Predicting renewal rates to help their leasing decision

September 10, 2020

Leading US REIT

Problem Statement

• Company had multiple nonintegrated data sources that were independently used for decision making by different teams; however, decisions were not based on an integrated data view
• Vacancy costs resulted due to the time lag between a resident moving out and a new resident moving in
• Potential gaps in the data that when filled could lead to better decision making

Analytics Led Approach

• Unified view of multiple data sources
• Correlation Analysis Information Value Analysis
• Sub Market Clustering, Logistic Regression Model
• Key Driver Impact Analysis, Amenity Impact Analysis
• Integrated data view, Mathematical Model
• Analysis Results

Business Impact

• Accurate predictions: Accuracy level was 80 – 85%
• Reduced vacancy cost: Clients had 3 months time to extend offers to profitable customers who were predicted to churn or find new customers
• Better understanding of key drivers: The impact of each key driver on the renewal was quantified

Critical Success Factors

• Increased model relevance: Separate models were developed for properties in each sub market group (similarly behaving sub market groups were identified through clustering)
• Better understanding of how amenities affected rent and helped the client revamp the decisions making process for pricing decisions with predicted propensities