Reading Your Website Analytics

Two free tools, answering two different questions

Most small businesses have two measurement tools and assume they are two versions of the same thing. They are not, and that confusion causes more wasted argument than anything else in this subject.

Google Search Console tells you what Google did with you. It reports on your presence in search results: how often you appeared, how often somebody clicked, what they typed, which pages Google has indexed, and whether anything is wrong from Google's point of view. It is a report about the search engine.

Google Analytics tells you what people did once they arrived. It reports activity on your website from every source - search, Facebook, a link in an email, somebody typing your address, the QR code on your truck.

Neither is the traffic number, and asking which one is right is the wrong question. They measure different populations at different moments with different losses, which is why they never agree.

Both are free. Search Console needs no code beyond a one-time verification, takes minutes to set up, and there is no business for which it is not worth having. If the choice is one tool or the other, for most small local businesses Search Console is the more useful.

Why the numbers never agree

The gap between the two is usually presented to a business owner as evidence that somebody made an error. It is not. Here is the actual mechanism, in seven parts.

1. Different populations. Search Console counts clicks on Google search results. Analytics counts page views where the tracking code ran. Neither is a subset of the other.
2. Lost tags. Ad blockers, privacy browsers, refused consent banners, people who leave before the page finishes loading, and ordinary script errors all remove hits from analytics. They do not remove clicks from Search Console. Analytics is therefore usually the lower number.
3. Redirects. A click on a URL that redirects can land in analytics as a different page, or as direct traffic with no source at all.
4. Google only. Search Console knows nothing about Bing, Facebook, direct visits, or the visit button on your Google Business Profile, which arrives in analytics as ordinary traffic rather than as a search click.
5. Different clocks. Search Console reports in Pacific Time and lags by a couple of days. Analytics uses whatever time zone the property was configured with.
6. Privacy filtering. Search Console deliberately omits rare search terms, so the list of queries never adds up to the totals shown above it. That is documented behavior, not a bug.
7. Attribution. Analytics assigns a visit to a channel using a lookback model that can credit an earlier interaction. Search Console credits the click to the click.

Google's own documentation defines the four Search Console metrics precisely and warns that what you see there will not match what you see when you search for yourself, because results depend on time, place, device and the searcher's history. That is worth remembering the next time somebody shows you a ranking screenshot.

What Search Console shows that analytics cannot

This is the case for logging into Search Console rather than only looking at visit counts.

• Impressions. How often you appeared at all, whether or not anyone clicked - the only free source of demand data about your own business. If you appeared nine hundred times and were clicked twelve times, the problem is your title and your reviews, not your position.
• The actual words people typed. Analytics has not shown search terms for well over a decade. Search Console is where those words live, and for a local business they are frequently surprising: people search for the thing you do, not the name of your business.
• Average position, and which way it is moving. Treat the absolute number lightly and the trend seriously.
• Indexing status. Which pages Google has, which it does not, and why. The message crawled but currently not indexed is the most useful diagnostic in small-business search, and nothing else will tell you.
• Manual actions - whether a human at Google has penalized the site. There is no other way to find out, and it matters most for businesses that once hired somebody to build links.
• What Google actually sees when it renders a page, through the URL inspection tool.

Google's performance report documentation defines all of this precisely. One recent addition: since 3 June 2026 Search Console has reported impressions from Google's generative AI experiences alongside ordinary results. As launched it is impressions only - no clicks, no click-through rate, no query detail - so treat it as a rough indicator of visibility rather than a measure of traffic.

The four or five numbers actually worth looking at

For a business with modest traffic, a useful report is a handful of numbers and one sentence about what changed. Everything else is decoration.

1. Phone calls. For a trade, a marina, a clinic or a restaurant, the conversion is a call, not a session. Calls are invisible to website analytics unless the tap-to-call link is instrumented or a tracking number is used. If your report measures sessions and ignores calls, it is measuring the wrong thing entirely.
2. Form submissions, and whether they arrived. Count them, and separately confirm the messages actually land in a mailbox somebody reads. A form that silently stopped delivering is a common and expensive failure.
3. Google Business Profile activity - calls, direction requests, website taps. For many businesses in this market this is the bigger channel by a wide margin, and it lives in a different report from your website analytics entirely.
4. Search Console clicks, with impressions beside them. Clicks alone tell you what happened. Clicks against impressions tell you whether the problem is being found or being chosen.
5. Which handful of pages get any traffic at all. At this size it is usually the homepage, one service page and the contact page. That is genuinely useful, because it tells you which pages are worth spending an afternoon improving and which nine are not.

Notice what is absent. Bounce rate, time on page, sessions per user and the rest are not wrong, they are simply too noisy to act on at low volume, and they lead to decisions with no evidence behind them.

At forty visits a month, read months, not days

This is the part that saves the most wasted effort.

If a site receives around forty visits a month, that is one or two a day. On any given Tuesday you might get five, and on the next Tuesday none. Neither is a signal about anything. Going from forty visits to fifty-five is not a 38 percent increase in business; it is the ordinary variation you would expect from a coin landing differently. Present it as a percentage and it sounds like a result, which is exactly why it gets presented that way.

At low volume, the shortest period you can reasonably interpret is a month, and the only comparison worth making is against the same period a year earlier, because seasonality around Cedar Creek Lake is enormous and comparing July to February tells you about the lake rather than about your marketing.

The same arithmetic disposes of a whole category of products. You cannot detect a ten percent improvement on forty visits. Split testing, session recordings and heat maps all need volume a site like this does not have and will not get, and at this scale they produce confident conclusions from noise - worse than having no data.

What a low-traffic site can legitimately learn is four things, and they are yes-or-no answers rather than trends: whether the phone number is being tapped, which few pages get any traffic, whether people find you by business name or by service, and whether anything is broken on a phone. Worth checking. Not worth a monthly presentation.

The mistakes that make a year of data meaningless

Four errors are common enough at this scale to be worth naming, and all four are avoidable.

Analytics installed more than once. A tracking code in the theme settings, plus a plugin, plus a tag manager container, all firing on the same page. The result is doubled page views, an impossibly good bounce rate, and a year of reports that mean nothing. This is the most common measurement fault we find on small sites, usually discovered when somebody wonders why traffic apparently doubled overnight two years ago.

Reporting impressions as if they were customers. An impression means your listing was on a screen somebody may never have scrolled to. It is useful as a denominator and misleading as a headline.

Reporting rankings. Search results are personalized by location, device and history - Google says so in its own documentation - so a ranking screenshot shows what one person saw on one device at one moment. Impressions and clicks are the honest version of that question.

Not measuring calls. Worth repeating, because it most often makes a report actively misleading. A business whose customers all call from a Maps listing can be doing very well while its website analytics look flat, and a report that never mentions calls will describe that business as failing.

One historical note that still catches people out: the previous generation of Google Analytics stopped collecting data in July 2023, and Google's documentation states that access to that historical data ended in 2024 and the data would be permanently deleted. If nobody exported it, a business's website history before 2023 no longer exists anywhere. The only lesson available is to export what you have before changing platforms.

When you do not need analytics at all

A genuinely honest section, and it applies to more businesses in this market than the industry would like to admit.

If your entire customer path runs through your Google Business Profile - people find you in Maps, read the reviews, and tap the call button without ever visiting the website - then website analytics measures almost none of your business. The profile's own performance report and some way of tracking calls are the right instruments. This is extremely common for trades.

Nobody needs a twenty-page report. For a business this size, a useful monthly summary is four numbers and one sentence about what changed. If a vendor's report runs longer than a page and does not contain the word calls, it is being padded.

An analytics retainer is not a product a low-traffic site should buy. At forty visits a month, a five percent improvement is two visits, and nothing at that volume lets you tell an improvement from chance. Paying monthly for analysis of numbers that cannot support analysis is a poor use of money that would otherwise buy content or photography.

What is worth doing, once: verify Search Console, install analytics correctly and exactly once, make the phone number a tappable link, and set a reminder to look quarterly. An afternoon of setup and fifteen minutes a quarter is, for most businesses here, the entire correct answer.