Which open-source web analytics tool fits your team
If your team needs marketing reports hosted on its own infrastructure, the choice comes down to three self-hosted options: Matomo for broad marketing reporting and consent-sensitive features, Plausible Community Edition for lightweight, cookieless reporting with a narrower feature set, and Umami for a minimal stack under the MIT license. Each one is a fair choice for a different kind of team, and none of them removes operating work from your side. Whichever you pick, your staff or your implementation partner will own installation, upgrades, capacity planning, backups and security patching, because all three are self-hosted and the vendors do not run the instance for you.
PostHog belongs in a different conversation. It is a product analytics suite with session replay, feature flags and surveys, and it is run mainly as PostHog Cloud, so teams that want to answer product questions such as feature adoption or funnel behavior inside an application should evaluate it alongside these tools rather than instead of them. If your core question is what visitors did on your marketing site and which campaigns drove them there, the three tools above are the closer fit. Before you shortlist, confirm which questions your marketing, legal and product owners need answered, because that answer decides more than any feature table.
Start by writing down three requirements: the monthly volume you expect, whether your regulators or your legal team expect a consent banner, and which reporting features your marketing team cannot work without, such as campaign attribution, ecommerce revenue goals or Looker Studio reporting. Then ask each vendor or your implementer to confirm current release support and the hosting footprint at your expected volume. Ask about the cost of operating each option as well as the licence fee, since Community Editions are free to download but still need staff time, infrastructure and a support path. Your implementation partner can turn those answers into a written scope and a quote.
Which analytics option suits which scenario
This table matches common analytics scenarios to the option that fits best, with the condition that makes each recommendation hold. Read each row as a starting point, then check the named condition against your own traffic, team and consent position before choosing.
| Scenario | Best fit | Reason and condition |
|---|---|---|
| Migrating from Google Analytics | Matomo On-Premise | The free Google Analytics Importer brings in aggregated reports only, so Visitor Log, Custom Reports and Segmentation will not work on imported data and some reports such as Funnels are not imported. Plausible can import GA4 daily aggregates, but it cannot import ecommerce revenue. Choose Matomo if you need the widest report set on your own servers. |
| Consent-sensitive EU site | Plausible Community Edition or Umami | Both are cookieless. Plausible states its service uses no cookies, so no banner is required, while Umami states it works without cookies or automatic collection of personal data. Matomo can also fit consent-exempt setups in some EU countries, but only if you keep to CNIL-style configuration and avoid User ID, heatmaps or session recordings. Confirm the exemption position with your data protection adviser. |
| Small team wanting the lightest stack | Umami | Umami runs on Node.js with PostgreSQL 12.14 or newer, and publishes a Docker Compose file with a bundled PostgreSQL database. Plausible Community Edition needs Docker, Docker Compose, ClickHouse support on the CPU and at least 2 GB of RAM. Matomo needs PHP and MySQL or MariaDB behind a web server, and it is sized from 4 CPUs and 8 GB of RAM at the low end. |
| Team needing heatmaps or session recordings | Matomo with the paid Heatmap and Session Recording plugin | Matomo's Heatmap and Session Recording plugin is a paid premium plugin sold per user under the InnoCraft EULA, with a 30-day free trial. Neither Plausible Community Edition nor Umami is positioned for this need in the supplied facts, so confirm current feature coverage with each vendor. Adding heatmaps or session recordings to Matomo also removes the CNIL consent exemption, so budget for consent handling. |
| Team needing product analytics | PostHog | PostHog is built for product analytics, session replay, feature flags and surveys, which go beyond the marketing reports these three tools focus on. Choose it when product funnels and feature usage drive decisions. PostHog is mainly run as PostHog Cloud, so confirm its self-hosting terms before planning an on-premise install. |
| Operator wants full control of upgrades and backups | Matomo On-Premise or Umami self-hosted | Both are installed and run in your own infrastructure. Plausible Community Edition security fixes arrive only when you apply each twice-yearly release, and the operator handles upgrades, capacity, uptime and backups. Confirm who will apply patches and test restores before choosing. |
| Need for cloud-only features | Plausible Cloud or Umami Cloud | Plausible Community Edition lacks marketing funnels, user journeys, ecommerce revenue goals, single sign-on, the Sites API, the Looker Studio connector and advanced bot filtering. Umami Cloud adds email reports and a streaming API. If these matter, a managed plan may fit better than self-hosting, and you can compare plans on the vendor pricing pages. |
What the license and edition terms mean for your deployment
Matomo's repository is licensed under GPLv3, and the On-Premise Community edition is free forever with unlimited users and hits, supported through forums. The license covers the core code, but several features you may expect are sold separately, so your total licensing picture depends on which plugins and hosting model you pick.
- Community edition: free forever with unlimited users and hits, with forum support. Any paid support contract for your own install is a separate purchase that you need to scope in writing.
- Heatmap and Session Recording: a premium plugin sold per number of users under the InnoCraft EULA, which is a proprietary license and not OSI-approved. It includes a 30-day free trial, so test it before you buy seats.
- Google Analytics Importer: free for On-Premise under GPL v3 or later. It imports aggregated reports only, so the Visitor Log, Segmentation and Funnels do not carry over.
- Consent position: Matomo's ePrivacy FAQ says that adding User ID, heatmaps or session recordings removes the CNIL exemption for certain configurations. Ask counsel to review this before you enable those features in an EU deployment.
- Matomo Cloud: paid tiers start at 29 per month excluding tax, with data stored in Frankfurt, Germany. Cloud is a hosted service, so the GPL terms describe the code, not the service you receive. Matomo pricing lists the current tiers.
Before procurement, list every plugin you need, mark each one as free, per-user or hosted, and have the EULA for any paid plugin reviewed by your legal team.
Plausible Community Edition is licensed under AGPLv3 or any later version. The AGPL extends source-sharing duties to software that users interact with over a network, so the license matters for a modified analytics service that your staff or visitors reach through a browser, not only for copies you distribute. Plausible's self-hosted page describes Community Edition as a long-term release published twice per year, supported by the community only, with security fixes arriving only when you apply them.
- Modified service obligation: if you change the code and users interact with it over a network, the AGPL generally requires offering those users the corresponding source of your modified version. Ask counsel how this applies to internal dashboards and to private patches.
- Release cadence: two long-term releases a year, with security fixes only after you apply them. Security release v3.2.1 was published on May 15, 2026, after v3.2.0 on January 26, 2026, so someone must track and install these.
- Features missing versus Cloud: marketing funnels, user journeys, ecommerce revenue goals, single sign-on, the Sites API, the Looker Studio connector and advanced bot filtering.
- Cloud pricing reference: Plausible Cloud's Starter plan lists $9 per month for 10k monthly pageviews, with custom Enterprise pricing for larger needs.
- Data portability: Community Edition users can export and import data to and from Plausible Cloud, which keeps a later move between models possible.
Confirm with Plausible in writing which support terms apply to your install, and check which Cloud-only features your reporting depends on before you choose the self-hosted path.
Umami is licensed under the MIT license, which is OSI-approved and permissive. Umami's pricing page states that the software is open source and can be self-hosted for free. The licensing question is simpler than for the AGPL, but the feature question is not: Umami Cloud adds email reports and a streaming API that the self-hosted version lacks, so reporting and integration needs decide whether the paid tier is worth it.
- MIT terms: you may modify, redistribute and run the code, provided you keep the copyright and license notice. No duty to publish changes applies to a self-hosted install, which is the main contrast with the AGPL option in this comparison.
- Cloud tiers: Hobby is $0 per month for up to 100K events with 6 months of retention. Pro is $20 per month for 1 million events with 2 years of retention. Business is $200 per month for 10 million events with 5 years of retention. Umami pricing has the current terms.
- Overage: $0.00003 per extra event on Pro and $0.00002 per extra event on Business. Enterprise pricing and retention are custom.
- Missing self-hosted features: email reports and the streaming API. If you need scheduled reports, decide whether to build them outside Umami or buy the Cloud tier.
- Release tracking: v3.4.0 was published on September 17, 2026, after v3.3.1 on August 20 and v3.3.0 on August 12. The sources supplied do not state a support window, so ask Umami for it in writing.
Before you choose between self-hosting and Cloud, compare your expected event volume and retention requirement against the tiers and confirm how long your audit or reporting needs data to stay available.
Matomo, Plausible Community Edition, Umami and PostHog in brief
Matomo, Plausible Community Edition and Umami all cover marketing-style web analytics that you can run yourself, while PostHog serves product teams that need behavior analysis. The cards below set out each license, current release, stack and the team each one suits, so you can narrow the shortlist before comparing running costs.
Matomo
Matomo is GPLv3 licensed, and its On-Premise Community edition is free with unlimited users and hits, with forum support. Matomo 5.14.1 is the latest stable release, and 6.0.0 is in beta. Version 6.0.0 needs PHP 8.1 or newer and MySQL 8.0 or MariaDB 10.6 or newer. Teams that need a full feature set should check which heatmap and session recording features are sold as paid plugins under a separate EULA, because those features remove the consent exemption in some EU markets.
Plausible Community Edition
Plausible Community Edition is AGPLv3 licensed and runs on Docker and Docker Compose with ClickHouse, needing at least 2 GB of RAM and a CPU with SSE 4.2 or NEON. It is a twice-yearly long-term release with community support only, so security fixes arrive only when your team applies them. The v3.2.1 security release was published on May 15, 2026. Choose it if your team can run Docker and ClickHouse and accepts that features such as SSO, the Sites API and ecommerce revenue goals sit only in Cloud.
Umami
Umami is MIT licensed and is a privacy-first platform that runs without cookies. It runs on Node.js with PostgreSQL, which the installation docs support from version 12.14. Self-hosting is free, but Umami Cloud adds features such as email reports and a streaming API. Umami v3.4.0 was published on September 17, 2026. It suits small teams that want a light stack with a bundled Docker Compose setup, provided they can handle their own upgrades and backups.
PostHog for product analytics
PostHog is MIT licensed outside its ee directory and is run mainly as PostHog Cloud, so self-hosting is the exception rather than the default path. It covers product analytics with session replay, feature flags and surveys. Choose it when your question is about user behavior inside a product, not campaign performance or marketing reports. Confirm the licensing of any ee features you plan to use before you deploy.
How to choose and pilot an analytics platform
This sequence moves a team from a list of reporting needs to a tested, staffed analytics install before go-live. Each step produces a written decision or a check that the platform works in your environment.
List the reports marketing and product teams need
Marketing and product leads write down the reports they use every week, such as traffic sources, campaign attribution, funnels, goal conversions, ecommerce revenue and session replays. Mark which items are marketing reports and which are product analytics questions, because a page-level tool and a product analytics suite answer different questions. The step is done when each report has an owner and a named decision it supports. Where a team needs funnels or session recordings, note that these may sit in paid add-ons or in a separate product analytics tool such as PostHog, which is covered elsewhere on this page.
Decide whether Google Analytics history must come across
Ask the analytics owner whether year-over-year comparisons or long trend lines depend on existing Google Analytics data. Matomo's Google Analytics Importer brings in aggregated report data only, so raw visit logs, Visitor Log, Custom Reports and Segmentation do not work for imported data, and some reports such as Funnels are not imported. Plausible imports historical GA4 stats as daily aggregates up to the first native Plausible visitor, supports up to 5 GA properties per site, and cannot import ecommerce revenue. The step is done when the team has a written decision to import, to start fresh, or to keep Google Analytics running in parallel for a defined period.
Confirm consent requirements with counsel
Ask your privacy counsel or data protection officer which markets you serve and which consent rules apply to each site. Matomo's ePrivacy FAQ says France's CNIL treats some analytics as exempt from consent only when it is limited to performance measurement, collects anonymous statistics and excludes marketing measurement, and that features such as User ID, heatmaps or session recordings remove that exemption. Confirm with counsel whether the same configuration applies in each target country, because the FAQ names France, Italy, the Netherlands and Spain and points to the CNIL configuration for the others. Umami and Plausible describe their tools as cookieless, which counsel should still verify against your own tag setup and consent banner. The step is done when counsel signs off on a written configuration per market.
Size the server against published guidance
Estimate monthly pageviews across all sites in scope, then compare against the published sizing. Matomo's requirements FAQ sizes a single server at 4 CPUs, 8 GB of RAM and 250 GB of SSD for up to 1 million pageviews a month, and recommends separate app and database servers with 8 CPUs and 16 GB of RAM each for up to 10 million pageviews a month. Plausible Community Edition requires Docker, Docker Compose, a CPU supporting SSE 4.2 or NEON and at least 2 GB of RAM. Umami requires Node.js 18.18 or newer for a source install and supports PostgreSQL from version 12.14. The step is done when the infrastructure team has a sized design and a list of the versions it will run.
Run a pilot on a staging domain
Deploy the shortlisted platform on a staging domain and send it test traffic that mirrors production, including campaign links, form submissions and any custom events. Compare the pilot's counts against the existing tool for the same period and document where they differ, since bot filtering, script blocking and sampling rules vary by product. Check that the reports on your list from step one can be built without workarounds, and that exports and imports work if you plan to move data later. The step is done when the team has a comparison table of reports, a list of gaps, and a go or no-go recommendation for each platform on the shortlist.
Confirm who owns upgrades and backups
Name the person or provider responsible for installation, upgrades, capacity, uptime, backups and security patches. Plausible's self-hosted page states that Community Edition operators handle these tasks themselves, and that the Community Edition receives security fixes only when the operator applies a release. Confirm which release line you will track and whether you will rely on community support or a paid support contract. Ask any implementation partner for a written scope covering these responsibilities. The step is done when the owner, the upgrade cadence and the backup test schedule are written down and accepted.
Check upgrade and backup readiness before go-live
Before switching production traffic, restore a backup into a separate environment and confirm the analytics data and configuration come back intact. Run one upgrade on the staging copy using the vendor's documented procedure and record any manual steps. Confirm that the team can read the vendor's release notes and security advisories and knows who acts on them within the internal change process. The step is done when a restore and an upgrade have both been tested on staging and the runbook names an owner for each task.
What running a self-hosted analytics install takes
Before you pick a self-hosted analytics platform, decide who will run it day to day, because the three options place very different work on your team. Plausible Community Edition operators manage installation, upgrades, capacity, uptime, backups and security themselves. Plausible's self-hosted page says the community is the only support channel, and it receives security fixes only when the operator applies a release. Plausible publishes Community Edition as a long-term release twice per year, so your team needs a standing task to track and apply each one. Confirm with Plausible whether any paid support exists for your install before you assume it does.
Matomo On-Premise sizing sets the hardware baseline. Matomo's requirements FAQ sizes a single server with at least 4 CPUs, 8 GB of RAM and 250 GB of SSD for up to 1 million pageviews a month. For up to 10 million pageviews a month, it recommends separate application and database servers, each with 8 CPUs and 16 GB of RAM. Matomo 6.0.0 needs PHP 8.1 or newer and MySQL 8.0 or newer or MariaDB 10.6 or newer, so check your current PHP and database versions before planning an upgrade. Umami is lighter on paper: a source install needs Node.js 18.18 or newer and PostgreSQL 12.14 or newer, and its Docker Compose file starts a bundled PostgreSQL database. Your real sizing depends on your traffic and retention settings, so load test with your own event volumes.
Recurring costs for all three are mostly staff time, hosting, backups and upgrades, and only some of them show up on an invoice. Staff time covers patching, monitoring, restore testing and handling reports of broken tracking. Hosting scales with traffic and retention, and backups need their own storage and tested restore procedures. Ask any implementation partner for a quote that lists each component separately, including who performs upgrades, how often backups are restored in testing, and what happens during a failed upgrade. Ask the vendor or community which upgrade path they support from your current version so your costs stay predictable over time.
Questions buyers ask before hiring help with web analytics
What drives the cost and effort of self-hosting Matomo, Plausible or Umami?
Infrastructure and people drive most of the cost, and both recur after go-live. Matomo's requirements FAQ sizes one server at at least 4 CPUs, 8 GB of RAM and 250 GB of SSD for up to 1 million pageviews a month, and recommends separate app and database servers with 8 CPUs and 16 GB of RAM each for up to 10 million. Plausible Community Edition needs Docker, Docker Compose and at least 2 GB of RAM, and Umami needs Node.js 18.18 or newer for a source install plus PostgreSQL 12.14 or newer. Labor covers installation, upgrades, backups, uptime and security patching, which the operator owns on all three. Ask any provider for a written scope that lists each of these tasks, the hosting assumptions and who is paid for each recurring item.
Can an in-house team run open-source web analytics without outside help?
Yes, if the team already runs web applications, containers and databases and can own patching. Umami and Plausible ship Docker-based paths, and Matomo runs on a standard PHP and MySQL or MariaDB stack behind a web server. The harder parts are capacity planning as traffic grows, keeping the Plausible Community Edition on a current release (its security fixes arrive only when the operator applies the release), and verifying consent treatment for each jurisdiction. A team without that background should either train on these tasks or buy a fixed-scope implementation and a retainer for upgrades.
How long do a pilot and a migration to self-hosted analytics take?
No supplied fact gives a reliable duration for a pilot or migration, so confirm the timeline in writing before work starts. The variables that set it are the number of sites, the volume of historical data you want to keep, and whether tags and consent banners change on every site. Plausible's Google Analytics import brings in GA4 history as daily aggregates up to the first native Plausible visitor, supports up to 5 GA properties per site, and cannot import ecommerce revenue. Matomo's Google Analytics Importer imports aggregated reports only, so Funnels and the Visitor Log do not carry over. Ask the provider to map each report you need against these limits before agreeing a schedule.
What does the work need from our side: hosting, a database owner and an upgrade contact?
Three client-side roles are needed before the first deployment. Hosting means a cloud account, data center or private network where the platform runs under your control, and DeployOpen works inside that infrastructure. A database owner must manage PostgreSQL for Umami, ClickHouse with Docker for Plausible, or MySQL or MariaDB for Matomo, including backups and restores. A named upgrade contact approves release windows, since Matomo 6.0.0 is in beta with new PHP and database minimums and Plausible Community Edition ships twice a year. Name these people in the statement of work so that the engagement does not stall on access requests.
Does Matomo's heatmap plugin affect cookie consent?
Yes, adding it can change your consent position. Matomo's ePrivacy FAQ says the CNIL recognizes some audience measurement trackers as exempt from consent only when they collect anonymous statistics, stay limited to purposes such as performance measurement, and exclude marketing. The same FAQ states that features such as User ID, heatmaps or session recordings remove that exemption. Matomo's Heatmap and Session Recording plugin is a paid premium plugin licensed under the InnoCraft EULA rather than the GPL, with a 30-day free trial. Have counsel review the exemption criteria for each country where you collect data before enabling the plugin, and confirm the tracking configuration with your data protection officer.
Can Plausible Community Edition export data to Plausible Cloud?
Yes, Plausible's self-hosted page says Community Edition users can export and import data to and from Plausible Cloud. That makes a move in either direction practical, but it does not carry every feature. The Community Edition page lists marketing funnels, user journeys, ecommerce revenue goals, single sign-on, the Sites API, the Looker Studio connector and advanced bot filtering as Cloud-only. Before you commit to self-hosting, list the reports and integrations your team uses today and check each against that list. Confirm the export format and any data gaps with Plausible directly, since the supplied material does not describe the export process in detail.
When is PostHog a better fit than Matomo, Plausible or Umami?
PostHog is the better fit when the team needs product analytics, such as funnels, user journeys, feature flags or session replay inside the product, rather than marketing and traffic reports. Matomo, Plausible and Umami focus on page-level traffic, campaigns and conversions. PostHog is MIT licensed outside its ee directory and is run mainly as PostHog Cloud, so check which components you can self-host and what they need before comparing the effort to run it. If your main question is how visitors arrive and convert on a website, one of the three web analytics platforms usually fits better, and a separate review of PostHog's self-hosting requirements is worth doing first.
How DeployOpen works on this
DeployOpen plans, implements, customizes, integrates, hardens, backs up, upgrades, documents and supports self-hosted open-source platforms inside your own cloud accounts, data centers or air-gapped networks. For web analytics, that means starting from your consent position, traffic profile and reporting needs, then building the chosen tool to match them.
How we scope
- Engagements are fixed-scope, with deliverables and timelines agreed in writing before work starts.
- Scope starts with which sites, pageview volumes and reports the analytics tool must cover.
- We confirm the consent approach and any cookie or tracker exemption with your privacy team before configuration.
How we build
- We install the chosen platform in your infrastructure, sized against the vendor's published requirements for your traffic.
- We configure tracking scripts, sites, goals and any Google Analytics import, then test data accuracy against a known source.
- We harden the database, web tier, admin access and backups, and check that no data leaves your environment unexpectedly.
How we hand over
- Deliverables include installation records, configuration notes, sizing assumptions and a runbook for your operators.
- We document upgrade steps, backup and restore procedures and the known feature limits of the chosen edition.
- Handover includes a walkthrough with the team that will own dashboards, users and monthly reporting.
How we support
- Ongoing support, upgrades and continuous testing are available as retainers after handover.
- Retainers can cover version upgrades when a new release changes requirements, such as PHP or database versions.
- Ask us to confirm the support path in writing, including who applies security releases and within what window.
