8 Best Free Website Analytics Tools for Small Websites

Umami's free hosted plan has a catch that deserves attention early: Hobby keeps six months of history. A dashboard can be easy to read and still be the wrong place to keep years of business records. In an October 2026 discussion about analytics retention, a website owner described discovering that older hosted history had disappeared. I'd choose self-hosting or a longer-retention plan before making year-over-year reporting part of the job.

For day-to-day traffic, the compact presentation is the reason Umami ranks first. You can look at pages, referrers, countries, devices, and named events without building a marketing workspace first. A local service business can start with its home, services, and contact pages, then add a contact-link event. That keeps the first report connected to an actual question: which pages lead someone toward getting in touch?

Umami also has reports for goals, funnels, journeys, retention, and other deeper analysis. It's grown beyond a page counter. I'd still choose it for a small site's traffic view before choosing it for a complicated multi-team measurement program.

The hosted Pro plan costs $20/month, including one million events, up to 20 websites, ten team members, and two years of retention. Extra events are billed separately. The open-source self-hosted software is free, while the server and upkeep are your responsibility.

Pick it if you want a clear traffic view and can choose a retention arrangement that matches your records. Skip it if you need advertising attribution and account integrations centered on Google products.

Google Analytics makes more sense when a website already has an advertising and measurement plan. Its event-based model connects visits with actions, and its Google product connections can matter more than having a tidy page-count screen. The official setup guide covers the account, property, data stream, and tag needed to start collecting.

I'd choose it for a store or lead-generation site whose owner expects to compare campaigns and conversion events. For example, a contact-page visit and a submitted inquiry answer different questions. Tracking only the page can make a campaign look successful while telling you little about whether the form worked.

The weakness is setup and reporting complexity. You need consistent event names, a clear conversion definition, and somebody willing to understand the reports. Adding the tag doesn't decide which business actions deserve measurement. A casual personal site can end up with much more reporting machinery than its owner will check.

The standard service is free; the enterprise Analytics 360 product is paid. Google's documentation is the place to follow current collection steps. I'd avoid copying an old Universal Analytics tutorial, because the current product uses a different model.

Choose this when the marketing connections justify the extra work. If the weekly question is simply which article brought visitors, Umami or Cloudflare can get you to that answer with less setup. Keep one measurement plan across the site so renamed buttons don't silently split the same action into several events.

Pick it if Google advertising and conversion reporting are part of the business. Skip it if you only need a short weekly traffic check and have no appetite for event configuration.

Matomo's On-Premise core is free, but free software isn't the same as a free analytics service. You'll supply hosting, database storage, updates, and report processing. Premium plugins can also add charges. I'd put those jobs into the decision before comparing it with a hosted plan.

The appeal is breadth. Matomo gives a website owner a fuller analytics system with traffic, acquisition, behavior, and goal reporting. It's a better candidate when you need a reporting process that grows beyond a few dashboard counters, particularly if the organization already operates PHP and a database.

The self-hosting documentation covers installation and maintenance. For a busy site, somebody needs to watch the database and the report-archiving work. If your website maintenance already goes unfinished, adding another application can make an inexpensive plan surprisingly costly in staff time.

Cloud hosting is paid, and some features in the Matomo ecosystem have separate plugin pricing. I wouldn't assume that every screenshot in a feature tour belongs to the free core. Start with the exact reports you need, then check the applicable edition.

Matomo is my choice for an owner who has both a reason to control the analytics installation and the capacity to look after it. Our Matomo vs Umami comparison explains the practical differences in reporting and maintenance.

Pick it if you need a broad reporting system and already have someone responsible for the server. Skip it if you want an account you can create and leave without database maintenance.

Cloudflare Web Analytics is a strong choice when you need basic website statistics and don't want to operate a reporting server. Its free service uses a browser beacon; that collection method matters because it measures browser activity, not every request arriving at your web server.

I'd choose it for a brochure site, a small publication, or a project whose owner checks traffic occasionally. The reports can answer where visitors arrived from and which pages attracted attention. That is enough to spot a popular tutorial or a neglected service page without adding a product-analytics program.

The weakness is the depth of the job it covers. A simple traffic dashboard won't replace a carefully designed funnel across account creation, activation, and subscription events. Someone who wants detailed cohorts or an experiment program should look at PostHog or Mixpanel.

The official overview explains the product and its collection approach. Other Cloudflare services have their own plans; choosing this free dashboard doesn't mean every performance or security feature from Cloudflare is included.

I'd also keep the numbers in context. Browser measurement and server logs count different activity, so disagreement doesn't automatically mean either system is broken. Use each report for the question it can answer. If a visitor's browser blocks the beacon, that visit may not appear in browser-based reporting.

Pick it if basic traffic reports are enough and hosted operation is a priority. Skip it if you need a detailed custom-event workflow or a product team's behavioral reports.

Microsoft Clarity combines heatmaps and session recordings in a free product that advertises no traffic limit. Its value is different from a page-count dashboard: it can help an owner examine how people interact with a page.

I'd add Clarity when the question is why people abandon a form, miss a button, or struggle on a phone. A traffic chart can show that a contact page is popular. A behavior view can point toward the part of the page that deserves closer inspection. That makes Clarity a complement to simpler analytics, rather than a direct replacement for every marketing report.

The weakness is that recordings need judgment. A pointer moving around a page doesn't explain the visitor's intention, and a heatmap doesn't prove that a layout change will improve results. I'd use it to choose a concrete design question, then check the page itself before making changes.

Because this collects richer interaction data, review the product's settings and your site's disclosure and consent arrangements before adding it. Don't put sensitive information into analytics labels. A form carrying private messages deserves more care than a public article page.

Clarity is free, with its product details on the official site. I'd keep the initial review narrow: one form, one navigation problem, or one important mobile page. A pile of recordings without a question becomes another queue nobody finishes.

Pick it if you need to investigate page behavior. Skip it if your only question is how many visits the website received last month.

PostHog's pricing gives product analytics a free allowance of one million events per month. Session replay and its other products have separate allowances. That's attractive for a small application, but you need to understand which products you're enabling before treating the whole account as cost-free.

I'd choose PostHog when a website behaves like a product: people sign up, use features, return, and sometimes pay. A sensible event plan might follow account creation, a first saved project, and a return visit. Those events can tell a team more than ranking the busiest URLs.

Its surrounding tools can also reduce the need to connect separate services for related product work. The trade-off is that the account has more scope than a small business traffic report. Teams need naming rules and a clear reason for collecting each event, otherwise reports become difficult to compare as the interface changes.

Paid usage depends on the product and volume. I'd check the pricing page with the actual features selected, then use billing controls available in the account. A free analytics allowance says nothing about an unrelated replay allowance or another enabled service.

PostHog earns a place for applications whose owners will act on feature behavior. For a static portfolio, I wouldn't take on that extra planning just to count visits to three pages.

Pick it if you want to measure product adoption and have a team that can maintain event names. Skip it if a small traffic dashboard already answers your questions.

Mixpanel's Free plan includes up to one million events per month, unlimited seats, and 10,000 session replays per month. Its free plan also limits saved reports per seat. I'd check that last detail before moving an entire team's reporting routine into the account.

Mixpanel centers the analysis on events. A product team can compare the steps between signup and a first completed task, or examine return activity among groups of users. This is a better starting point for a funnel question than asking a page counter to reconstruct the whole journey.

The strength is the ability to explore actions with a shared event model. The weakness is the work needed to create that model and keep it accurate. If one developer sends project_saved and another sends saved_project, the dashboard can split a single business action into two unrelated records.

Growth pricing is usage-based, with the current terms on the pricing page. I'd start with a few important events and the reports the team will actually use. Sending every possible interaction can increase volume while making the original question harder to find.

For an ordinary content site, Mixpanel is more work than I'd choose. For a product with a clear activation path, it's a serious candidate alongside PostHog. The choice then depends on how much surrounding tooling the team wants in the same account.

Pick it if your main questions concern funnels, retention, and product actions. Skip it if you need a report that a nontechnical site owner can check without an event plan.

GoatCounter's terms allow reasonable free public usage, including personal sites and small-to-medium businesses. The older claim that its free hosted service is only for noncommercial websites is too narrow. Very large traffic volumes fall outside that free arrangement.

I'd choose it for a project that wants a small set of traffic answers and little temptation to build an elaborate tracking program. Its official site describes hosted and self-hosted options. You can decide whether someone else should operate the service or whether you'd prefer to keep that work on your own infrastructure.

Its smaller scope is also the weakness. If you need advertising integrations, session recordings, or a product team's funnel workspace, another entry here has a closer match. Don't choose a minimal tool and then spend weeks trying to make it behave like a larger reporting suite.

The free hosted service has fair-use terms, while self-hosting carries operating costs. I'd read those terms before putting a high-traffic publication on it. A small personal blog and a site sending millions of pageviews a day create very different workloads for a hosted service.

Pick it if modest public traffic reporting is the whole job. Skip it if you expect high-volume product analysis or a broad marketing integration setup.