Telling a tested review from a rewritten specification sheet

The theory of identifying genuine product testing is well covered elsewhere. This is about the version you meet in practice.

The signature of a rewritten sheet

Every claim in the piece can be traced to the manufacturer's own materials, with no figure appearing that was not published. Descriptions of performance are adjectival rather than comparative, since no comparison was made.

Images are official product renders rather than photographs of a unit in a real setting. The structure follows the manufacturer's feature list in the manufacturer's order, which is a strong tell. Faults are absent or generic, because a writer who has not used the product cannot find specific ones.

What genuine testing produces

It produces measurements the manufacturer did not publish, such as timings, temperatures, consumption or noise readings. It produces comparisons against other products tested in the same conditions, which is far more informative than absolute figures.

It produces specific irritations that only appear in use, such as a control that is awkward or a part that is hard to clean. It produces photographs showing the unit in a real environment, including the parts marketing images avoid. It produces statements about what could not be tested, which is the mark of someone who knows the limits of their method.

Assessing a stated method

A published method can be examined for whether it measures what it claims and whether conditions were controlled. Sample size matters, and most consumer reviewing tests one unit, which cannot distinguish a bad example from a bad design. Repeatability matters, and a single run of a test tells you less than several with the spread reported.

After the warranty ends, blinding matters where subjective judgement is involved, and it is rare outside formal testing organisations. A method that can be criticised is more valuable than a conclusion that cannot, which is the whole argument for publishing one.

Video reviewing and its particular tells

Footage of the product being used, disassembled or measured is evidence that the reviewer had it. Continuous shots showing a complete operation are harder to fake than cut sequences of short clips.

Screen-recorded specification lists with a voiceover are the video equivalent of a rewritten sheet. Reviewers who show failures, mistakes and setup problems are demonstrating that filming happened during real use.

Long-term follow-up videos are the most valuable and the least common, for the same economic reasons that apply to text.

Machine-generated and templated content

Automatically generated product pages have proliferated and typically combine specification data with generic phrasing. Repeated sentence structures across many products in a category are the clearest indicator of templating. Content that never mentions a specific measurement, a specific fault or a specific comparison is unlikely to reflect any use.

Author pages with no history, no photographs and no other subject expertise are a further signal. This is a fast-moving area and the signals change, so the durable test remains whether specific new information is present.

Where a claim here has not been tested directly it is inference, and it should be read as inference.

Building a shortlist of sources

Identify two or three sources in each category that publish methods and buy their own samples, and return to them. Note which reviewers describe failures and limitations, since that willingness is the most reliable long-term indicator. Prefer comparative testing over individual reviews wherever it exists, because it controls conditions across products.

After the warranty ends, keep a note of sources that proved wrong, which is uncomfortable and improves calibration faster than anything else. This site does not test products; it explains what to look for, and the testing has to come from someone who owns a meter.

The takeaway

Look for a number the manufacturer never published; that is the shortest test of whether testing happened.

Buy for the failure you can live with, not the feature you will use twice.