Website analytics for small businesses: what actually matters

The analytics tools most small businesses end up with were designed for companies with a data team: someone to configure them, someone to build the reports, and someone whose actual job is to look at the reports on Monday morning. If you are running the business, the website and the marketing, you have none of those people, and the tool does not adapt — it just presents you with forty screens and waits.

So the usual outcome is a dashboard nobody opens, a vague sense that traffic is “fine”, and a cookie banner annoying every visitor in exchange for data you never use. That is a bad trade, and it is worth fixing even if you change nothing else.

This post is what we would tell someone who asked us at a dinner party, in the order we would tell them. There is a product at the end of it, and you can ignore that part.

1. Is it going up? Not this week against last week — that is mostly noise. This month against the same month last year, or a rolling quarter. One number, checked rarely.
2. Where do people come from? Search, social, someone else’s link, an AI assistant, or typed straight in. This is the number that tells you which of your efforts is doing anything at all.
3. What do they do when they arrive? Specifically: do they do the one thing you want — enquire, book, buy, subscribe. If you have never defined that thing, do that before you read another report.
4. Which pages do the work? Almost every small site has two or three pages doing nearly all the useful business, and they are usually not the pages that got the most attention.
5. What broke? A form that stopped submitting, a page that got slow, traffic that fell off a cliff. Analytics earns its keep on the day it tells you about a problem you had not noticed.

That is the list. If a tool answers those five without a project, it is a good tool for you, whoever makes it.

Say twelve people arrived from LinkedIn last month and one of them enquired. Your dashboard will cheerfully print 8.3%, and you will compare it to the 4.1% next to it, and you will conclude something about LinkedIn.

You should not. With twelve visitors, one extra enquiry would have made it 16.7% and one fewer would have made it 0%. The number is not measuring LinkedIn; it is measuring which twelve people happened to turn up. Do this across a dashboard and you will spend real money reallocating your effort on the basis of coin flips.

The honest response is to refuse the division rather than dress it up. Coruve will not print a rate for a row with fewer than 30 visits — it shows a dash and the reason instead. That is a deliberately unimpressive piece of design, and it is the one we would most want a small-business owner to copy into whatever tool they use: before you act on a percentage, look at the count underneath it.

The most common mistake after “never opening it” is opening it every day. Daily checking on a small site is nearly all noise, and it trains you to react to randomness — you will start attributing Tuesday’s dip to something you did on Monday, and you will be wrong most of the time.

A schedule that suits most small businesses:

• Once a month, for twenty minutes. The five questions above, this month against last year. Write down one thing you will do differently. That is the whole ritual.
• Whenever you ship something. A new page, a price change, a campaign. Look a week later, at that thing specifically, and be honest about whether the sample is big enough to judge.
• When something is alerted to you. Which is the case for getting a tool that tells you rather than waits for you.

The last one matters more than it sounds. A tool you have to remember to visit is competing with everything else in your week, and it will lose. Anything that surfaces the interesting thing without being asked — an email, a summary, an alert when a form stops working — is worth more than a dashboard with twice the features.

Consent banners exist largely because analytics tools set cookies and collect personal data. If your analytics does neither, the usual reason for the banner goes away. Plenty of small businesses run a banner out of caution while collecting data they never look at — paying an annoyance tax on every single visitor for nothing.

We are not your lawyers and this is not legal advice; your obligations depend on what else your site does, and other tools you embed may have their own requirements. But it is worth actually asking the question, because “everyone has one” is not a reason and the banner is not free — it costs you a fraction of every visitor before they have read a word.

1. Can you set up a goal without a developer? This is the single biggest predictor of whether you will get value. If tracking “someone clicked Book Now” requires code, it will not happen.
2. Does it work without a cookie banner? See above.
3. Does it tell you when it does not know? A tool that prints a confident number for every cell is not more capable than one that sometimes shows a dash — it is less honest.
4. Can you get your data out? CSV at minimum. It is your data, and export is the cheapest insurance against being stuck.
5. Where is the data stored, and who owns the company storing it? If this matters to your customers, get a specific answer rather than a badge.
6. Will you actually open it? The best analytics tool is the one that tells you the interesting thing without being asked. A tool you have to remember to visit will lose to the rest of your week, every week.

Coruve is built for exactly the reader of this post: a small business without a data team. It sets no cookies and needs no consent banner. It catalogues the buttons, links and forms your visitors actually use and names them in plain English, so setting up a goal is a few clicks rather than a developer ticket. It writes you a daily brief instead of waiting for you to remember it exists. And it refuses to divide numbers that are too small to divide, which is the behaviour described further up this page.

It also separates visitors sent by AI assistants from the crawlers those companies run — which very few tools do, and which is becoming a real question rather than a novelty. That is the subject of our complete guide to AI referral traffic if it is relevant to you.

Things it does not do, so you do not have to find out later:

• No session recordings. There are click and scroll maps, but nothing plays back an individual visitor’s session. If watching recordings is what you want, a different tool is the right answer.
• It is not open source, and it is not self-hosted by you.
• It will not rescue a site with almost no traffic. No analytics tool will. If you get a handful of visitors a week, the honest advice is to spend the month on getting visitors rather than on measuring them.

The pricing page has the current tiers, including a free one, and the free AI traffic checker will show you your assistant traffic without an account at all.