Know Which Tenants Will Stay
Explainable renewal-probability and early-warning risk scores for every tenancy across a Hong Kong retail portfolio.
- ThinkCol built custom models that score every tenancy on renewal probability and early-warning financial and reputational risk.
- Risk scores are distilled from financial-news sentiment and Cantonese social listening across 120+ local sources, scoring tenants on financial health, reputation, perception, and customer satisfaction.
- In validation, 82% of forecasted store sales were accurate within a 20% margin, and 70% of renewal outcomes were predicted correctly.
An asset management team needed to assess tenant risk across a full Hong Kong retail portfolio. Non-renewal notices landed too late to backfill space, leaving units empty for six to twelve months, and negotiations often launched without a view of projected sales or risk.
ThinkCol built custom models that score every tenancy: a renewal probability model trained on the portfolio's own lease, rent, and sales history, plus an early-warning risk score distilled from financial-news sentiment and Cantonese social listening across more than 120 local sources — scoring tenants on financial health, reputation, perception, and customer satisfaction, each with the reasons behind it.
Scores land as a ranked, explainable watchlist so asset managers can focus on the highest-value, highest-risk tenancies first. In validation, 82% of forecasted store sales were accurate within a 20% margin, and 70% of renewal outcomes were predicted correctly.
Why customise? The models read the portfolio's own renewal history and Hong Kong-specific tenant signals, not a generic credit score.
