How We Review

This page sets out, step by step, how we research a product, how we reach a verdict, and what our work can and cannot tell you. We publish it so you can judge our conclusions for yourself.

Our six-step process

Step 1: Identify the product and the question

We start by pinning down exactly what is being examined: the product, model or product type, who makes it, and what buyers are trying to find out. That is usually one of three questions. Is a reported problem real? Is the product worth the money? Is there a better option for someone affected by its weak point?

Step 2: Gather the evidence

We collect what is publicly available and note where each piece came from:

• Manufacturer information: specifications, manuals, warranty terms and the company’s own claims
• Owner feedback, across more than one retailer where possible, and retailer questions and answers
• Official recall notices and safety records, such as those published by the US Consumer Product Safety Commission
• Regulatory guidance where it is relevant, for example from the Federal Trade Commission
• Credible news reporting and the published findings of independent testing organisations, credited and linked

We summarise what we read in our own words. We do not copy customer reviews, and we do not reproduce another publication’s work.

Step 3: Identify recurring patterns

One angry review is an anecdote. The same fault described by unrelated buyers, in their own words, is a pattern worth reporting. When we sort through feedback we separate:

• One-off complaints from recurring ones. A single bad unit tells you little. A fault that keeps coming up tells you more.
• Serious issues from minor inconveniences. A safety hazard and a stiff zip are not the same kind of problem, and we do not present them as if they were.
• Product problems from seller problems. Late delivery, damaged packaging and a poor third-party seller are real frustrations, but they are not faults in the product.

We describe what we found in measured terms, such as “some buyers report” or “a recurring complaint is”. We do not publish counts, percentages or failure rates, because public reviews cannot support them: unhappy buyers are more likely to write a review than content ones, so a list of complaints is not a measure of how many units fail.

Step 4: Consider the positive evidence

We do not only look for what goes wrong. We record what satisfied owners praise, what the product does well for its price, and anything the company has said publicly about a reported issue. Every review we publish sets out what buyers like alongside what they complain about.

Step 5: Compare alternatives

Where a product has a recurring weakness, we look for options that may suit the buyers it affects. We suggest an alternative only when the evidence indicates that it avoids the main complaint, and for each one we say who it is best for, why it is worth considering and what its own drawback is. Alternatives are not ranked by commission, and we do not present any of them as flawless.

Step 6: Reach a verdict

The verdict has to follow from the evidence above it. Each one answers four questions in plain language:

• Who may this product suit?
• Who may want to consider alternatives?
• What are the main drawbacks?
• Is there an alternative that may be a better fit?

Our verdict labels

Reviews carry one of five labels. They are deliberately in words, not numbers, and they are not all negative: a site that only ever found fault would not be worth trusting.

Guides and recall explainers do not carry a label, because they are not verdicts on a single product.

When we give a numeric rating

Most of our articles carry a verdict label and no score. A number suggests a precision that research into public information rarely justifies, so we only give one when there is enough evidence to explain every part of it. When we do, the product is scored from 1 to 5 in up to six areas:

• Quality: materials, build and finish.
• Performance: whether it does the job it is sold to do.
• Value: what you get for the price compared with rivals.
• Durability: how it holds up over time, according to longer-term owners.
• Ease of use: set-up, daily use and maintenance.
• Customer experience: warranty terms, returns and how the company handles problems.

The overall rating is an editorial judgement informed by those scores, not an average, because a serious weakness in one area can outweigh small strengths elsewhere. The article explains the reasoning, and an area is left out if we do not have the evidence to score it.

Sources and evidence

Research articles end with a section called Sources & Evidence that lists where the facts came from, with links so you can check them yourself. Inside the article, a specific complaint or incident carries its own source where we have one. We try to keep reported information and our own interpretation visibly separate: what a recall notice says is reported fact, and whether the product is still worth buying is our opinion.

What our work cannot tell you

• How an individual unit will perform. Products vary, and so do the ways people use them.
• How often a product fails. Nobody outside the manufacturer has that figure, and public reviews cannot supply it.
• Whether a particular review is genuine. We read reviews critically, but manipulation is not always visible from the outside.
• What has changed since the date at the top of the article. Products are revised, and recalls are announced without warning.

Our articles are opinion based on research. They are not professional, legal or safety advice.

How affiliate links work

Some product links are affiliate links. As an Amazon Associate we earn from qualifying purchases, at no extra cost to you. Commissions never decide what we cover, what we conclude or which alternative we suggest. Affiliate links are marked for search engines as sponsored, and any article that contains them says so at the top and again beside the links. See our affiliate disclosure.

Sponsored content

We do not publish paid reviews, and companies cannot pay to change a verdict or to be listed as an alternative. If we ever publish content that a company has paid for, it will be clearly labelled as sponsored at the top of the page.

Updates and corrections

Products change, and so does the evidence. When we revise an article in a meaningful way we update the date shown at the top and add a short note saying what changed. We do not change a date just to make an article look fresh. If we get a fact wrong we correct it and say so. See our corrections policy.

How companies can respond

A company that believes our coverage is inaccurate or incomplete can send a response. We review every one, correct anything that is wrong, and may publish the company’s statement alongside the article, clearly labelled as its own. A response does not buy a better verdict, and we do not remove criticism that the evidence supports. See company response.

What we will and will not do