Case Article–Forecasting at FoodMart: A Case on Cultural Diversity in Supply Chain Management

1. Introduction

Diversity, equity, and inclusion (DEI) is a much-discussed topic in management, both in business schools and in the broader working world. The latest accreditation guidelines from the Association to Advance Collegiate Schools of Business specifically list diversity and inclusion as a guiding principle, and they further integrate diversity and inclusion into three of the nine accreditation standards (AACSB 2020). Many authors have argued that there is a business case for diversity: the idea that diverse teams improve firm performance (Robinson and Dechant 1997, Carter et al. 2007, Campbell and Mínguez-Vera 2008). In addition, authors such as Bell et al. (2009) claim that there is a moral and ethical imperative for diversity education in business schools.

However, there is no clear consensus in the research on what techniques actually work for “teaching” DEI topics, especially in a manner that leads to concrete change in the business world (Anand and Winters 2008). Kulik and Roberson (2008a) present evidence that most existing methods do work to increase knowledge of DEI but little to change attitudes or behaviors, although in Kulik and Roberson (2008b), these same authors note that the very existence of DEI training communicates organizational values to the learners. More pessimistically, Emerson (2017) and Noon (2018) note that members of the dominant group often react defensively to DEI training focused on recognizing unconscious biases, rendering it worse than ineffective. There is, therefore, an argument to be made that DEI education should be more subtly integrated into education and training (Devine et al. 2012, Tropp and Godsil 2015, Madva 2017), and it is that concept that motivates the creation of this case.

Unfortunately, teaching materials and similar resources for bringing DEI topics into the supply chain management classroom remain quite sparse. David and Coates (2021) survey several of the most popular case publishers, including Harvard Business Publishing, INSEAD, and Ivey Publishing, finding that cultural and demographic differences are largely absent from the most popular teaching cases across all disciplines, including in supply chain management. Searching two major journals that support teaching in the field of supply chain management, INFORMS Transactions on Education and Decision Sciences Journal of Innovative Education, turns up scarce examples where DEI topics are present in the cases, regardless of whether they comprise a primary learning goal. Those that have been published in recent years incorporate DEI topics in the context of artificial intelligence (AI). Chen et al. (2022) presents a semester-long project aimed at using AI to solve various social issues. A case by Pachamanova et al. (2022) incorporates discussion of how machine learning algorithms have the potential to recreate human biases when dealing with sociodemographic data. As compared with these examples, this case addresses a very different area of supply chain management (forecasting), and it is, therefore, a valuable contribution to the available teaching materials for supply chain management instructors.

It is widely agreed that forecasting is an important topic in business school education (Hanke and Weigand 1994, Loomis and Cox 2003, Beal and Wilson 2015, Kros and Rowe 2016), and many examples of activities for teaching forecasting can be found in management and economics teaching journals. Although most focus on specific analytical methods and tools (Chu 2007, Kros and Nadler 2008, Torres et al. 2018, Brusco 2022, Diamant 2024), those such as Gel et al. (2014) and Gavirneni (2008) focus more on the value of having students apply qualitative thinking and subjective judgement to data sets. Although Leiter (2023) is from the political science discipline, the author specifically focuses on teaching students to forecast without teaching specific models and finds it effective in helping students understand the underlying principles that drive the predicted behavior. This case falls into this latter category, explicitly asking students to investigate why a standard seasonal model does not apply to the historical sales data. Further, the case includes a “mystery item,” the identity of which students try to guess, and thereby, it introduces an element of gamification, a feature other authors have found to drive student interest in forecasting (Snider and Eliasson 2013, Legaki and Assimakopoulos 2018). This case, therefore, provides instructors in the supply chain management discipline a starting point for discussing a DEI topic in the classroom, specifically the value of cultural diversity in the workplace, while simultaneously engaging them in learning about forecasting.

2. Forecasting at FoodMart

The forecasting at FoodMart case asks teams to forecast products based on sales history that is influenced by various cultural celebrations. Because the celebrations do not necessarily occur in the same week each year according to a standard Gregorian calendar, students need to investigate the sales history and try to understand the underlying reasons for the demand patterns in past years. Students are given five years of sales history data, broken down into weekly buckets, for four products. They are also given a statistical forecast said to be generated by the forecasting software. This statistical forecast is a straight calculation of seasonality for each product (i.e., for each week, the five years of history are averaged). Three of the products are identified in the spreadsheet, and the fourth is a “mystery” product that students are asked to identify.

2.1. Learning Goals

After completing this activity, students should be able to

• identify irregular patterns in historical sales data and determine their root causes;

• evaluate a statistical forecast created without human intervention; and

• explain how culturally diverse teams can be valuable in a forecasting process.

3. Classroom Experience

This game was used twice in an introductory undergraduate course in supply chain management geared primarily toward freshmen and sophomores. The first class consisted of 51 students, 49 of whom were present for the activity. Approximately 60% of the students were from the United States, and 40% were international students. The second class consisted of 55 students, 49 of whom were present for the activity; approximately 70% from the United States, and 30% were international. Both years, in the class session prior, students had been introduced to basic forecasting concepts and components of time-series forecasting, including trends, seasonality, product life cycles, and randomness, through a lecture and several in-class examples. This game was presented as an in-class activity, which was not specifically graded but was included in each student’s participation grade for the semester.

Both the case and the data set were shared with students via the Learning Management System, as is standard for this course (students bring their laptops to every session). Students worked in the groups that they had been assigned for the entire semester. These teams had formed using CATME, a web-based tool for forming teams based on a variety of factors. In this particular course, the instructor chose algorithm parameters that would prioritize grade point average (GPA), schedule compatibility, and self-identification as members of underrepresented races/ethnicities and genders (Layton et al. 2010). The software, therefore, seeks to match students of similar GPA and availability for group meetings and to ensure that each group contains either zero or two or more students of underrepresented races/ethnicities and genders.

3.1. Student Performance

The framing of the questions as mysteries to be solved by the students (either finding the root cause or finding the mystery item) drove student engagement as compared with exercises that ask for numeric output. Although undergraduate students often feel frustrated at being given a task without a recommended set of steps for completing the task, the low-stakes nature of the game as an in-class activity allowed students to focus on problem-solving. Students spent the group time actively working on the activity, and most of the discussion overheard during the time allotted was related to holidays and groceries, two topics obviously central to the case. Across the two class sessions, only one group did not seem to know how to approach the activity at all and were able to get back on track with minimal prompting by the instructor, specifically asking the team what happens in the third or fourth week of November in the United States.

In the first use of this game, no group was able to identify the root causes of the historical data for all three of the known products, and about one third of groups identified the mystery item. More detailed information was collected in the second use of this game. The 49 students present divided into 14 groups, 13 of which identified the relationship between American Thanksgiving and orange juice. Eight groups identified the relationship between Medjool dates and Ramadan, and three identified the relationship between Mexican Coke and Passover; however, only one group identified both. During the full-class debrief, the instructor asked how many individual students knew the underlying patterns before discussing the data with their groups. Ramadan and Passover were recognized by 13 students and 3 students, respectively, so the majority of groups relied heavily on the knowledge of only one or two students for each product. The four groups that did not identify either were asked about their thought process; three of them said they had assumed they were holiday related but could not find a holiday that made sense, whereas the last group said they thought the Mexican Coke pattern was something related to outdoor activities (we are situated in a climate where the first “outdoor” weather of the year typically falls in March or April).

In guessing the mystery item, five groups (16 students or 33%) correctly guessed that the mystery item was a snack food, such as hot dogs or chips. Another four groups (13 students) guessed chocolate, candy, or flowers. Based on the discussion once the answer was revealed, the latter guesses were based on students viewing February and May spikes in sales as being related to Valentine’s Day and Mothers’ Day, respectively, rather than the Super Bowl and Memorial Day. When further pressed to explain the July (U.S. Independence Day) and September (Labor Day) spikes, they admitted that their guesses did not fit the data as well as the correct answer.

During the debrief period, students were asked to think about what type of team they would assemble if they were about to be given another set of items to identify. In one class session, the first response specifically made mention of working with people from other cultures, whereas in the other session, the first response was more generally about working with people “who know stuff you don’t.” One student pointed out that even if you do not know all the holidays for every culture, just stopping to think about your different customers would be important, and from there, students were able to give additional examples. These were primarily in the realm of product design as many students had seen high-profile failures on social media. Overall, students in both sections were able to make the connection between diversity and business value.

3.2. Student Learnings

In the first class where this was used, the standard procedure was for students to turn in an index card at the end of each class session with their name and either a question about the day’s material or the main thing they had learned. For the class session in which this game was used, almost all students (46 of 49) provided the latter, and the remaining 3 asked questions related to administrative matters. In the second class where this was used, students were specifically asked to write a key learning for the class session on their index card. The key learnings for both years were sorted into four categories.

• Importance of a diverse team. Responses in this category explicitly mentioned the value of diverse teams in creating forecasts.

• Importance of multicultural knowledge. Responses in this category focused on the value of knowing about holidays in other cultures.

• Importance of human knowledge. Responses in this category discussed the value of human knowledge as compared with fully computer-generated forecasts.

• Other. Responses in this category included general forecasting processes, Excel functions, and anything else not included above.

The total number of responses in each category across both classes is summarized in Table 1. Over one third of students focused on diversity or multiculturalism, and another third focused on the importance of human knowledge in forecasting. These takeaways match closely with the planned learning objectives and demonstrate the value of this activity.

Table 1. Summary of Students’ Key Learnings

3.3. Possible Variants

This activity is well suited for adaptation to various learning audiences, and instructors wishing to do so may consider options such as the following.

• For students with more advanced forecasting knowledge, other components of time-series forecasting, such as trends or monthly seasonality, could be added. Students would, therefore, need to isolate various patterns in the historical sales data to find the ones affected by cultural celebrations.

• For classes outside of the United States, other cultural celebrations could be substituted as appropriate. It is recommended that the first item be something very recognizable to students from the dominant culture and that subsequent items be less so.

4. Conclusion

This case is a tool for instructors who wish to introduce DEI concepts in their courses in a context relevant to supply chain management students. The case material and the questions asked of students help them learn about the value of diverse teams while completing a quantitative analysis drawn directly from real-world examples. This integration allows an instructor to emphasize the importance of cultural diversity while potentially bypassing the typical resistance seen with solely DEI-focused activities.