Propensity To Buy (P2B) Model: How to Target the Audience Most Likely to Buy
Is this case study relevant to you?
Explore how our solutions can implement to your business
What problems does the business want to solve?
A retailer selling high-end accessories wants to implement differential marketing campaigns, for example, ads based on visitor behavior. The goal is to target the audiences most likely to buy.
A Propensity To Buy (P2B) Model is a practice of predictive modeling that tries to find the probability of each customer or business will or will not buy a certain product over a period of time in the future.
For every model that the FocusKPI team has built, we test more than 10,000 variables. Types of variables include:
-
Time spectra (weekly, monthly, quarterly) of bid, bin, successful bid/bin, gmb, revenue
-
Time spectra of on-site activity such as visit duration time, number of total page, view item page, watch page, bid page, bin page, search page, sessions
-
Time spectra of category in which user purchased
-
User’s profile (like age, income, etc) or lifetime information (like recency – days since last bid/bin, successful bid/bin, tenure – months since registration)
-
Trend variables (velocity variables such as bid, bin by category)
-
Time interval (averaged number of days between bids or bins, and corresponding RMS)
Using our proprietary modeling tool Focus Score, we can rank variable contribution and identify top predictors out of 10,000 variables for each model.
As seen above, our team has identified 10 variables for a P2B model with the order of contribution, which enables the company to identify potential buyers and to target them in the next promotional campaigns.
Slide right to view all tabs
What problems does the business want to solve?
A retailer selling high-end accessories wants to implement differential marketing campaigns, for example, ads based on visitor behavior. The goal is to target the audiences most likely to buy.
A Propensity To Buy (P2B) Model is a practice of predictive modeling that tries to find the probability of each customer or business will or will not buy a certain product over a period of time in the future.
For every model that the FocusKPI team has built, we test more than 10,000 variables. Types of variables include:
-
Time spectra (weekly, monthly, quarterly) of bid, bin, successful bid/bin, gmb, revenue
-
Time spectra of on-site activity such as visit duration time, number of total page, view item page, watch page, bid page, bin page, search page, sessions
-
Time spectra of category in which user purchased
-
User’s profile (like age, income, etc) or lifetime information (like recency – days since last bid/bin, successful bid/bin, tenure – months since registration)
-
Trend variables (velocity variables such as bid, bin by category)
-
Time interval (averaged number of days between bids or bins, and corresponding RMS)
Using our proprietary modeling tool Focus Score, we can rank variable contribution and identify top predictors out of 10,000 variables for each model.
As seen above, our team has identified 10 variables for a P2B model with the order of contribution, which enables the company to identify potential buyers and to target them in the next promotional campaigns.