Open-source alternatives to Google Analytics 4 in 2026

TL;DR

• Google Analytics 4 is free in money but expensive in sampling, consent-banner overhead, and lost data ownership. Standard reports are unsampled; Exploration reports start sampling above 10 million events per query.

• Three open-source replacements cover the full spectrum: Matomo (full feature parity, official GA4 importer), Umami (clean dashboards, lightweight), Ackee (absolute minimum, GraphQL-first).

• At production scale, the math favours managed open-source. Matomo Cloud starts at €22/month for 50,000 hits. Managed Matomo on DANIAN runs €9/month with no hit cap.

• All three tools default to a cookie-less tracking mode. When configured properly, that mode does not require a cookie banner in many jurisdictions — a configuration property, not a certification.

• Matomo's official GA4 Importer pulls historical data via the Google Analytics Data API. Migration does not start a new analytics record from zero.

Why people are leaving Google Analytics 4 in 2026

The honest summary is short. GA4 is free monetarily, and the cost has moved into ops time and data quality. Marketers stay because the bill is zero. They leave when an audit, a client question, or a 10-million-event sampling threshold forces them to count what zero actually buys, and to write it down.

The trigger is rarely the platform itself. The trigger is a privacy-policy review with a client, a question from an auditor, or the marketing freelancer doing diligence on a customer's site. The conversation goes: where does this visitor data go? Who sees it? Can we show that to a prospect? GA4 has answers, and they involve Google Cloud, US-based infrastructure, and Consent Mode v2. Those answers are workable; they are also work, and that work shows up on the freelance marketer's invoice.

Then there is the data itself. Standard GA4 reports are unsampled. The moment you build an Exploration report — the kind that answers a real question — you cross into sampled territory once a query touches more than 10 million events (Google Analytics support docs). The "(other)" row collapses any dimension above roughly 25,000 unique values, hiding the long tail. Demographic data is thresholded for privacy. None of this is hidden by Google; the icons are right there in the report header. It is hidden by being normal — most teams stop noticing the orange "data quality" indicator after a week.

Consent Mode adds another layer. Visitors who decline cookies are not measured; they are modelled. Industry estimates put the under-counting at 30–50% on sites that enforce consent strictly. The decision to base marketing budget on modelled traffic is a real one — defensible, sometimes correct, but not the same as observed traffic.

The escape ladder inside Google's ecosystem is real and expensive. Linking GA4 to BigQuery exports raw events and bypasses sampling, but it adds Google Cloud Platform spend and engineering work to maintain the schema. Server-side tagging via Google Tag Manager moves the tracking call off the browser. That improves data accuracy, at the cost of a Cloud Run instance and the time to operate it. Each step on the ladder costs more than the last and locks the team further into Google's stack. The buyers reading this post often realised the cumulative bill had quietly grown larger than a managed open-source instance would have cost in the first place.

For the buyers who tried to pay their way out of these constraints, GA 360 exists. Pricing is custom and reportedly starts around $50,000/year on annual contracts; large estates run upwards of $150,000/year. That number is the floor of "stop sampling me" if the rest of the stack stays inside Google's ecosystem.

So the search begins. "Open-source alternative to Google Analytics." "Self-hosted analytics." "Replace GA4 without breaking my dashboards." The pages this post covers are what the buyer finds.

What "alternative to GA4" actually means in 2026

Three paths exist, and naming them up front saves the reader a side quest. The cheaper-SaaS path moves the bill, not the model. The self-host path moves the data home, at a real cost in time. The managed-open-source path keeps the data ownership of self-hosting and sheds the operations cost. This post is about the third path.

Path A — A cheaper SaaS. Plausible Cloud, Fathom, Simple Analytics. These are managed services run by their respective companies, often on EU infrastructure, with simpler dashboards than GA4. They solve some of GA4's problems and do not solve the data-ownership question. They are a real option; they are not what this post covers.

Path B — Self-host the open-source tool. You install the software on your own server. You keep all the data. You patch and back up the system, and you absorb the operational time cost. Free monetarily. Real cost in hours and on-call risk.

Path C — Managed open-source. Someone else runs the open-source tool for you on infrastructure you can pick. You keep the data ownership of self-hosting and shed the patching, the backups, the upgrade cycle, and the 2 a.m. alerts. This is the path the rest of this post covers.

Three open-source tools fit Path C cleanly. Each represents a distinct philosophy. Pick by what you actually use, not by what GA4 happens to expose.

The shortlist — three open-source GA4 replacements

Matomo — full feature parity, including the GA4 importer

Matomo is the long-running heavyweight. It shipped originally as Piwik in 2007, rebranded to Matomo in 2018, and is maintained by InnoCraft Ltd. Roughly 1.5 million sites run it. The feature surface is the closest you can get to GA4 from outside Google.

License. GPL-3.0. GitHub stars. Around 21,500 on matomo-org/matomo as of May 2026.

What you get: real-time visits, sessions, goals, e-commerce reports, sales funnels, multi-channel attribution, and a tag manager. Heatmaps and session recording are paid plugins (€199/year and €149/year on the official marketplace). A/B testing is also a paid plugin. Everything that does not need ML-grade compute is in the box; the paid plugins exist because Matomo is a small company funded by them rather than by ads.

The piece that matters for migration is the GA4 Importer plugin. It is official, it is free, and it pulls historical data from the Google Analytics Data API. The importer covers up to 13 months of GA4 history — Google's API limit, not Matomo's — and brings over pageviews, sessions, events, conversions, and the standard set of dimensions. After the import, your Matomo dashboard shows trend data going backwards, not just forwards. Marketers who were ready to abandon GA4 because the audit was uncomfortable rarely want to lose the year of trend data they actually use; the importer is the bridge.

Matomo supports a cookie-less tracking mode out of the box. When configured properly — anonymised IPs, no cross-site identifiers, no consent-bound cookies — that mode does not require a cookie banner in many jurisdictions. This is a configuration property of the deployment, not a legal claim about your site. Verify it with someone who reads regulations for a living before relying on it.

The wider plugin marketplace is the second reason teams pick Matomo. Beyond the heatmaps and session-recording add-ons named above, there are paid plugins for media analytics, form analytics, custom reports, advanced segmentation, and roll-up reporting across multiple sites. Free community plugins cover GeoIP databases, custom alerts, multi-currency e-commerce, and a hundred smaller jobs. The point is not that you will use most of them. The point is durability. When a marketing team needs a specific GA4-shaped capability six months from now — say, a custom dashboard that combines two campaigns into a single funnel — there is usually already a plugin for it, or the documented extension API to build one.

Cost stack. Self-hosted on your own server: free, plus your time. Matomo Cloud (run by Matomo themselves): the Business plan starts at €22/month for 50,000 hits across up to 30 websites. Larger Cloud tiers scale up to €822/month for 5 million hits. 
Managed Matomo on DANIAN: €9/month with no hit cap, no per-website surcharge, and full plugin compatibility. → managed Matomo hosting

Best for: marketers who need the full GA-shaped feature surface and a real migration path off GA4.

Spin up managed Matomo on DANIAN — 7 days free, no card.

Umami — clean dashboards, lightweight footprint

Umami is the modern middle ground. Built by Umami Software, Inc. in San Francisco, it sits between Matomo's everything-included surface and Ackee's deliberate minimalism. The first hour with Umami feels noticeably lighter than the first hour with GA4 — that is the design intent.

License. MIT. GitHub stars. Roughly 25,000 on umami-software/umami as of May 2026, with active commits through April.

What you get: real-time view, pageviews, sessions, custom events, segments, goals, and a clean filter model where every filter shows up in the URL — share the URL with a colleague, they see the same view. Version 3 added three things worth naming: Boards for composing custom dashboards from charts and metric components; Session Replay built on rrweb with configurable masking levels for privacy; and Web Vitals tracking that captures LCP, INP, CLS, FCP, and TTFB from real user browsers. The tracking script weighs about 2 KB.

The Next.js + PostgreSQL (or MySQL) stack is the practical reason Umami feels fast on small servers. It is one container plus a database. There is no Java daemon, no separate ingestion service, no message queue. Operators who have run Matomo at scale notice the difference; teams without a dedicated operator notice it more.

Umami uses no cookies and stores no PII by default. Visitor identifiers are derived from a daily-rotating hash of IP, user-agent, and a server-side salt. The cookie-less default is what allows the same "no banner required in many jurisdictions, when configured properly" framing as Matomo.

A practical note for teams that have run Matomo before: Umami's resource footprint is genuinely smaller. A medium-traffic site (a few hundred thousand pageviews per month) typically runs comfortably on the base €9/month tier without resource pressure. The same site on Matomo benefits from one resource upgrade once historical data builds up. Neither tool will stretch a budget; the difference matters mostly for teams running analytics on the same instance as something else, or who care about the ingest-script size on their pages. The 2-KB Umami tracker is half the size of the Matomo equivalent.

Cost stack. Self-hosted: free. Umami Cloud (run by Umami): a free tier covers 10,000 events per month across three websites; paid plans start at $0/month for hobby tier and reach $20/month for the higher Pro tier, event-metered above the included caps (umami.is/pricing). 
Managed Umami on DANIAN: €9/month per instance. → lightweight Umami analytics

Best for: teams who want a clean, fast dashboard without GA4's complexity tax — and who will value the v3 additions more than they value Matomo's plugin marketplace.

Ackee — absolute minimum, dev-friendly

Ackee is the smallest serious option and the one most likely to get dismissed unfairly. It does less than Matomo or Umami on purpose. The audience it serves is the developer or small site that wants pageviews, referrers, and a clean API to read them with — and nothing else.

License. MIT. GitHub stars. Around 4,500 on electerious/Ackee as of May 2026. Maintained by Tobias Reich and a community of contributors.

What you get: views, durations, referrers, top pages, devices, browsers, operating systems, languages, and locations. A multi-step anonymisation pipeline removes any persistent visitor identifier at ingestion. The interface is deliberately spare — there is no funnel builder, no goal-conversion model, no e-commerce report, no segmentation engine. If your job involves any of those, pick Matomo. If your job is "tell me which posts get read," Ackee is the right size.

The piece worth flagging is the GraphQL API. Everything the Ackee dashboard shows you is served by it; everything you might want to plug into a custom dashboard is exposed through it. For developers who want to embed analytics into a status page, an admin console, or a Notion-shaped operations doc, that API is the actual product. Most analytics tools treat the API as an afterthought; Ackee treats it as the front door.

The stack is Node.js 22+ and MongoDB 4.4+. Cookie-less by default — same configuration story as the others.

The list of features Ackee deliberately does not have is worth naming, because that list is the case for using it. There is no funnel builder. No goal-conversion model. No e-commerce integration. No campaign attribution. No heatmaps. No session replay. No A/B testing. The deliberate absence is the design — Ackee solves the problem of "how many people visited which pages, where did they come from, how long did they stay" and stops. For a developer running a personal site, a documentation portal, an open-source project page, or a small SaaS landing page, that is the entire question.

Cost stack. Self-hosted: free. No project-run cloud version exists; the only managed paths are third-party. Managed Ackee on DANIAN: €9/month per instance. → Ackee for minimalist tracking

Best for: developers and operators of small sites who want pageviews and an API, not a full marketing-analytics platform.

Comparison table

The numbers turn the post on a single axis: at any production-class workload, the managed-open-source path is the cheapest of the named alternatives. GA4 is free until you need to escape its sampling; the BigQuery linkage and the Consent Mode v2 work that escape costs as much in time as it saves in licence fees. Matomo Cloud is fair pricing for a vendor-run service. Managed open-source on DANIAN is the same Matomo, Umami, or Ackee, on infrastructure you pick from 21 datacenter locations, with patching and backups handled, for €9 per app per month.

How to pick — three questions

The fastest filter is to ignore the feature-comparison list and ask three questions about what you actually use. Most analytics audits end with a sentence the marketer does not say out loud — we use about a quarter of what GA4 exposes. Picking by your real usage rather than by feature-list parity is how you avoid carrying GA4's complexity tax over to the new tool.

Question one: do you need feature parity with GA, including funnels, goals, e-commerce, multi-channel attribution, and a working migration path off GA4? If yes, pick Matomo. The GA4 Importer is the deciding factor; nothing else has it.

Question two: do you want the smallest learning curve and a clean, modern UI for pageviews, sessions, events, and Web Vitals — without GA4's complexity? If yes, pick Umami. The v3 additions (Boards, Session Replay, Web Vitals) make it the most pleasant tool of the three for a marketer who wants to stop fighting their analytics platform.

Question three: do you just need pageviews, referrers, and a clean API to read them with? If yes, pick Ackee. It is the smallest tool that solves the actual problem, and the GraphQL API is genuinely the best in this category.

Picking by usage rather than feature-list parity is how teams end up with Umami when they thought they needed Matomo, and with Ackee when they thought they needed Umami. The cost of switching down is small. The cost of staying on a tool too big for the job is the time spent ignoring the parts you do not use.

FAQ

Is Matomo really a 1:1 replacement for GA4?

For most marketing-team workflows, yes. Real-time, sessions, goals, funnels, e-commerce, multi-channel attribution, and tag management ship in the box. The GA4 Importer pulls up to 13 months of historical data via the Google Analytics Data API. The honest gap is Google Ads attribution depth — Google's own platform integrates with its own ad system in ways no third party can match.

Will I still need a cookie banner?

Possibly not, if you configure the cookie-less tracking mode correctly on Matomo, Umami, or Ackee. All three default to or support a no-cookie, no-PII tracking model. In many jurisdictions, that configuration removes the requirement for a consent banner. This is operational guidance, not legal advice — check with someone who reads regulations professionally before publishing the change.

How does the Matomo GA4 Importer actually work?

It uses the official Google Analytics Data API to pull historical reports — pageviews, sessions, events, conversions, and the standard dimensions — into your Matomo instance. Google's API exposes up to 13 months of history. The plugin runs as a scheduled task in Matomo and writes the imported data into the regular reporting tables, so dashboards and segments work the same way they would for natively-collected data.

Why aren't Plausible or Fathom on this list?

They are excellent products in a different category. Plausible and Fathom are cheaper-than-GA4 SaaS analytics platforms, run by their respective companies on their own infrastructure. They are not what most teams mean by "open-source alternative" because the buyer does not host the software themselves. If a managed SaaS at $9–$19/month is the right answer for you, those are worth a look.

If I self-host one of these, what does it actually cost?

On a $24/month production-class VPS plus backup and monitoring, the infrastructure runs about $44/month. The operational time — patching, backup verification, upgrades, on-call — adds five to ten hours of setup and one to two hours per month. At a freelance sysadmin rate of €60–120/hour, that's another €60–240/month. The math comes out around €100–280/month effective for a serious self-hosted setup.

Can I run more than one of these at once?

Yes, and a few teams do. A common pattern is Matomo for marketing reporting (the funnel data, the campaign attribution) plus Umami for engineering and product teams (the lightweight dashboards, the Web Vitals tracking). At €9 per app per month on DANIAN, running both is cheaper than most single-tool SaaS plans.

What happens to my data if I leave DANIAN later?

The data is yours. The open-source software is yours. Daily backups are kept off-site, exports run on demand over SFTP, and the file format on disk is whatever the upstream project produces — Matomo's MySQL dump, Umami's PostgreSQL dump, Ackee's MongoDB export. You can take that to another host or to your own server and the application keeps working. We publish a "How to leave DANIAN" guide for the same reason a cleaning company leaves the keys on the table — the customer is here because they choose to be.

Matomo vs Umami — which should I actually pick?
Matomo if you came from GA4 and need funnels, e-commerce reports, multi-channel attribution, and the GA4 Importer for 13 months of history. Umami if you want clean dashboards, Web Vitals, and the lightest possible footprint with no migration baggage. Matomo carries more weight; Umami feels lighter for the same job.

Which open-source analytics tool is the easiest to set up?
Umami, by a clear margin. The stack is one Next.js container plus a database — paste a 2-KB tracking script on your site and you have a dashboard within minutes. Matomo takes longer because there is more to configure. Ackee is fast but requires Node.js 22+ and MongoDB 4.4+.

Are Matomo, Umami, and Ackee actively maintained in 2026?
Yes, all three. Matomo ships releases roughly monthly under InnoCraft Ltd. Umami had v3 commits through April 2026 with active web-based Session Replay work. Ackee's pace is slower — community-maintained by Tobias Reich and contributors — but commits land regularly. None of the three is at risk of going dormant in the next twelve months.

How long does a GA4-to-Matomo migration actually take?
Plan for three to five working days end to end on a typical site. The GA4 Importer plugin pulls 13 months of history in the background — that's hours, not days. Real time goes to re-tagging the site, rebuilding the segments and goals you actually use, and verifying counts match GA4 within ±5% before switching off.

Can I run GA4 and Matomo in parallel during migration?
Yes, and it's the recommended approach. Both tracking scripts can fire on the same page; neither interferes with the other. Run them side-by-side for two to four weeks, compare daily numbers in both dashboards, fix any discrepancy you find, and only remove the GA4 snippet once Matomo numbers match within tolerance.

Can I import Universal Analytics data into Matomo?
Universal Analytics stopped collecting in July 2024 and Google retired the API the same month. The official Matomo importer only covers GA4. If you exported your old UA data to BigQuery or CSV before sunset, you can load it into Matomo via the database directly — manual work, but the data is recoverable.

Does Matomo support e-commerce tracking?
Yes — out of the box, no paid plugin required. The e-commerce module covers products viewed, cart additions, completed orders, abandoned carts, conversion rates by traffic source, and average order value. WooCommerce, Magento, PrestaShop, and Shopify integrations exist as official or community plugins. Umami supports e-commerce events too, but the report depth is lighter.

Do these tools support real-time visitor tracking?
All three do. Matomo's "Visitors in Real-time" widget refreshes every few seconds and shows live actions per visitor. Umami's real-time view streams active sessions on a single screen. Ackee shows current online visitors via its GraphQL subscription endpoint. Latency from page-load to dashboard appearance is typically under fifteen seconds on all three.

Can I track multiple websites in one Matomo instance?
Yes — a single Matomo instance handles unlimited sites. Each gets its own ID and dashboard, and you can roll reports up across sites with the (paid) Roll-Up Reporting plugin or by querying the API. Umami also supports multiple websites per instance. On Ackee, each tracked domain is configured as a separate "domain" object.

Does Umami have a public API for embedding stats?
Yes. Umami exposes a REST API documented at umami.is/docs covering websites, sessions, events, pageviews, and metrics — every chart in the dashboard is driven by it. You authenticate with an API key and pull JSON to embed stats in admin panels, status pages, or Notion dashboards. Ackee's API is GraphQL-first; Matomo's is XML/JSON.

How much does the Matomo tracking script weigh — will it slow down my site?
Matomo's tracker is around 9–12 KB gzipped depending on enabled features — smaller than GA4's gtag plus the Tag Manager bootstrap. It loads asynchronously, so first paint is unaffected. Umami's tracker is roughly 2 KB. Ackee's is similar to Umami. None register meaningfully on a Lighthouse score for a typical site.

Does Matomo integrate with WordPress, Shopify, or Webflow?
Yes, all three. WordPress has an official Matomo plugin handling tag injection, dashboard widgets, and user-role permissions. Shopify uses a snippet you paste into the theme's theme.liquid file. Webflow uses its custom-code panel — paste the tracking snippet site-wide. Umami and Ackee follow the same pattern: a snippet in the head tag works everywhere.

Where does my analytics data live when I host it on DANIAN?
You pick the region at deploy time from 21 datacenter locations across six continents. The container, the database, and the daily off-site backups all live in the region you choose. There is no cross-region replication unless you configure it yourself. If you change your mind, the data is yours — exports run on demand over SFTP.

What support do I get if my Matomo instance breaks?
24/7 live chat and email support is included in the €9/month price — no separate support tier, no professional-services upcharge. Typical work covers DNS edits, plugin upgrades that conflict, database tuning, backup restores, and migration help. The same engineer who deploys also debugs; there is no first-line gating. Chat replies usually arrive within fifteen minutes.

Can I give my clients read-only access to their own dashboards?
Yes — Matomo has the most mature permission model: role-based access at the per-site level, with View, Write, and Admin tiers, plus single sign-on via SAML in the (paid) plugin. Umami offers user accounts with website-level access. Ackee has a single admin user, so multi-client setups need one Ackee instance per client.

Conclusion — what to do this week

The migration question is rarely whether to leave GA4. It is which open-source replacement fits the team that is leaving. Three concrete actions answer that question quickly, and each one takes less than an afternoon if the trial app is already running.

Pull the GA4 audit and identify which of the four data-quality issues is biting you. Sampling above the 10M-event threshold, the (other) row hiding the long tail, demographic thresholding, and consent-modelled traffic — name which one prompted the question. The answer narrows the choice.

Try Matomo, Umami, or Ackee on a 7-day trial. Pick by the three-question filter above. Most readers will land on Matomo because the GA4 Importer is the deciding factor; readers who care about clean dashboards and Web Vitals will land on Umami; developers will land on Ackee.

If Matomo is the pick, install the GA4 Importer in week one. It runs in the background and gives you 13 months of history before you finish the migration plan. The window to import is the window before you switch off GA4 traffic — start it early.

The €9 figure does the rest of the work. Pick the region from 21 datacenter locations, deploy, point the tracking snippet at your new instance, and let the open-source tool do what GA4 quietly already does — count visitors — without the sampling, the consent-banner workarounds, or the BigQuery bill.