Review Methodology

Last updated: August 17, 2026

How every number on this site is obtained, checked and published — and, just as importantly, the conditions under which a number is thrown away rather than published with a caveat.

1. Where every figure comes from

Specifications are collected from retail product listings across each category, at a scale large enough to describe the category rather than a handful of products. Each figure is recorded together with the exact source text it was read from, so any published number can be traced back to the words that produced it.

There is no other source. We do not test, measure or install. Where a manufacturer’s own documentation contradicts a retail listing, the manufacturer’s page is the better source and you should treat it as authoritative over anything here.

2. Two gates every figure must pass

Gate one: coverage

A specification is published for a category only where it appears on at least 40% of that category’s products and on at least eight products in absolute terms. Both thresholds matter. A median drawn from four listings is not a category baseline, and a field present on a small minority describes the brands that happen to fill in a form rather than the market.

Where a field misses the gate, it is listed as withheld with its actual coverage stated, so you can see the difference between “we did not look” and “the data is not there”.

Gate two: units

This is the check that matters most, and it exists because retailers file different physical quantities under the same specification slot.

Every extracted figure is re-read against the source text it came from, and the unit in that text is compared against the units the field can legitimately hold. Where fewer than 80% of a field’s values survive, the field is dropped for that category entirely.

3. What we publish instead of averages

Comparison pages publish quartiles — the 25th percentile, the median and the 75th percentile — and not means.

The reason is specific rather than stylistic. Every contamination pattern we have found in retail specification data lands in the upper tail: a capacity stated in the smaller unit, a mains voltage recorded where a battery voltage belongs, an accessory listed as a machine. A mean is maximally sensitive to exactly those values and a median is barely affected by them. Where a maximum in our data is obviously not a real product, we say so beside the table rather than quietly trimming it.

4. What we refuse to publish

5. Where a specification is silent

The correct value is “Not stated”. It is never a plausible default.

This sounds obvious and is the single most common way specification content goes wrong: a column is created, most products populate it, and the remainder are quietly filled with whatever the category typically has. The result reads as data and is invention. Legitimate rows on this site carry “Not stated”, and a table with several of them is telling you something real about how a category documents itself.

6. Category premise checks

Before a category page is written, we confirm the corpus can actually support it. A category whose signature specification is not published at a usable rate does not get a comparison page, and the finding is stated on the hub instead. Publishing a thin table would imply coverage the data does not have.

7. Updates and corrections

Each page carries the date its figures were last refreshed. Specifications change without notice, so a page is a snapshot rather than a live feed. Errors are corrected in place with the date updated; we do not silently revise. Report anything that looks wrong to editorial@rvmonitordata.com with the page and product named.