Web analytics is the practice of measuring how people use a website — the job Google Analytics, Matomo and SlimStat all do. Each of them collects a record of every visit, stores it, and shows it back as a report: where visitors came from, which pages they opened, and what they did before leaving.
The field divides in two, and most of the confusion around it comes from mixing the halves.
| Dimension | On-site analytics | Off-site analytics |
|---|---|---|
| What it measures | Visits, pages, clicks and goal completions on your site | Rankings, mentions, shares and competitor visibility |
| Where data comes from | A tracker or server log you control | Search engines, social platforms, third-party estimates |
| Who holds the data | You, if the tool writes to your own database | The external platform, always |
| Typical question | Why do visitors leave the checkout page? | Who outranks us for this keyword? |
| Example tools | SlimStat, Matomo, Google Analytics | Google Search Console, Ahrefs, Semrush |
Everything below is on-site analytics, because that is the half you install, own and can change.
What actually gets recorded
An on-site analytics tool keeps a log of small facts about each visit:
- how someone found you — a search, a link, a newsletter, or no referrer at all
- which pages they opened, and in what order
- how long each page held them
- which page they left from
- whether they did the thing you wanted: signed up, bought, downloaded
None of that is interesting on its own. It becomes useful the moment you compare it: this week against last, one landing page against another, mobile against desktop.
How a visit becomes a number
Every analytics tool, however it is packaged, does three jobs. Break any one of them and the reports go quiet.
Collect. A small script in the page notes that a visit happened. Some tools can do the same work server-side, with no JavaScript in the browser at all — which is how a self-hosted tool keeps counting when an ad blocker strips the script.
Store. The record is written somewhere durable, so it survives the visitor closing the tab. Hosted services send it to the vendor’s servers; a self-hosted WordPress plugin such as SlimStat writes it straight into a table in your own site database.
Show. The stored rows get grouped into counts, charts and comparisons a person can read.
Follow one pageview through all three:
- A visitor opens
/pricing/at 3:14 PM from Berlin. - The tracker notes the URL, the time, the referring page, and a rough location derived from the IP address.
- The browser sends that note in the background, after the page has rendered, so the visitor waits for nothing.
- Storage accepts the note as one new row.
- The dashboard shows it — first as one live hit, later as part of “top pages this week”.
The whole round trip is invisible to the person browsing. What changes is what you can see afterwards.
Where the practice came from
In the late 1990s, site owners read raw server log files: dense, technical, and usually parsed by hand. Google Analytics arrived in 2005 and made JavaScript tracking with a hosted dashboard the default. The recent shift is back the other way, toward tools that keep the data on the owner’s own server — driven by GDPR and by how much of the web now blocks third-party scripts.
The metrics worth watching
Six numbers cover most small sites. Start with one or two, get comfortable, then add more.
| Metric | What it tells you | What to watch out for |
|---|---|---|
| Pageviews and sessions | Raw traffic volume | Bots inflate both unless filtered |
| Unique visitors | How many different people came | Every tool counts these differently |
| Traffic sources | Who sent them | ”Direct” is often an untracked referral |
| Bounce rate | Share who left after one page | High on a checkout page means trouble |
| Session duration | How long people stayed | The final page of a visit is unmeasurable |
| Conversion rate | Share who completed a goal | Only as honest as the goal you defined |
Beyond those sit behavior flow, exit rate, on-site search terms and custom events — worth adding once the basics are answering questions rather than raising them.
What it looks like in practice
Online stores read checkout behavior to find where carts are abandoned. Writers watch time-on-page to see which articles hold attention. Nonprofits follow donation flows. Marketers check which campaigns produce sign-ups rather than clicks.
A worked example. Suppose a recipe blog has steady traffic but almost no newsletter sign-ups. The reports show most visits arriving on mobile from Pinterest, an average visit under 30 seconds, and few readers scrolling past the ingredients list. Those three facts point at one change rather than five: the sign-up form sits below the fold on the device nearly everyone is using. Move it, then compare next month against this one — possible only because you were already recording.
Common mistakes
- Tracking everything “just in case”, then reading none of it
- Ignoring how differently mobile visitors behave
- Collecting insights and never acting on one
- Overcomplicating the setup before the first question is answered
- Forgetting the legal side — GDPR, CCPA and consent for anything identifying
Choosing a tool
- Solo blogger — something lightweight you will actually open, such as Plausible
- WordPress site or store — a plugin that reports inside wp-admin, such as SlimStat Analytics, which has 70,000+ active installs and a 4.8 rating from 817 reviews
- Marketing team with existing reporting — Google Analytics or Matomo, for the integrations
For a WordPress site, self-hosting instead of Google Analytics has three practical advantages: the data never leaves your server, which simplifies GDPR and CCPA questions; there is no second dashboard to learn; and you see every hit in real time rather than a sampled estimate.
SlimStat’s free plugin covers the real-time view, server-side tracking mode, geolocation, exclusion filters and the raw access log. Heatmaps, scheduled email reports, network-wide stats for multisite, funnels and exports are Pro features.
Getting started in an afternoon
- Pick one goal — more newsletter sign-ups, fewer abandoned carts, whatever pays.
- Install a tool that matches your comfort level. On WordPress: Plugins → Add New → search for it → Install → Activate.
- Confirm it is collecting. Open your own site in a private tab and look for yourself in the real-time view.
- Exclude yourself, by IP address or user role, before your own visits distort the numbers.
- Watch three to five metrics that relate to your goal, once a week.
- Make one change based on what you saw, then check whether it worked.
Then repeat. That loop is the whole discipline.
Frequently asked questions
Do I need coding skills to use web analytics?
No. Most tools are point-and-click, and a WordPress plugin needs no code at all — you install it the same way you install anything else.
How often should I look?
Weekly is a sensible rhythm. Daily only while a campaign is live, because day-to-day movement on a small site is mostly noise.
Does analytics slow a site down?
Not measurably, with one well-built tracker. Running four overlapping tools does slow a site down — a setup problem, not an analytics problem.
Can I track visitors without cookies?
Yes. Cookie-less counting is common now, and it is what makes consent-free measurement possible in much of Europe. It trades accuracy about returning visitors for a far simpler compliance story.
Is web analytics legal under GDPR?
It depends on the tool and the configuration, not on analytics as a category. Storing data yourself, anonymizing IP addresses, and not sharing with third parties are what make the answer easy.
What is the difference between self-hosted and cloud analytics?
Self-hosted keeps the records in your own database, on your own hosting. Cloud analytics stores them on a vendor’s infrastructure and shows them back to you. The trade is control and privacy against convenience and scale.