PropertyMetrics / residential evidence deskMethodData sources
PropertyMetrics / Build a local comparable pool without confusing distance with similarity

Build a local comparable pool without confusing distance with similarity

How to widen a sparse postcode search, preserve distance labels and remove nearby transactions that do not answer the subject-property question.

Published by PropertyMetrics Research DeskLast reviewed: 13 August 2026

2 minute read. Practical research method for use with public property records.

Start with a candidate pool, not a valuation set

An exact postcode can contain only a handful of completed transactions, particularly for one property type. A nearby-postcode search can make the evidence review more useful, but it changes the question from ‘what sold in this postcode?’ to ‘which nearby sales might deserve inspection?’

PropertyMetrics keeps the subject postcode, the nearby postcode and centroid distance on every candidate. That distance is a discovery aid. It does not measure street character, school catchments, station access, noise, outlook or the route between properties.

Remove legal and physical mismatches first

Filter tenure and broad property type before looking at price. A leasehold flat and a freehold terrace can be geographically close while answering different market questions. Next compare floor area, bedroom count, plot, parking, condition, lease terms and any material extension.

The public Price Paid feed does not contain those attributes. Add them from particulars, EPC records, surveys or direct inspection, and record the origin rather than filling gaps from memory.

  • Keep the exact-postcode set available as a separate reference.
  • Prefer a smaller explainable set to a larger automatic average.
  • Do not use centroid distance as an adjustment formula.
  • Record why every selected transaction remains useful.

Treat time as another difference to explain

A transaction several years old can be physically similar but needs a time-period discussion. A recent sale can be physically irrelevant. Avoid choosing on date alone and avoid applying a general house-price index mechanically to one property without explaining the chosen index and geography.

When the candidate set crosses a changing market period, show both recorded and adjusted figures. Keep the recorded completion price immutable.

Stop when the evidence remains heterogeneous

If every candidate needs a large or poorly evidenced adjustment, the pool has not become stronger merely because it is larger. Document the weak match and seek better particulars, a wider but carefully controlled search, or professional valuation evidence.

Primary references