How VantaSpec sources prices
Where our prices come from, what we verify before publishing one, and what we deliberately refuse to show. Including the checks that exist because we got it wrong.
A price comparison site is only worth using if its prices are real. This page explains exactly where ours come from and what we check, because the failure modes in this category are not obvious from the outside.
Where prices come from
Some retailers send us a product feed through an affiliate network, which we read every day. The rest we check ourselves, a retailer a night in rotation, reading the price from the retailer's own product page. For each product we hold a link to a specific page at a named UK retailer, and every price links back to it.
When we read a page we prefer structured data, the machine-readable product information retailers publish for search engines, over the rendered page, because it is what the retailer intends to be authoritative and it does not silently break when they redesign.
What we verify before publishing a price
The link must point at a product page. A price whose link goes to a search results page is not a price for that product. We require the product to be identified in the URL's path.
This matters because it is the signature of fabricated data. When we audited an earlier version of this catalogue, 1,580 of 1,591 prices were not real: nine retailers, values clustered around a plausible mean, and every link pointing at a search query rather than a product. Those records are gone, and the check that catches them runs on everything we ingest.
The page must sell the same kind of product. A wrong match will otherwise pass every other test: a real retailer, a real product page, the right brand, a real price. We compare the retailer's own description of what they are selling against the category we filed the product under, and reject the offer if they disagree.
This one is not theoretical either. A PCIe sound card was matched to a gaming headset and published with the sound card's photograph, because the listing mentioned a headphone amplifier. It cleared every other check we had. The check that catches it now also discards the image, because both came from the same page.
The value must be plausible for the category. Prices far outside the sane range for a product type are rejected rather than shown.
What we refuse to show
- Prices we cannot trace to a specific product page.
- Images from sources we cannot verify. A product image must come from the manufacturer's own site, the retailer page we read the price from, or a recognised product data provider. We do not take images from search results, marketplaces, or review sites.
- Products identified only by a series name. "Asus TUF Gaming" is a product line, not a product. A page about a line no one can buy is not useful.
What we are still working on
Being straight about the limits:
Prices are refreshed, not live. Feeds are read daily and the rest are re-checked in a nightly rotation, so a price can be up to a few days old. Each one carries the time it was last checked: follow the link to confirm before buying.
Coverage is uneven. We hold specifications for far more products than we hold prices for, because a price requires a UK retailer to currently stock the item.
We do not list every retailer. We list the ones whose product data we can read reliably enough to trust.
Why we publish this
Anyone can claim their prices are accurate. The checks above exist because specific things went wrong and were caught, and describing them honestly is more useful than a promise. If you find a price that is wrong or out of date, we would rather hear about it. Get in touch.