Kies de Nederlandse taal
Course module: JM0190-M-6
Entrepreneurial Marketing
Course info
Course moduleJM0190-M-6
Credits (ECTS)6
CategoryMA (Master)
Course typeCourse
Language of instructionEnglish
Offered byTilburg University; Tilburg School of Economics and Management; TiSEM: Marketing; TiSEM: Marketing;
Is part of
M Data Science and Entrepreneurship (joint degree)
prof. dr. M.C. Kaptein, PDEng
Other course modules lecturer
P. Provodin
Other course modules lecturer
Z. Puha
Other course modules lecturer
Academic year2020
Starting block
SM 2
Course mode
Registration openfrom 19/01/2021 up to and including 20/08/2021

Entrepreneurial Marketing: modern quantitative methods in marketing.
General course description
In this course we will introduce novel data science methods as these are discussed and used in the modern (quantitative) marketing literature. Throughout the course, we will adopt a “marketing as interventions” approach: i.e., we will think about successful marketing campaigns as selecting “the right message for the right customer and the right time” and explore sequential learning methods to address this problem.

The course consists of three “blocks”; the first two blocks are more theoretical in nature, while the third block is applied. The first two theoretical blocks will consist of lectures introducing a new topic and subsequently the plenary discussion of 1 or 2 contemporary research papers related to that weeks topic. Hence, the course will have a very strong focus on the modern quantitative data science literature as it is emerging within marketing (and, admittedly, some of the papers will be challenging to read; we will discuss them together to try to understand them well). The applied block will consist of a (group) assignment in which the learned materials are applied.

In the first block we will introduce our main approach, and we will discuss the contextual Multi-armed bandit problem (and it’s potential solutions) as a way of formalizing the “marketing as interventions” approach. We will cover the basic problem setup, various (applied) strategies to address the problem, and a number of practical tools that allow students to experiment with bandit policies. Note that during this block will not require an in depth knowledge of the mathematical proofs: rather, we will focus on the main ideas presented in the papers that we will discuss.

In the second block we will cover several related topics: first, we will discuss causal inference from both Rubin’s and Pearl’s points off view. Next, we will discuss churn minimization and uplift modeling. Finally, we will close off with a discussion of active learning and its use in (batch mode) uplift modeling.

The third block consists of an applied project that is provided by a company called Exsell (; Exsell offers a platform that allows their clients to optimize online content. We will, in groups, analyze their product and provide suggestions for improvement: we will think about how to extend their product to include the “marketing as interventions” approach that we have covered in the theoretical blocks. Effectively, you will be creating a prototype for the second version of their product that should incorporate the lessons learned in the theoretical blocks and guide Exsell towards developing a platform for online, sequential, selection of marketing interventions.

After passing this course, the students are able to:
  1. Frame sequential learning problems in marketing as (contextual) Multi-Armed Bandit problems.
  2. Explain the exploration-exploitation tradeoff that arises in sequential learning.
  3. Implement various contextual bandit policies such as UCB and Thomson Sampling and discuss their applications in marketing.
  4. Explain the difference between causality and correlation and discuss various approaches (Rubin’s Causal Model & SEMs) for modeling causal relationships.
  5. Understand active learning and describe the pro’s and con’s of various approaches to active learning in a marketing setting.
  6. Implement DS / ML models to model customer churn.
This course will be taught fully online. During the first two blocks we will discuss the assigned papers in online, interactive, sessions. Note that you are expected to contribute actively to the discussions: we expect an active involvement of all the students who select this elective. The final, applied, block will also take place online; during this block you will meet once a week with your supervisor to discuss your progress.

Block 1 will start on February 3rd, with the midterm on March 10th.
Block 2 will start March 17th, with the midterm on April 14th.
Block 3 (the applied assignment) will start April 14th, with final presentations on May 13th.

After each block of theory you will take a midterm exam. The midterm exams are “open book” and will consist of 5 essay questions (where each answer will be limited to 250 words). These “mini-essay mid-term exams” will be graded on an individual basis and the average grade you receive for the two midterm tests will constitute 60% of your final grade.

The final assignment, which you will carry out in groups of 3-4 students, will be graded during the final presentation. The final assignment grade will constitute 40% of your final grade.

Every week we will be covering one or two recent research papers; you are expected to read these scientific papers before you arrive to the class. Below you will find the articles you will need to read for the first three weeks of the course (on Canvas we will make sure to announce the papers for the subsequent weeks at least two weeks before the lecture).

Week 1:
1. Kaptein, M., McFarland, R., & Parvinen, P. (2018). Automated adaptive selling. European Journal of Marketing

Week 2:
1. Chapelle, O., & Li, L. (2011). An empirical evaluation of thompson sampling. In Advances in neural information processing systems (pp. 2249-2257). 
2. Eckles, D., & Kaptein, M. (2019). Bootstrap thompson sampling and sequential decision problems in the behavioral sciences. Sage Open, 9(2), 2158244019851675. 

Week 3:
1. Li, L., Chu, W., Langford, J., & Schapire, R. E. (2010, April). A contextual-bandit approach to personalized news article recommendation. In Proceedings of the 19th international conference on World wide web (pp. 661-670). 
2. Li, L., Chu, W., Langford, J., Moon, T., & Wang, X. (2012, May). An unbiased offline evaluation of contextual bandit algorithms with generalized linear models. In Proceedings of the Workshop on On-line Trading of Exploration and Exploitation 2 (pp. 19-36). 

Week 4:
TBA (the remainder of the literature will be announced on Canvas).

Contact person
Z. Puha
Timetable information
Entrepreneurial Marketing
Required materials
Recommended materials
Final grade

Kies de Nederlandse taal