Methodology

Our aim is a spec table you can trust over the retailer's own listing. Here is exactly how the data gets here.

How specs are compiled

1. Discovery. We enumerate the product lines each brand actually sells in a category, from manufacturer sites and distributor catalogs.
2. Harvest. We fetch each source page and store a dated snapshot — our provenance backbone, so we can re-check a value without re-fetching.
3. Extract. Specs are extracted from those snapshots into a structured, validated format. Every value keeps a link to the exact source it came from.
4. QA. Automated checks (schema validation, sanity ranges, identifier verification) plus a human/agent spot-check of a sample per category. We verify identifiers against their source and do not publish fabricated values.
5. Load. Validated data is loaded into the catalog you browse here, and re-run as sources change.

Provenance on every value

Each spec carries its source. Product pages show a Sources section and inline source links — click through to see the manufacturer or distributor page a value came from. If a value isn't on a source page, we leave it blank rather than guess.

When sources disagree

Manufacturers, distributors, and marketplaces sometimes list different specs for the same item. We resolve conflicts by precedence — manufacturer > distributor > marketplace — and retain the losing value in the record rather than deleting it, so a correction is always auditable.

Neutrality

The same process applies to every brand. No brand pays for placement, and rankings are honest even when they're unflattering. See about for our independence and disclosure stance.