Incremental Response Modeling Based on Segmentation Approach Using Uplift Decision Trees
Data mining methods have been successfully used in direct marketing to model the behavior of responders. But these response models do not take in account, the behavior of customers who would take an action irrespective of marketing action. Redundant marke
- PDF / 430,914 Bytes
- 10 Pages / 439.37 x 666.142 pts Page_size
- 68 Downloads / 166 Views
2
Specialist, Fidelity Investments, Bangalore, India [email protected] M.SC, Indian Institute of Technology Kanpur, Kanpur, India [email protected] 3 Lead, Fidelity Investments, Bangalore, India [email protected]
Abstract. Data mining methods have been successfully used in direct marketing to model the behavior of responders. But these response models do not take in account, the behavior of customers who would take an action irrespective of marketing action. Redundant marketing communications sometimes annoy the customer and reduce the brand value of the company. Accurate targeting of customers also reduces direct marketing cost. Incremental response modeling aims to predict the behavior of customers who respond positively only in the case of marketing. In this paper, we propose a two-step approach for incremental response modeling. In the first step, we segment the data using uplift decision trees using traditional and modified divergence metrics. Then, in the second step we use the standard incremental response modeling methods. Experiments on real world direct marketing campaign data showed that the proposed method outperforms the uplift decision trees. Keywords: True-lift modeling
Net lift Uplift decision trees
1 Introduction Direct marketing cost is one of the biggest contributors of overall marketing cost. Most companies use data mining tools to develop predictive models to identify customers who are likely to respond to direct marketing campaigns. But improper use of these models may lead to unexpectedly low response rates. This is mainly because many of the customers get annoyed due to repetitive phone calls, emails, etc. Companies thus want to target their customers precisely to retain the brand value and improve customer satisfaction. Traditionally, response models try to predict the likelihood of response for a particular customer, given a marketing action. These models do not account for likelihood of response irrespective of marketing action. Incremental response modeling as shown in Fig. 1 (also known as uplift modeling or differential response modeling or true uplift modeling) removes this drawback and works towards choosing the customers © Springer International Publishing Switzerland 2016 P. Perner (Ed.): ICDM 2016, LNAI 9728, pp. 54–63, 2016. DOI: 10.1007/978-3-319-41561-1_5
Incremental Response Modeling
55
effectively. Incremental response models are used to identify the customers where there is a true uplift [5]. We want to identify customers who respond to the direct marketing campaign positively and remove the customers from the targeting list who will take an action irrespective of marketing or not.
Respond if targeted Yes
No
Yes
Spontaneous response (Targeting not required here)
Don’t target at all (customers getting annoyed, leading to negative response)
No
Customers to be targeted
No effect of targeting (only increasing the cost to the company)
Respond if not targeted
Fig. 1. True response uplift.
In true uplift modeling, historical treatment
Data Loading...
