on holiday, at work or planning a renovation at different times of year, and most small companies have a rhythm across the year that they forget the moment they look at a chart. Compare the same month with the same month last year: then the season is the same on both sides, and what remains is what actually changed.
Read next
You type your web address from memory; you have never typed your own company name into Google. Your customer does, and does it last of all, just before calling. What meets them there: your site under a heading you did not choose, the state figures, a competitor advert, an old Facebook page. Who put each row there, what you can change, and how to run that search as a stranger.