1280-UC Predicting Marketing Campaign Acceptance and Identifying Customer Segments

Location

https://campus.kennesaw.edu/colleges-departments/ccse/events/computing-showcase/fa26-cday-program.php

Document Type

Event

Start Date

20-11-2026 2:00 PM

Description

This study examined whether customer demographics, purchasing behavior, and prior campaign responses could help predict who would accept a marketing offer. We analyzed 2,237 cleaned records from Kaggle’s Customer Personality Analysis dataset, in which 14.9% of customers accepted the most recent offer. We compared logistic regression, decision tree, random forest, and gradient boosting classifiers using a stratified 80/20 train–test split and five-fold cross-validation. We selected each model’s decision threshold from out-of-fold training predictions before evaluating it on the test set. Random forest achieved the highest test F1 score (0.5874) and precision (55.26%), while logistic regression achieved the highest mean training cross-validation F1 (0.5756) and test recall (68.66%). A separate K-means analysis identified two customer segments. The higher-spending segment had an observed acceptance rate of 21.76%, compared with 10.05% in the lower-spending segment. These findings suggest that response scores could help marketing teams prioritize customers when outreach capacity is limited, while customer segments could guide controlled tests of tailored offers.

This document is currently not available here.

Share

COinS
 
Nov 20th, 2:00 PM

1280-UC Predicting Marketing Campaign Acceptance and Identifying Customer Segments

https://campus.kennesaw.edu/colleges-departments/ccse/events/computing-showcase/fa26-cday-program.php

This study examined whether customer demographics, purchasing behavior, and prior campaign responses could help predict who would accept a marketing offer. We analyzed 2,237 cleaned records from Kaggle’s Customer Personality Analysis dataset, in which 14.9% of customers accepted the most recent offer. We compared logistic regression, decision tree, random forest, and gradient boosting classifiers using a stratified 80/20 train–test split and five-fold cross-validation. We selected each model’s decision threshold from out-of-fold training predictions before evaluating it on the test set. Random forest achieved the highest test F1 score (0.5874) and precision (55.26%), while logistic regression achieved the highest mean training cross-validation F1 (0.5756) and test recall (68.66%). A separate K-means analysis identified two customer segments. The higher-spending segment had an observed acceptance rate of 21.76%, compared with 10.05% in the lower-spending segment. These findings suggest that response scores could help marketing teams prioritize customers when outreach capacity is limited, while customer segments could guide controlled tests of tailored offers.