How we review

Version 2.0 · effective 22 September 2026

The most useful thing a small publication can tell you is where its knowledge ends. This page describes what we actually do, and — more importantly — what we do not do, so you can weigh what you read here against sources that do more.

What we actually check

1. Feature set against the vendor’s own current pages. Which tier includes what, which features are platform-specific, how many devices a licence covers. This is where most published comparisons go wrong, because they copy an older summary instead of reading the source. It is where we found the errors we corrected on the Norton page.
2. The gap between the marketing and the product. Brand-level marketing describes the whole range; the thing you are buying is one tier of it. We check what the specific tier includes, not what the brand is associated with.
3. Commercial terms. Introductory price versus renewal price, term length, how automatic renewal is switched off, what the refund policy says and what statutory rights interact with it.
4. Platform and support reality. Which operating system versions are still supported, which features are Windows-only.
5. What the published independent testing shows — read at the laboratory, in the current round, and cited as a pointer rather than reprinted as a score.
6. What the product does not do, and what the reader still has to do themselves. This section is in every review, because it is the part advertising omits.

What evidence we accept

• Primary vendor documentation for feature and support claims.
• Published laboratory methodology and results from AV-TEST, AV-Comparatives and SE Labs for protection claims.
• The instruments themselves for anything legal — the GDPR, the ePrivacy Directive, consumer rights directives.
• Official public threat reporting, such as ENISA’s annual Threat Landscape and Europol’s IOCTA, for the shape of the threat environment.

We do not accept a vendor’s press release, an influencer’s claim, or another affiliate page as evidence of anything.

What we deliberately do not publish

• Scores and star ratings. A number implies a measurement. We have not made one.
• Laboratory scores quoted out of their round. They decay; a figure from an old round on a page with a buy button is marketing dressed as evidence.
• Prices. They differ by country, tier and promotion and change constantly. We describe the pricing mechanism — introductory term, then renewal at a higher standard price — and send you to the vendor for the number.
• Specifications we have not verified. Including minimum system requirements, which change as vendors drop support for older releases.

Limits we declare

• We are a small publisher, not a laboratory.
• We earn commission on the product this site’s main article is about. That is disclosed at the top of the page, before the content, and under every button. You should read what we write with that in mind — which is exactly why we also print the product’s limitations and the free alternatives on the same page.
• We have no visibility of what any reader does. With no analytics, we cannot A/B test a conclusion into existence, and we could not tell you how many people read this sentence.

How to check us

Every article lists its sources. Where a claim is attributed to a vendor page or a laboratory, go and read that page or that report — they are free. If what you find disagrees with what we wrote, tell us at info@pravator.online and we will follow the corrections procedure.