Methodology · version 0.3
How the register score works
Each branch gets a score out of 100 from four parts. Scores compare a branch with others in the same postcode districts and price bands, so a Bath agent is never marked down for not being in London. Missing parts are left out and the rest rescaled, and every score shows how much of the data is in. A provisional score shows at 50% coverage; a branch enters the public ranking at 75%.
Sale price as a share of first asking price, median days from listing to sold subject to contract, share of listings that sell, and share cut in price. Built by matching each listing on the agent's own website to its HM Land Registry sale.
Google, Trustpilot and ESTAS ratings blended with a Bayesian average (prior 4.3 stars worth 30 reviews). Branch reviews count in full; brand-wide reviews count at 10%, capped at 100. Reviews from the last 24 months weigh double.
Share of local stock and new instructions in the last 14 days, square-root scaled so the biggest branch does not run away with it.
Redress scheme, Propertymark, client money protection, Companies House status, and whether the agent has claimed and kept its profile up to date.
How listings are collected
A crawler visits each agent's own website once a day, respecting robots.txt and waiting between requests. It recognises the common agency website platforms, reads structured data where the site provides it, falls back to the site map and listing cards, and records price, status, address, bedrooms and property type. Day-to-day changes give new instructions, price reductions and sales agreed. Each month, completed sales from HM Land Registry are matched back to the listings to measure price achieved and time to sell.