Citation

Mithra, Krishnamoorthy Sathya and Karthikeyan, Ganesh and Nagarajan, Deivanayagampillai and Vellaichamy, Parthasarathy (2026) Demand forecasting for Adidas using long short-term memory (LSTM) networks. In: Data-driven Decision Making and Soft Computing. CRC Press, pp. 259-277. ISBN 978-100363483-6, 978-104106297-4, 978-104106314-8

Text Data‑driven Decision Making and Soft Computing; Applications and Advances.pdf - Published Version Restricted to Repository staff only Download (1MB)

Abstract

In today’s fast-evolving and highly competitive world, accurate demand forecasting is crucial for managing an organization’s supply chain, reducing inventory cost, and optimizing production. Adidas, a global brand in the athletic footwear and apparel industry, faces challenges in predicting sales trends due to fluctuating consumer demand, market uncertainties, and external disruptions. Such complexities are hard to handle with traditional forecasting methods, which leads to the need for advanced techniques. This study integrates long short-term memory (LSTM) networks, a deep learning approach well-suited for time series forecasting, with fuzzy logic, which enhances decision-making by handling uncertainty and imprecise data. By leveraging this hybrid approach, Adidas can achieve more precise demand predictions, minimize surplus inventory, and improve strategic decision-making in supply chain management. Adidas is a German multinational corporation (MNC) that designs and manufactures sports shoes, clothing, and accessories, and was founded in 1949 by Adolf Dassler. The Adidas Group consists of Reebok sportswear company, TaylorMade Golf Company, FC Bayern Munich, and Runtastic, which are headquartered in Olympiaring, Herzogenaurach, Germany. The German national football team’s victory in the 1954 World Cup brought Adidas’ name to football pitches everywhere, as they made the lightweight football boots for the national team. According to the Adidas brand value is more than $17.1 billion. Today, Adidas has 1746 concept stores, 779 factory outlets, and 316 concession corners and others around the world. The Adidas Group’s net sales amounted to about €14.53 billion in 2014 and have more than 53731 employees from around the world. The mission of the Adidas Group strives to be the global leader in the sporting goods industry with brands built on a passion for sports and a sporting lifestyle (Jayawardhana, 2016). Adidas experienced an accumulation of unsold inventory over the course of 2022, which was caused by a 20% decrease in consumer demand. The issue is further worsening due to the controversy of Kanye West, Adidas’ current Brand Ambassador. Kanye West expressed an antisemitic (anti-Jews) comment, triggering a dislike from Adidas’ consumers who don’t like racism. This results in a slowdown of new goods’ production and accumulation of old goods within Adidas’ storage. Essentially, this disrupts the goods production and storage within Adidas. For a company as big as Adidas, disruption in good production and storage can result in multiplied loss, because Adidas has so many warehouses, and produces so many goods, even on a weekly basis. A disruption in storage and production will eventually affect all of these other warehouses and stores, creating a bottleneck domino effect. To prevent this from happening, Adidas must create a good supply chain condition. More than just to prevent the disruption, however, supply chain management is also important for Adidas to increase its capability in serving thousands of consumers every day (Witjaksono, 2023). Demand forecasting is a prominent business use case that allows retailers to optimize inventory planning, logistics, and core business decisions. One of the key challenges in demand forecasting is accounting for relationships and interactions between articles. Most modern forecasting approaches provide independent article-level predictions that do not consider the impact of related articles. Recent research has attempted to address this challenge using Graph Neural Networks and showed promising results. This paper builds on previous research on Graph Neural Networks and makes two contributions. First, we integrate a Graph Neural Network encoder into a state-of-the-art DeepAR model. The combined model produces probabilistic forecasts, which are crucial for decision-making under uncertainty. Second, we propose to build graphs using article attribute similarity, which avoids reliance on a predefined graph structure (Kozodoi et al., 2024). Satisfied customers become brand advocates, promoting the brand through positive word-of-mouth and repeat purchases. They are also more likely to provide constructive feedback that helps the brand evolve and improve. Conversely, dissatisfied customers can damage a brand’s reputation through negative reviews and social media complaints. The abstract emphasizes the importance of integrating customer satisfaction into brand strategy. This can be achieved by actively collecting customer feedback, addressing concerns promptly, and exceeding expectations throughout the customer journey. Finally, the research highlights potential future directions by exploring the role of emerging technologies like social media and artificial intelligence in shaping brand–customer relationships and driving customer satisfaction. A critical analysis of the Adidas Group’s success requires delving into both its branding strategies and customer satisfaction. Examining their rich heritage and focus on innovation alongside the effectiveness of their multi-brand portfolio is crucial. Understanding how they leverage celebrity endorsements and sponsorships to build brand loyalty is also important. On the customer satisfaction side, research should explore how product quality, design, and technology drive positive experiences. Additionally, the effectiveness of their customer service channels and how they utilize feedback for improvement deserves scrutiny. A balanced approach necessitates evaluating criticisms of brand saturation and ensuring their image aligns with evolving consumer preferences like sustainability. Benchmarking against competitors like Nike will provide valuable insights. Finally, incorporating data from market research, customer reviews, and brand perception studies, along with established marketing theories like Net Promoter Score (NPS) and Service Quality Model, Hopfield Neural Network (SERVQUAL), will solidify your analysis, revealing the Adidas Group’s strengths and opportunities for continued success (Singh & Srivastava, 2024). A demand forecasting method based on multi-layer LSTM networks has been proposed. The proposed method automatically selects the best forecasting model by considering different combinations of LSTM hyperparameters for a given time series using the grid search method. It has the ability to capture nonlinear patterns in time series data while considering the inherent characteristics of non-stationary time series data. The proposed method is compared with some well-known time series forecasting techniques from both statistical and computational intelligence methods using the demand data of a furniture company. These methods include autoregressive integrated moving average, exponential smoothing, artificial neural network, K-nearest neighbors, recurrent neural network (RNN), support vector machines, and single-layer LSTM (Abbasimehr et al., 2020). LSTM models forecast the sales in the marketplace and are compared with different machine learning models to predict the demand in the future. With the recent advancement of deep learning architecture, it is possible to handle large, voluminous market data. In this article, we have proposed an LSTM network that takes the historical sales data of the products as input and forecasts the demand of each product for the next three time series (Lakshmanan et al. 2020). The enhancements in digital technologies entail a shift in consumer behavior and expectations toward brands and retailers. Especially, consumers belonging to Generation Z are enabled through digital means to self- control their consumer journey. While this benefits E-commerce channels, “brick-and-mortar” stores need to redefine their role to serve consumers with more meaningful value propositions that go beyond the purchase of tangible products. The reopening of the Adidas brand flagship store in London in October 2019 gives cause for the design of a new service concept with the objective of harnessing the strengths of physical retail and increasing relevance for Gen Z consumers. An experiential service that empowers consumers during their journey toward self-actualization differentiates from the offerings of convenience-driven E-commerce channels and allows consumers to invest in their personal growth and to contribute to a higher purpose (Sepoetro, 2018). Fuzzy Logic was initiated in 1965 by Lotfi A. Zadeh, professor of computer science at the University of California, Berkeley. Basically, fuzzy logic is a multivalued logic that allows intermediate values to be defined between conventional evaluations like true/false, yes/no, high/low, etc. Notions like rather tall or very fast can be formulated mathematically and processed by computers, in order to apply a more human-like way of thinking in the programming of computers. Fuzzy systems are an alternative to traditional notions of set membership and logic that have their origins in ancient Greek philosophy. The precision of mathematics owes its success in large part to the efforts of Aristotle and the philosophers who preceded him. In their efforts to devise a concise theory of logic and later mathematics, the so-called “Laws of Thought” were posited. One of these, the “Law of the Excluded Middle,” states that every proposition must either be True or False. Even when Parmenides proposed the first version of this law (around 400 B.C.), there were strong and immediate objections: for example, Heraclitus proposed that things could be simultaneously True and not True. It was Plato who laid the foundation for what would become fuzzy logic, indicating that there was a third region (beyond True and False) where these opposites “tumbled about.” Other, more modern philosophers echoed his sentiments, notably Hegel, Marx, and Engels. But it was Lukasiewicz who first proposed a systematic alternative to the bi–valued logic of Aristotle. Even in the present time, some Greeks are still outstanding examples for fussiness and fuzziness (note: the connection to logic got lost somewhere during the last 2 milleniums). Fuzzy Logic has emerged as a profitable tool for the controlling and steering of systems and complex industrial processes, as well as for household and entertainment electronics, and for other expert systems and applications like the classification of SAR data (Hellmann, 2001). A fuzzy set is a class of objects with a continuum of grades of membership. Such a set is characterized by a membership (characteristic) function that assigns to each object a grade of membership ranging between zero and one. ©

Item Type:	Book Section
Uncontrolled Keywords:	Commerce, Data handling, Decision making
Subjects:	H Social Sciences > HF Commerce
Divisions:	Faculty of Engineering and Technology (FET)
Depositing User:	Ms Rosnani Abd Wahab
Date Deposited:	03 Sep 2026 04:53
Last Modified:	03 Sep 2026 04:53
URII:	http://shdl.mmu.edu.my/id/eprint/16611

Downloads

Edit (login required)