Data methodology

Every product, specification, price and image on Comparezi records where it came from, when it was retrieved and when it was last verified.

Where data comes from

• Manual entry by the team, from manufacturer documentation and official listings.
• Imports from licensed feeds and files (for example merchant or manufacturer product files).
• Demo data, used only while the site is being built. It is fictional and every demo page says so. Demo pages are never submitted to search engines.

We do not use scraping as a data source.

Making values comparable

Values are stored in base units (bytes, grams, millimetres, watts, milliamp-hours) and shown in the unit that suits each category, so a 5,000 mAh battery means the same on every page. Brand names and model numbers are normalised during import, and duplicate products are matched on GTIN/EAN/UPC or brand plus manufacturer part number. Uncertain matches go to a person to review; they are never merged automatically.

Prices and deals

Each price we see is recorded with its time. A deal is shown only when the current price is at least 5% below the 30-day average of daily lows, or is the lowest we have recorded. Flat prices are never shown as deals. Prices can change after we record them; the seller's price at checkout is the one that counts.

How long we keep data

Each source has its own retention rule set by its licence. Some sources allow data to be kept for only a short time or do not allow images to be stored; those rules are applied per source, not site-wide.

Freshness

Records that have not been verified within their source's refresh period are flagged for re-checking, and offers that stop updating are withdrawn rather than left showing an old price.