Business Objective
Our client, one of the world’s largest logistics Real Estate Investment Trust wanted to understand the factors that affect real estate prices and attain an optimal pricing point for each property.
The client wanted us to identify and quantify key pricing drivers for each individual market and create a price elasticity model to estimate the likelihood of conversion at different price points. They also wanted the model to be capable of providing recommendations on target pricing to maximize revenue.
Challenges
- Limited data due to long term leases with limited properties
- Missing/incorrect data for multiple acquired properties
- Prices and drivers varied a lot between markets/sub-markets and were impacted by wide variations in properties size, condition, and features
Solution Methodology
- Cleaned and collated internal and external data sources encompassing lease, negotiation, property, customer attributes, spatial and macroeconomic data among others.
- Created a 3-tiered modeling approach incorporating linear regression, logistic regression, and Bayesian models to identify price drivers, predict win probability for each given price point, and adjusting the price curve based on market/competitor data.
- Fed the output of the above optimal price prediction engine directly into the revenue maximization model and optimized using NLOPTR (Non-Linear Optimization) to arrive at the optimal win probability.
Business Impact
- The Pricing Optimization model showed a lift of 3% while the Revenue Maximizer model results showed a lift of 4% over the baseline data
- R Shiny based web tool helped the leasing team to curate competitor comparable properties and adjust price elasticity curves accordingly
- The scenario planner enabled the client to find an appropriate balance between pricing and occupancy to maximize rent & revenue. The model predicted the required changes in Optimal Rental Price to improve Win Probability after factoring in the impacts of Covid-19
