Using Purchase History to Identify Customer “Projects”
Whether it is gathering ingredients for a special meal or assembling the tools and supplies needed for a craft project, customers frequently purchase a collection of products that they need to complete a specific project. One might expect that when a consumer is in the midst of such a project, she will be more open to product suggestions that might help shape her plans and achieve her goal in a satisfactory manner. Yet today’s marketers have few tools to help identify collections of products that are associated with projects, or customers who seem to be engaged in such an activity. Behavioral customer segmentations are typically static and basket analysis seldom straddles multiple purchase occasions that might be associated with the same project.
Analytics at Wharton is pleased to announce a rich data set from a Fortune 500 Specialty Retailer that will allow researchers to study this problem. The data set contains 60,000 customers along with all the individual items each customer purchased over a 24-month window. In addition to purchases, the dataset includes detailed product hierarchies and product attributes, store/location information, and email campaigns by the firm to the households. While many researchers could provide basic segmentation strategies, the corporate sponsor seeks novel approaches to segmentation which recognize customer needs change over time, people consume bundles of products for “projects,” and that this can be identified by the items in each customer’s shopping cart.
In addition to the primary research question, the sponsor is open to proposals on other avenues of research utilizing this detailed purchase and direct marketing history. These avenues could include: product recommendations, product cross-sell/promotion analysis, geographic purchasing habits, or other analysis.
Note: This Research Opportunity remains open for proposal submissions. Interested researchers can submit proposals online through the Archived Proposal Submission Portal. Researchers are encouraged to review proposal submission guidelines before submitting their proposal. Additional questions can be directed to ai-analytics.wharton.upenn.edu.
Research Teams
GRANTEES OF THE IDENTIFYING CUSTOMER PROJECTS DATA:
MEASURING THE EFFECTIVENESS OF LOYALTY PROGRAMS USING COLLABORATIVE FILTERING
Wayne Taylor, UCLA
WHAT COME NEXT? SEQUENCE-BASED PROJECT AND PRODUCT RECOMMENDATIONS
Thomas Lee, UC Berkeley
Sayantan Mukhopadhyay, UC Berkeley
Kristine Yoshihara, UC Berkeley
MODEL-BASED PROJECT DISCOVERY
Bruno Jacobs, Erasmus Universiteit Rotterdam
Dennis Fok, Erasmus Universiteit Rotterdam
Bas Donkers, Erasmus Universiteit Rotterdam
DISCOVERING PURCHASE PATTERNS OF CUSTOMERS VIA A HIERARCHICAL THEME DICTIONARY MODEL
Jun Liu, Harvard University
Ke Deng, Tsinghua University
Kate Li, Suffolk University
Dmitry Zinoviev, Suffolk University
Zhen Zhu, Suffolk University
USING PURCHASE HISTORY TO IDENTIFY AND RECOMMEND CUSTOMER “PROJECTS”: A MODULAR APPROACH
Panagiotis Sarantopoulos, Athens University of Economics and Business
Katerina Pramatari, Athens University of Economics and Business
Dimitris Papakiriakopoulos, Technological Educational Institute of Athens
Panos Markopoulos, University of Cyprus
UNDERSTANDING WHAT YOUR CUSTOMERS DO WITH WHAT THEY PURCHASE: A STRUCTURAL APPROACH
Ludovic Stourm, University of Pennsylvania
Eric Bradlow, University of Pennsylvania
Raghuram Iyengar, University of Pennsylvania
UNCOVERING GOAL STRUCTURE FROM CONSUMER PURCHASE HISTORIES
Gary Russell, University of Iowa
Wagner Kamakura, Rice University
USING A FLEXIBLE CROSS-CATEGORY CONSIDERATION MODEL TO IDENTIFY UNOBSERVED PROJECTS: COMMERCIAL CUSTOMERS VS END-USE CUSTOMERS
Matthew Osborne, University of Toronto
Andrew Ching, University of Toronto
IDENTIFYING CUSTOMER-GENERATED PROJECTS FROM TRANSACTIONS DATA OVER TIME
Suzanne Walchli, University of the Pacific
Gerald Post, University of the Pacific
Disclaimer
The Research Opportunity webinars are posted here for a research-focused audience and should not be quoted, paraphrased or otherwise utilized without written permission of Analytics at Wharton. Any media inquiries or requests for quotes about the projects should be directed to analytics@wharton.upenn.edu.