Six Names for the Same Crowd

Cluster Stability and Incremental Predictive Value in a Six-Segment Open-Home Buyer Model Across 48,206 Registrations at 27 Sales Offices

A research poster from the School of Continuous Improvement re-derives the six-segment buyer model an estate agency group assigns from open-home sign-in registrations, using the group's own k-prototypes specification on 48,206 registrations across 3,984 inspections and 27 sales offices over fourteen months. Median bootstrap cluster stability is 0.40 in the derivation sample and 0.31 on held-out registrations, with five of the six segments below the 0.5 mark at which a cluster is conventionally treated as dissolved, and 46.8 per cent of registrants change segment between consecutive quarterly refreshes (adjusted Rand index 0.19). Entered alongside a count of open homes attended in the prior ninety days, segment membership moves discrimination from 0.673 to 0.681, a difference whose confidence interval spans zero, while the group's allocation-coverage indicator records 99.6 per cent of registrations as segment-allocated throughout.

@misc{slop_yrnz28,
  author       = {Sten Okwuosa and Mirela Hanke and Renke Sabel},
  title        = {Six Names for the Same Crowd: Cluster Stability and Incremental Predictive Value in a Six-Segment Open-Home Buyer Model Across 48,206 Registrations at 27 Sales Offices},
  year         = {2026},
  publisher    = {Slop University},
  doi          = {10.5555/slop.yrnz28},
  url          = {https://slop.university/outputs/slop-poster-open-home-sign-in-yrnz28/},
  version      = {1.0},
  note         = {Research poster},
}