Built for Planned Furniture Promotions: a research pipeline that turns a raw industry mailing list into a living database of retailers matching PFP's exact criteria — 50+ years in business, 10,000+ sq ft, still operating — with a cited source behind every claim.
Each stage is cheaper than the next and shrinks the work the next stage must do — so expensive research only ever runs on records that already passed. Every fact is cached and timestamped, making the list refreshable instead of disposable.
9,012 raw rows → 7,641 unique retailers. Stable IDs keyed to street address + phone, never to name — same-name collisions are the #1 error in industry lists.
Website, business listings, and social activity confirm each store still operates. Conflicting signals are flagged, never silently dropped.
The main filter. Reads each store's own history page, news coverage, and registries — capturing true continuous operating age, not just the "Since 19##" on the homepage.
A second, independent research pass tries to refute every positive finding before it reaches the list. Caught a false claim in the pilot — see Data Quality below.
Showroom size from the store's own claims, cross-checked against open building-footprint data — a defensible estimate for nearly every U.S. address before buying any data.
Brands carried, buying groups, generational ownership, succession signals — rolled into dual rankings tuned to PFP's offer.
A stratified sample spanning 42 states, run end-to-end through Stages 1–2B in July 2026. Hover any element for detail.
Every row is a real, currently-operating U.S. furniture retailer verified to have been in continuous operation for 50+ years. Click any row to see the evidence behind the claim — the quoted source, and the liveness proof tied to the store's own address and phone. The full run projects ~1,410 of these.
| Retailer | City | St | Est. | Years | Confidence | Source |
|---|
A store that's been family-run for half a century is either a rock-solid, well-capitalized operator — or an owner quietly approaching retirement. Those are opposite sales conversations. So every qualified lead gets scored through both lenses, producing two ranked call lists from the same underlying data. Draft weights below; final tuning happens against PFP's own best-client history.
Each score runs 0–100 and is multiplied by a data-confidence factor, so a thin record can never outrank a well-documented one — leads with insufficient evidence go to a research queue instead of quietly sinking to the bottom of a list. Signals are collected during enrichment (Stage 4) from pages the pipeline is already reading, so scoring adds almost no marginal cost.
The pilot proves the method. Each phase compounds on the cached research before it — nothing is ever re-bought or re-researched.