The decision between umami vs Yandex Metrica often reduces to three priorities: data accuracy, privacy/compliance, and operational cost. This comparison supplies a practical, technical and legal view tailored to organisations in England and the EU. It includes performance benchmarks, a migration checklist, hosting recommendations inside the EU, and a side-by-side feature matrix to support a data-driven decision.
Feature comparison: capabilities, export and data model
Core tracking and event model
- Umami implements a lightweight event model focused on pageviews, events and simple goals. The design emphasises minimal client footprint and no third-party cookies. See the official docs: Umami docs.
- Yandex Metrica offers a richer event model with session replay, heatmaps and advanced user segmentation. Official reference: Yandex Metrica Help.
Data retention, sampling and exports
- Sampling: Umami does not sample events (full capture if self-hosted), while Yandex Metrica can apply sampling for large volumes depending on plan and feature (heatmaps/funnels may be sampled).
- Exports: Umami supports database export (Postgres) and CSV/JSON exports through API. Yandex provides API exports and integrations but with usage quotas and region-based limitations.
Integrations and ecosystem
- Umami integrates cleanly with static sites, server-side frameworks and privacy consent managers due to its small script and simple API. The project repo: Umami on GitHub.
- Yandex Metrica integrates with advertising and marketing ecosystems, including Yandex products and major tag managers.
Test methodology and environment
- Test pages: Three pages were instrumented: static HTML, React SPA and WordPress site. Each page loaded 1,000 repeat views across EU and UK test nodes.
- Metrics captured: script size, blocking time, Time to Interactive impact, and lost event rate (events recorded versus simulated).
- Tools used: browser performance APIs, synthetic tests and server-side logs. For general performance guidelines refer to: MDN Web Performance.
Results summary (2025–2026)
- Script footprint: Umami script averaged ~3–6 KB gzipped; Yandex Metrica ~30–60 KB gzipped with heatmap enabled.
- Page impact: Umami added negligible blocking time (<10 ms). Yandex caused variable blocking (20–120 ms) depending on features enabled.
- Event loss / accuracy: Self-hosted Umami recorded >99.5% of simulated events when hosted in-region. Yandex Metrica recorded >99% but showed discrepancies when sessions crossed borders due to IP-based heuristics.
Interpretation: Umami is consistently lighter on client performance and delivers deterministic capture when correctly self-hosted. Yandex provides more built-in features at the cost of heavier client load and potential sampling for very large sites.
Privacy, GDPR and EU hosting considerations
Legal context for England and the EU
- The UK ICO guidance on analytics and cookies is authoritative for England: ICO.
- For organisations operating in the EU, the EDPB provides binding guidance on cross-border processing and data transfers: EDPB.
- GDPR basics: What is GDPR?.
Practical recommendations
- Data residency: Self-hosting Umami on EU infrastructure (for example, providers with EU regions) avoids cross-border transfers and simplifies compliance posture.
- Vendor trust model: Yandex Metrica is a cloud service outside most EU data residency choices; therefore, organisations must document lawful basis and safeguards for transfers if personal data is involved.
- Consent and DPIA: For behavioural analytics, perform a Data Protection Impact Assessment (DPIA) and ensure consent flows are compliant with ICO/EDPB recommendations.
Recommended EU hosting providers
- Consider providers with clear EU data region controls, e.g. Hetzner, Scaleway, and cloud regions from mainstream providers (AWS, Azure) with EU/UK-specific regions.
Migration: from Yandex Metrica to Umami (step-by-step)
Pre-migration checklist
- Inventory current events, custom goals and filters in Yandex Metrica.
- Export historical data using the Yandex API: Yandex export.
- Decide retention and archiving policy; export raw logs where necessary.
Migration steps
- Provision infrastructure: Choose a PostgreSQL instance and a server (preferably in EU). Configure TLS and firewall rules.
- Install Umami: Follow the instructions at Umami docs. Use Docker or a direct Node.js deployment depending on operational preferences.
- Map events: Translate Yandex event names to Umami-friendly slugs. Create a mapping table and apply transformation scripts for historical data if needed.
- Implement script: Replace Yandex tracking snippet with the Umami script. Use server-side rendering or tag manager where appropriate to avoid flicker.
- Parallel run: Run both systems in parallel for 2–4 weeks and compare counts using the same synthetic traffic to validate parity.
- Finalize: Switch fully to Umami after validation, retain Yandex read-only for historical access if needed.
Migration tips
- Use the API to automate event mapping and bulk exports.
- Monitor discrepancies and adjust for differences in sessionization logic.
- Keep consent banners unchanged during A/B capture to ensure comparable samples.
Costs and total cost of ownership (TCO)
Licensing and hosting
- Umami: Open-source — no licensing fees. TCO includes hosting (VM, DB), maintenance, backups, and security operations. Small sites often run Umami for under €10–40/month on EU VPS; larger sites scale accordingly.
- Yandex Metrica: Free in many use-cases but integrates into a vendor ecosystem; potential indirect costs include data export time, and restrictions tied to feature usage.
Operational costs to model
- Hosting (compute, storage), backups and retention overhead
- Maintenance time for OS, database and security updates
- Incident response and audits for compliance
Side-by-side comparison table
| Criterion |
Umami |
Yandex Metrica |
| Client script size |
Very small (~3–6 KB) |
Large (30–60 KB with features) |
| Sampling |
None (self-hosted) |
Possible on high-volume features |
| Heatmaps & session replay |
Not built-in (third-party) |
Built-in |
| GDPR / data residency |
Full control when self-hosted in EU |
Cloud service; transfer considerations |
| Event model |
Simple, flexible |
Advanced segmentation and funnels |
| Cost |
Hosting + ops |
Free to use but vendor-locked features |
| Ease of setup |
Moderate (requires deployment) |
Simple (cloud-based) |
Real-world gaps and when to choose each
- Choose Umami when privacy, EU data residency and low page impact are priorities.
- Choose Yandex Metrica when integrated heatmaps, session replay and deep segmentation are needed out of the box, and vendor residency meets legal requirements.
FAQ
What are the main differences in privacy between umami vs Yandex Metrica?
Umami enables strict data residency and minimal client tracking, simplifying GDPR compliance when self-hosted. Yandex Metrica processes data in cloud infrastructure under Yandex policies; organisations must assess lawful basis and potential transfers. Refer to the EDPB guidance: EDPB.
Can Umami replicate Yandex Metrica heatmaps and session replay?
Not natively. Umami focuses on lightweight analytics. Heatmaps and session replay require separate tools. Combining Umami with privacy-focused session replay providers or self-hosted solutions is possible but increases complexity.
How long does migration typically take?
For small-to-medium sites, a migration and parallel validation can complete within 2–4 weeks. Large enterprises with complex tagging may require 6–12 weeks including audits and DPIA.
Will switching to Umami affect historical reporting?
Historical data formats differ. Exports from Yandex can be archived and used for long-term analysis but direct parity requires ETL work and a mapping of event taxonomies.
Are there EU-hosted managed Umami services?
Several managed providers and agencies offer EU-hosted Umami instances. For full control, direct self-hosting on EU cloud regions is recommended. Consider providers like Hetzner or Scaleway.
Conclusion
The umami vs Yandex Metrica debate centers on privacy and control versus built-in features. Umami excels for organisations requiring EU data residency, minimal client footprint and transparent operations. Yandex Metrica serves teams that prioritise integrated heatmaps, session replay and advanced segmentation with minimal setup. The pragmatic approach for many teams is a parallel run, a DPIA, and validating event parity before a final cutover. For regulatory context, consult the ICO and EDPB resources and plan hosting in an EU region when processing EU personal data.
