DoVPN editorial methodology
How DoVPN tests and reviews VPNs
A VPN review should show what was tested, what the provider says, and what remains uncertain. DoVPN records those separately. We do not award a final score until the required hands-on tests and critical evidence are complete.
The same privacy, security, performance, app, support, and transparency checks apply to every VPN.
One declared use case, such as privacy or censorship resistance, supplies the final thirty points.
A final score also requires every critical criterion. Missing evidence stays visible instead of becoming a zero.
We review one product at a time
A provider name can cover several products with different infrastructure and responsibilities. A managed subscription, a free service, and a self-hosted client do not share the same server evidence. Each review therefore names its product scope, plan, apps, test dates, and purpose before scoring begins.
How the score works
The universal core is worth 70 points. A purpose module is worth 30 points. Every criterion has a fixed weight and a 0 to 5 anchor. Its contribution is the weight multiplied by the anchor score and divided by five. This prevents a reviewer from changing the importance of a criterion after seeing the result.
The 70-point core supports provider-to-provider comparison. Full 100-point scores are most useful when the reviews use the same purpose module. DoVPN publishes the core and purpose subtotals so readers can see the difference.
| Core criterion | Points | Final-score requirement |
|---|---|---|
| Logging policy and data handling | 7 | Critical evidence required |
| Independent security assessment | 7 | Coverage threshold |
| Operator identity and jurisdiction | 4 | Critical evidence required |
| IP, DNS, IPv6, and WebRTC leak resistance | 10 | Critical evidence required |
| Kill switch and connection-failure safety | 7 | Critical evidence required |
| Protocol and cryptographic design | 5 | Critical evidence required |
| Speed and connection stability | 8 | Coverage threshold |
| Apps, setup, and everyday usability | 5 | Coverage threshold |
| Pricing, renewal, and refund transparency | 4 | Coverage threshold |
| Support quality | 3 | Coverage threshold |
| Security and availability incident handling | 5 | Coverage threshold |
| Technical and organizational transparency | 5 | Coverage threshold |
| Universal core | 70 |
Evidence is labelled, not blended
Provider documentation is useful for prices, supported platforms, stated policies, and technical design. It does not prove real-world behavior. Independent reports can test important claims, but their scope and date still matter. DoVPN records each result as verified, a provider claim, an independent report, conflicting, not tested, or not applicable.
Every scored result links to its evidence and carries a check date. A grade describes evidence quality, while the 0 to 5 anchor describes product performance. They answer different questions and are never combined into one opaque number.
Repeatable hands-on tests
Test records include the app and operating-system version, device, network, server, protocol, location, time, and raw evidence. We keep failures as well as successful runs. A single clean screenshot cannot replace repeated testing.
- Leaks: three runs covering public IP, DNS, IPv6, WebRTC, reconnect, and a network change.
- Kill switch: forced tunnel failure, app termination, reconnect, and Wi-Fi or mobile-network changes.
- Speed: an unprotected baseline followed by repeated nearby, middle-distance, and distant server measurements.
- Apps: installation, account setup, connection state, settings, recovery, and accessibility checks on claimed platforms.
- Support and refunds: recorded technical and billing questions, plus a real refund request when the review says the refund path was tested.
Purpose modules
Before testing, the reviewer selects one primary purpose. The current modules cover general use, privacy, censorship circumvention, streaming, and torrenting. Each uses its own fixed 30-point criteria. A censorship review, for example, needs evidence from named restricted networks. Protocol marketing alone is not enough.
App-store data and screenshots
Apple App Store and Google Play metadata can show current app versions, release dates, rating counts, distribution, and the seller or developer name. Store screenshots can also document the interface at a point in time. Our collection process saves the storefront, retrieval time, source URL, and original image.
Store ratings belong to the listed app, not automatically to the paid product under review. They may combine free, self-hosted, and subscription users. DoVPN displays that limitation and never turns a third-party store rating into a DoVPN score.
Updates, corrections, and preview reviews
Price evidence normally expires first. Security tests, policies, audits, incidents, and app releases have their own refresh periods. Automated checks flag stale records, but a reviewer must confirm material changes and rerun affected tests.
An unfinished review can be rendered as a private working preview. It stays out of search indexing, the sitemap, site navigation, author listings, and publication feeds. It shows the missing evidence and carries no final score. Publication requires the coverage threshold, every critical result, current evidence, completed editorial checks, and a clear verdict.
Commercial relationships do not change the result
DoVPN may earn a commission from some provider links. Commercial status is kept outside the evidence and scoring records. A provider cannot buy a score, suppress a documented failure, or review the article before publication. Factual corrections are welcome and are checked against the underlying evidence.
To report an error or provide new evidence, use the contact page. See the affiliate disclosure for funding details.
