How we research and score

A review is only useful if you can tell how it was produced. This page describes the rubric we score against, the process every review goes through, and the objective thresholds a review must clear before it goes live.

What we score

Each review carries a breakdown rather than a single opaque number, so you can weight the dimensions that matter to you. A product that scores well on value and poorly on software is a different recommendation than the reverse, even at the same overall score.

Real-world performance: How the product behaves in the situations people actually buy it for, not just on a spec sheet. Sustained output, battery endurance, thermal behaviour and measured throughput carry more weight than peak numbers.
Design and build: Materials, serviceability, port selection, controls and how well the hardware holds up to repeated handling.
Value: Scored against current street price and the closest competitors, not MSRP. A product can lose points purely for being outpriced by an equivalent rival.
Ecosystem and software: App quality, update history, subscription requirements and how much the product depends on a vendor service that could change.

What the numbers mean

Scores are out of 5. Most reviews land between 3.5 and 4.5, but we publish lower scores when a product underdelivers — a low number is a useful warning, not a reason to skip the review. Here is what each band actually signals:

4.5–5.0: Class-leading. We struggled to find a meaningful complaint, and we would buy it with our own money.
4.0–4.4: Strong recommendation with a caveat we name explicitly in the review. Most readers will be happy.
3.5–3.9: Good, but beaten on at least one dimension by something we would point you at first. Buy it on sale.
3.0–3.4: Compromised. Usually a product that works but is priced against much better competition.
Below 3.0: We would steer you away. Either it underdelivers on its own pitch or a cheaper rival does the same job.

Value is scored relative to the rest of its category, so an expensive product is measured against cheaper rivals that do most of the same job. That is why flagships routinely post the lowest value scores on the site even when we recommend them.

How a review is produced

1AuditEvery published review is scored continuously for depth, evidence density and originality. Anything that regresses re-enters the queue.
2Evidence briefSpecifications, pricing and measurements are collected from manufacturer documentation and established testing outlets before any prose is written.
3DraftThe review is written against the evidence brief, with a mandatory competitive comparison and an explicit "who should buy" verdict.
4Editorial reviewA second pass checks that the piece answers a real buying question, commits to a recommendation, and does not hedge.
5Fact checkClaims, specifications and retailer links are verified. Unverifiable claims are cut rather than softened.
6Quality gateAn automated gate blocks publication until the thresholds below are met. It is a hard gate, not a warning.

The publish gate

The gate is a floor, not an average. Every review on the site clears every number in the left column; a draft that misses one is sent back rather than published with a caveat. The right column is what we actually aim for — missing a target puts a review in the revision queue, but we would rather publish something short and dense than pad it out to hit a word count.

MeasureHard floorTarget
Minimum length900w1800w
Distinct sections56
Answered buyer questionsoptional8
Named competitors compared13
Concrete numeric claims612
Maximum similarity to any other review22%—

The similarity ceiling exists to stop near-duplicate reviews. If two products genuinely warrant the same analysis, they belong in one comparison, not two reviews. A question count is deliberately not a gate: it is the easiest number on this page to game, and an FAQ bolted on to clear a threshold helps nobody.

How we make money

Verdict Hound is reader-supported through affiliate links and display advertising, which means a retailer link earns us a commission only where we have an active partnership. We do not accept payment for coverage, for a higher score, or for a place in a ranking, and advertisers get no advance sight of a review. Scores are set before retailer links are attached. Read the full affiliate disclosure.

Corrections

If a specification, price or claim in a review is wrong, we want to fix it. Reviews are revised in place and the update date on the article reflects the change.