How Worthivo chooses.

Worthivo surveys public owner discussion wherever it accumulates—retailer reviews, forums, community threads, comment sections, published tests—looking for recurring patterns rather than treating a single comment as consensus.

That survey decides what gets investigated. It does not decide the outcome. Attention is not evidence that owners were satisfied: a marketing budget buys discussion, and so does a seeded unit, an influencer window, or a commission-paying marketplace. A product’s presence on this site means it was worth investigating. What the investigation found is on the page.

Owner language is paraphrased unless Worthivo has explicit permission to quote it.

Before research starts, the discussion itself is screened, and the answers and their dates go in the research record. Does it appear across sources that do not cite one another and share no commercial interest in the product? Does it include use over time, not only unboxings? Do owners describe specific, varied circumstances, or repeat similar phrasing? Is any of it traceable to a launch window or an undisclosed seeding campaign? What does the dissent look like?

The screen routes to a decision. Too little discussion of any kind to research, and the product is dropped and the decision logged—no page. Substantial discussion with documented evidence that the discussion itself was procured or falsified is a candidate for Not Worthivo. That finding is about the discussion, not about the product, and a page carrying it says so.

A disclosed review-sample or press-loan programme is normal industry practice. It is weighed as a question about how independent a source is, never as evidence that a company faked its own praise. Where authenticity cannot be checked, that is recorded as unchecked. Inability to verify is never treated as proof of manufacture, and never counts against a product.

The favourable and the unfavourable material are read in the same pass, with the same scepticism, against the same bar. A softer test for negative findings would make the two verdicts that can earn a commission cheap and the one that earns nothing expensive. The bar is identical in both directions for that reason.

A finding moves a verdict only where it appears across at least three sources independent of one another, spanning more than one venue. No claim about a product rests on a single account in either direction—one delighted owner is not a recommendation, one angry owner is not a defect—unless that source is an identified institutional record: a recall notice, a regulator finding, a manufacturer service bulletin, a published standards test. Where the question is how the product holds up, the evidence includes owners reporting after extended use. The page names the model, revision, market, date range, and the longest ownership the evidence actually covers.

Specifications are checked against the manufacturer or another identified primary source. Every fact carries its source and its date, kept visibly separate from Worthivo’s judgment. Worthivo publishes no figure implying it surveyed or aggregated owners, and no retailer or marketplace star rating, review count, or rank position appears on a Worthivo page in any framing.

Research ends in a stated verdict, and the reader is given it in the opening rather than after four sections of caveats. A verdict attaches to the product once, and every page covering that product shows the same one.

Worthivo Recommended. Weighed against the whole evidence set, the evidence runs favourable for the ordinary buyer of what the product is sold to do. An affiliate link is permitted, and disclosed.

Worthivo Conditional. The evidence resolved, and runs favourable only inside a condition the research can name from that evidence—a condition that does not exclude the ordinary buyer. An affiliate link is permitted, and disclosed. Conditional is not a gentler negative, and it is not somewhere to park research that never resolved.

Not Worthivo. Either a corroborated pattern of owner-reported problems ran across independent sources and, weighed against the favourable evidence, defeats the product for the ordinary buyer—or the attention around it is documented as manufactured. There is no affiliate link, anywhere on this site, including as an alternative on another product’s page. The page states that there is none, and why.

Worthivo can publish a Not Worthivo, and will. A product with no affiliate programme is judged exactly as any other, and the absence of a link is never a reason a page fails. A product that reached Not Worthivo is never dropped from a comparison or a shortlist to leave only the linkable options standing.

The verdict is fixed in the research record, and dated, before the page is written. Then Worthivo looks for what would overturn it, whichever way it went: for a favourable verdict, recalls, later-revision regressions, long-term owners who turned against it; for an unfavourable one, firmware fixes, corrected listings, satisfied long-term owners, superseded revisions. The searches and their results are recorded, including the ones that found nothing.

Where that pass turns up something that clears the bar, the verdict returns to the record and is decided again before drafting resumes. Publishing the contrary evidence does not substitute for re-deciding. Where a verdict changes, the record keeps the original, the new evidence, the direction, the date, and who decided.

Where a fault pattern meets the bar, the specific findings go to the manufacturer in writing, dated, and the page publishes either the response or the fact that Worthivo asked on one date and had no reply by another. Silence is published as silence, and is never itself a reason for a verdict. Every verdict, all three tiers, is signed off against the record by someone who did not write the page.

A product does not need to suit everyone. Whatever the verdict, the page names who it fits, who should skip it, and what inconvenience a buyer has to accept.

Every complaint a page raises has to end somewhere: in a buyer it rules out, in a stated reason it did not move the verdict, or in the verdict itself. A complaint left hanging as general doubt is a failure. So is a complaint narrowed until it stops counting—where a corroborated problem defeats what the product is sold to do for the ordinary buyer, that is not a boundary, it is a Not Worthivo.

Worthivo has not hands-on tested the products it writes about, and says so on every review.

A number is not a verdict. Worthivo’s conclusion is one of the three verdicts above, written in words, and no score produces one, outranks one, or stands in for one. No figure drawn from commercial value—commission rate, cart value, research priority—reaches the record, the draft, or the page.

The earlier method published an editorial score alongside each product. No page carries one now: the scores were retired when this standard came into force, because a number that looks like a verdict competes with the verdict whatever the small print says. Third-party star ratings are never represented as Worthivo’s own aggregate rating, and are not published in any other framing either.

Worthivo adopted this standard on August 22, 2026. Any page published before that date is re-researched under it or taken down. Until that happens it carries no verdict badge, because there is no verdict.

A badge appears only where the research record and the independent sign-off behind it exist, and for a page still awaiting re-research neither does. The earlier method treated a product’s presence here as the finding that it was good: a page could not conclude against a product, because a product that would have failed was never written up. Evidence gathered that way is not a verdict under this standard, however carefully it was gathered. It predates the authenticity screen, the falsification pass, the approach to the maker, and sign-off by someone who did not write the page.

Those pages keep the affiliate links and the disclosure they were published with. Each one is re-researched under this standard or taken down. Until that happens, read them as what they are: work done to an older method, still on the site, and not a verdict.

Source dates appear near time-sensitive evidence. A material correction changes the page’s updated date and says what changed.

Affiliate compensation does not purchase a Worthivo verdict or ranking. Commercial interest may choose what Worthivo investigates. It may not appear in what Worthivo concludes. The verdict decides the link; the link never decides the verdict.