Marketing channel allocation
23 Mar 2023
Overview
Real Estate asset marketing is performed by Asset Managers with little granularity due to high volumes and fast pace of management
This leads to increased costs as decisions cannot be drilled down to the particular cases. The answer to “which channel is best?” for an asset is usually “all channels”
Following this strategy sets a performance baseline, but costs are way above what it is really needed to maintain that performance
The Challenge
The customer was spending +1.8M€ in asset publishing, as the strategy was to publish all assets in all Real Estate portals
The goal was to maintain lead generation performance while reducing publishing costs
Real Estate asses are sold on a 1:1 ratio. Generating 1,000 leads for an apartment does not imply the potential of selling 1,000 apartments, as assets differ from one another
The Approach
Liquidity model: ML model to anticipate, for an asset with certain features (including price), the amount of leads that it could generate in the market
Placement model: ML model to predict which would be the Real Estate portal that would bring the lead that acquires the asset
Saturation pace model: model that identifies the pace at which each asset saturates lead generation. Lead generation follows a log curve, where 80% of leads are generated in the first X days
Cascading algorithm: algorithm that generates a publishing cascade based on the combined output of the previously mentioned models, with the goal of maintaining asset sales while reducing publishing costs
The Results
Cost reduction of 300.000€ for the assets under consideration (-16% on last year’s budget)
Cascade implemented for asset inflow
Marketing channel allocation
23 Mar 2023
Overview
Real Estate asset marketing is performed by Asset Managers with little granularity due to high volumes and fast pace of management
This leads to increased costs as decisions cannot be drilled down to the particular cases. The answer to “which channel is best?” for an asset is usually “all channels”
Following this strategy sets a performance baseline, but costs are way above what it is really needed to maintain that performance
The Challenge
The customer was spending +1.8M€ in asset publishing, as the strategy was to publish all assets in all Real Estate portals
The goal was to maintain lead generation performance while reducing publishing costs
Real Estate asses are sold on a 1:1 ratio. Generating 1,000 leads for an apartment does not imply the potential of selling 1,000 apartments, as assets differ from one another
The Approach
Liquidity model: ML model to anticipate, for an asset with certain features (including price), the amount of leads that it could generate in the market
Placement model: ML model to predict which would be the Real Estate portal that would bring the lead that acquires the asset
Saturation pace model: model that identifies the pace at which each asset saturates lead generation. Lead generation follows a log curve, where 80% of leads are generated in the first X days
Cascading algorithm: algorithm that generates a publishing cascade based on the combined output of the previously mentioned models, with the goal of maintaining asset sales while reducing publishing costs
The Results
Cost reduction of 300.000€ for the assets under consideration (-16% on last year’s budget)
Cascade implemented for asset inflow
Marketing channel allocation
23 Mar 2023
Overview
Real Estate asset marketing is performed by Asset Managers with little granularity due to high volumes and fast pace of management
This leads to increased costs as decisions cannot be drilled down to the particular cases. The answer to “which channel is best?” for an asset is usually “all channels”
Following this strategy sets a performance baseline, but costs are way above what it is really needed to maintain that performance
The Challenge
The customer was spending +1.8M€ in asset publishing, as the strategy was to publish all assets in all Real Estate portals
The goal was to maintain lead generation performance while reducing publishing costs
Real Estate asses are sold on a 1:1 ratio. Generating 1,000 leads for an apartment does not imply the potential of selling 1,000 apartments, as assets differ from one another
The Approach
Liquidity model: ML model to anticipate, for an asset with certain features (including price), the amount of leads that it could generate in the market
Placement model: ML model to predict which would be the Real Estate portal that would bring the lead that acquires the asset
Saturation pace model: model that identifies the pace at which each asset saturates lead generation. Lead generation follows a log curve, where 80% of leads are generated in the first X days
Cascading algorithm: algorithm that generates a publishing cascade based on the combined output of the previously mentioned models, with the goal of maintaining asset sales while reducing publishing costs
The Results
Cost reduction of 300.000€ for the assets under consideration (-16% on last year’s budget)
Cascade implemented for asset inflow