Link to the University of Pittsburgh Homepage
Link to the University Library System Homepage Link to the Contact Us Form

Business Analytics for Non-profit Marketing and Online Advertising

Chang, Wei (2013) Business Analytics for Non-profit Marketing and Online Advertising. Doctoral Dissertation, University of Pittsburgh. (Unpublished)

Primary Text

Download (1MB) | Preview


Business analytics is facing formidable challenges in the Internet era. Data collected from business website often contain hundreds of millions of records; the goal of analysis frequently involves predicting rare events; and substantial noise in the form of errors or unstructured text cannot be interpreted automatically. It is thus necessary to identify pertinent techniques or new method to tackle these difficulties. Learning–to-rank, an emerging approach in information retrieval research has attracted our attention for its superiority in handling noisy data with rare events. In this dissertation, we introduce this technique to the marketing science community, apply it to predict customers’ responses to donation solicitations by the American Red Cross, and show that it outperforms traditional regression methods. We adapt the original learning-to-rank algorithm to better serve the needs of business applications relevant to such solicitations. The proposed algorithm is effective and efficient is predicting potential donors. Namely, through the adapted learning-to-rank algorithm, we are able to identify the most important 20% of potential donors, who would provide 80% of the actual donations.
The latter half of the dissertation is dedicated to the application of business analytics to online advertising. The goal is to model visitors’ click-through probability on advertising video clips at a hedonic video website. We build a hierarchical linear model with latent variables and show its superiority in comparison to two other benchmark models. This research helps online business managers derive insights into the site visitors’ characteristics that affect their click-through propensity, and recommends managerial actions to increase advertising effectiveness.


Social Networking:
Share |


Item Type: University of Pittsburgh ETD
Status: Unpublished
CreatorsEmailPitt UsernameORCID
Chang, Weiwec19@pitt.eduWEC19
ETD Committee:
TitleMemberEmail AddressPitt UsernameORCID
Committee ChairShang, Jennifershang@katz.pitt.eduSHANG
Committee MemberVenkatesh, Vrvenkat@katz.pitt.eduRVENKAT
Committee MemberMontgomery,
Committee MemberRajgopal , Jayantrajgopal@pitt.eduRAJGOPAL
Committee MemberHegde, G.G.hegde@katz.pitt.eduHEGDE
Date: 2 July 2013
Date Type: Publication
Defense Date: 3 December 2012
Approval Date: 2 July 2013
Submission Date: 22 May 2013
Access Restriction: No restriction; Release the ETD for access worldwide immediately.
Number of Pages: 99
Institution: University of Pittsburgh
Schools and Programs: Joseph M. Katz Graduate School of Business > Business Administration
Degree: PhD - Doctor of Philosophy
Thesis Type: Doctoral Dissertation
Refereed: Yes
Uncontrolled Keywords: data mining, Bayesian statistics, direct marketing, business analytics, learning to rank, customer relationship management
Date Deposited: 02 Jul 2013 13:48
Last Modified: 15 Nov 2016 14:12


Monthly Views for the past 3 years

Plum Analytics

Actions (login required)

View Item View Item