Pick the analytics tool that fits, even if it is not ours
We read the source and the pricing pages of eight privacy analytics products, then wrote down where each one beats us. Every comparison below names a real reason to choose the other tool.
~30% more visits recorded than consent-gated, cookie-based tools, because the script sets no cookies and is not blocked by a consent decline or an ad blocker.
Problems we built for, and where we stop
What people actually run into with analytics tools, read from the research, before any feature list.
“decisions about landing pages, content and ad spend are being made from numbers with a messed up denominator”
What we do: Every pageview carries a bot label with a reason code; the dashboard shows the top reasons held back.
Where we stop: We label and do not block. Three reason codes fire today. A label is a statement about a signal, not a verdict. Reason-per-bot is not unique: Swetrix, Rybbit and PostHog show one.
“my tag broke and I found out weeks later”
What we do: An automatic notice when a reliably active site goes quiet.
Where we stop: Total silence only, not a drop or a partial break.
“a ten-minute read counted as a zero-second bounce”
What we do: Bounce needs one pageview and under 10 engaged seconds, and the threshold is shown.
Where we stop: Equals GA4's threshold, so new only against the privacy tools.
“GA4 users dropped after the banner and the modelling never brought them back”
What we do: Runs beside GA4 and counts on our own servers.
Where we stop: The capture rate is an unaudited estimate until capture-audit_plan.md finishes.
“an upgrade that will not boot, RAM and ClickHouse for a small site”
What we do: Hosted, no operations.
Where we stop: Not open source and not self-hostable.
“every tool says privacy matters, and very few say anything meaningful about what they actually collect”
What we do: The privacy pages list what is stored and for how long.
Where we stop: The wording is hedged on consent by design, and the visit ID is pseudonymous, not anonymous.
Checked at the source
How we compare, including where we lose
Eight privacy analytics products, checked at source. Where one of them beats us, the row says so.
- Free plan
- VeritaMetrics: 1 sites, 50,000 pageviews a month
- Other products: Umami Cloud gives 1 site. Fathom has no free plan.
- Script size, gzipped they win
- VeritaMetrics: ~8KB
- Other products: Plausible is smaller at 1.3KB. PostHog is 89.0KB.
- Captured without tagging
- VeritaMetrics: Clicks, forms, outbound links, downloads, scroll depth
- Other products: Umami captures none of these.
- Page speed from real visits
- VeritaMetrics: LCP, INP and CLS
- Other products: Plausible collects none. Matomo sells a plugin at 89 EUR a year.
- Rage and dead clicks
- VeritaMetrics: Both, on every plan
- Other products: Of the eight we compared, only PostHog has both.
- Events that fail to send
- VeritaMetrics: Saved on the device and re-sent on the next page load
- Other products: PostHog retries in memory, so a closed tab loses them.
Competitor figures checked 3 August 2026. Products move, so see the full comparison for the current detail on each one.
What Sets VeritaMetrics Apart
Privacy by Default
Our snippet sets no cookies, and no IP address is stored on an analytics row.
Learn moreComplete Data Capture
Ad blocker resilience and a no-JS pixel fallback keep more of your traffic in view.
Learn moreRanked growth fixes
Growth diagnostics that scan your analytics and surface the most impactful improvements.
Learn moreVisit timeline
Privacy-safe journey reconstruction without recording screens. Debug funnels without surveillance.
Learn moreDetailed Comparisons
vs Google Analytics 4
The free, ubiquitous incumbent. Powerful and tied into Google Ads and BigQuery, but it sets cookies, needs a consent banner, and its script is roughly 145KB.
“Bot traffic from Singapore and China is in my reports and I cannot block it.”
vs Plausible Analytics
A cookieless, open-source, EU-hosted favourite with a very light script. Simpler by design, so it leaves out session replay, automatic events, and automated site-health scoring.
“ClickHouse needs several containers and idles using over 4 GB of RAM, for a workload that fits in 150 MB elsewhere.”
vs Fathom Analytics
A simple, cookieless tool with EU isolation on by default and a light script. Deliberately narrow: no funnels, session replay, or automatic event capture.
vs Matomo
The mature open-source platform, free to self-host with heatmaps, A/B testing, and more. Defaults to cookies, needs consent setup, and carries a heavy script.
“Bots are in my reports and the segment does not remove them.”
vs PostHog
A broad product-analytics suite with session replay, feature flags, and experiments, and a large free tier. Heavy (~73KB) and cookie-setting by default.
“Self-hosting PostHog was basically unusable: the deploy script varies from somewhat working to completely broken.”
vs Simple Analytics
A cookieless, EU-domiciled tool with a light script and a conversational AI. No session replay, revenue tracking, or self-hosting.
vs Umami
Open-source and free to self-host, with a small script and recent session replay. Automatic capture is limited, and its Postgres backend has no query rollups, so faster reporting means switching to a self-managed ClickHouse install instead.
“A long read is counted as a bounce with zero seconds on it.”
vs Cloudflare Web Analytics
Free with no traffic cap, cookieless, with a no-JS edge mode, Core Web Vitals, and a privacy statement that explicitly rules out fingerprinting. Basic: no events, funnels, session replay, or CSV export.
“Cloudflare congratulated me on 10k visits last month, but the traffic turned out to be non-human.”
At a Glance
| Feature | VeritaMetrics | GA4 | Plausible | Fathom | Matomo |
|---|---|---|---|---|---|
| Cookieless by default | |||||
| Consent banner claim (vendor claim) | Usually, see /privacy | Claimed, no caveat | Claimed, with caveat | Requires configuration | |
| GDPR position (vendor claim) | Legitimate interest (Art. 6(1)(f)), see /privacy | EU collection, US processing | Requires configuration | ||
| Data storage region | London, UK | Collection only | EU isolation by default | EU Cloud or any region | |
| Live visitor count | Realtime report (last 30 min) | ||||
| Script size (gzipped) | ~8KB | ~ 148 KB | ~ 2.5 KB | ~ 2 KB | ~ 47 KB |
| Ad blocker resilience | Proxy required | Custom domain (undocumented) | Proxy required | ||
| No-JS fallback (pixel) | |||||
| Automatic event capture | Enhanced measurement (8 events) | Outbound, downloads, forms (default on) | Pageviews, downloads, outlinks | ||
| Rage & dead click detection | |||||
| Bot classification, with a reason per visit | Aggregate only, not per visit | Undocumented | |||
| Bounce that counts an engaged single-page read | |||||
| Session replay (DOM recording) | Paid plugin (€219/yr) | ||||
| Visit timeline (event list) | |||||
| Automated insights | Rule-based, no LLM | Automated, custom, anomaly, trend | Rule-based, no AI | ||
| Funnels | Manual (Explore) | Business tier | Paid plugin (€199/yr) | ||
| Revenue attribution | Tracking only, no attribution | Manual tagging | Business tier (manual) | Event values only | |
| Data export | CSV / JSON / XLSX | CSV + BigQuery (1M/day) | CSV (API on Business) | CSV only | Excel / XML / JSON / CSV |
| Self-hosting option | On request (contact sales) | Community Edition (AGPL) | Fathom Lite (frozen) | ||
| Open source | AGPL | Lite only | GPL core | ||
Also considered
Other privacy-first analytics tools worth knowing about.
Pirsch
German and cookieless, from $6/mo. Funnels and goals, no session replay. A Go-native, server-side option with a paid self-host tier.
GoatCounter
Free and open-source (EUPL), self-hostable as a single Go binary. Deliberately minimal: pageviews and referrers, no funnels or replay.
Counter.dev
Pay-what-you-want and open-source (AGPL), EU-hosted. The lightest script here at about 1.1KB, but pageviews only.
Before you choose
- How does cookieless web analytics actually work?
- A visit is counted from a one-way hash instead of a stored identifier. We cut the IP address to its first three blocks, hash it with the User-Agent, the date and a salt that rotates daily, and keep the result. Nothing is written to the visitor device, so there is no cookie to block or clear, and the hash changes every day. For visitors who block JavaScript, a 43-byte pixel records the page view on our server instead.
- Is cookieless analytics as accurate as Google Analytics?
- It is usually more complete, because the two lose data in different ways. Ad blockers and consent refusals remove visits from a cookie-based tool before it counts them, which industry estimates put at 10 to 40% of visits. A cookieless tool counting on its own servers keeps those. What it gives up is cross-day identity: we cannot tell you that the same person returned last week, because the daily hash is designed not to support that.
- Do I still need a cookie banner if I switch?
- That depends on your jurisdiction and on what else you run, and it turns on ePrivacy Article 5(3) and PECR rather than on cookies alone. Setting no cookies answers part of the question and not all of it. We set out our position, and what it depends on, on our privacy pages. It is not legal advice.
- Can I run this alongside Google Analytics?
- Yes, and most teams do. Our snippet counts independently of GA4 and Tag Manager, so nothing needs removing and your Google Ads and Search Console setup keeps working. You can run both for a period and compare the two counts before deciding anything.
See how these features work
Jump to another guide.
Ready to Switch?
Set up VeritaMetrics in under 2 minutes. No cookies, no compromises.