The use of Big Data in Fashion Forecasting

Abstract

This research paper investigates the impact Big Data on the fashion industry. The concept of data analysis has introduced new perspectives and opportunities to the business world to determine the existing trends and customer behavior in a scientific methodology. Data analytics not only provides a method to analyze the data sets collected from the consumer interactions, but also enables organizations to make informed business decisions for their future operations. The effectiveness of Big Data relies on the quality of the customer data collected from various sources, especially the digital media. The aim of this paper is to determine the effects of Big Data in the fashion industry by the means of fashion trend forecasting. Based on the available academic studies and analyzed business practices around the world, it was determined that the use of Big Data in the fashion industry greatly improves the fashion forecasting of brands and gives a significant competitive advantage over their rivals.

Introduction

With the rapid developments in the digital technologies, the traces of the mankind have turned into digital footprints. The enormous amount of data is being generated continuously whenever people use their smartphones or computers by leaving a digital trail of information behind. This profoundly complex and simultaneously produced “big data” has introduced various new opportunities to the business world. Today, the companies are effectively using big data to determine the customer insights and behavior through the collected online traces of people. The analysts are finding ways to improve their products and services. These improvements not only shape people’s lives also shape the experiences people are having with the world around them.

Considering the continuous changes in customer demands, the fashion industry purely relies on its ability to discover and develop new trends. Eventually, analyzing and determining the customers’ interests and insights towards various designs, products and brands becomes a crucial success factor in the fashion industry. According to the article “Big Data in Fashion: Transforming the Retail Sector” , even though the fashion industry mainly depended on intuition and creativity for direction in designing, buying and merchandising in the past; today, the use of big data is the integral part of their business activities including market identification, trend analysis, understanding the customer, converting high ticket purchases, lifting new designers, measuring influencers’ impact and improving cross selling. As stated before, rapidly changing customer trends mean rapid change in the target audience, therefore these fashion companies need to keep up with the current demands. The collected data of the people lead the companies to create new collections and styles.

Taking into account of the recent data-based developments across the globe, investigating the role of big data in the fashion industry is essential. By embracing the data related approaches, most of the fashion brands will indicate what products will work for their audiences and their targeted markets from season to season. Therefore, this research paper focuses on how the use of big data affects the fashion industry by the means of fashion trends forecasting. The fashion forecasting can be defined as predicting the colors, fabrics, textures and other styles that will be presented on the runway and in the stores for the upcoming seasons. This forecasting is done by the analysis of existing data to spot the approaching trends. The existing data usually includes facts and numbers as well as purchasing and sales statistics from previous years. These findings contribute to the collection of styling ideas to bring out an emerging trend to influence the fashion market. The goal is to foresee the future recognition of the prospective fashion styles.

Big Data presents a niche opportunity for the fashion forecasting in terms of analyzing the current demands and trends. Therefore, after describing the old method of fashion trend forecasting, this academic paper will focus on the use of big data in the fashion trend forecasting in terms of its advantages and current applications. The leading research tools of this academic paper are academic studies, and journal articles. In addition, websites such as Vogue Business, Data Science Central and Launch Metrics are reviewed in order to understand the point of views of different fashion brands in terms of big data’s impact on the future fashion trends.

Fashion trend forecasting before Internet

The fashion trend forecasting was a very straight forward process in the pre-internet world. The designers, fashion houses and independent fashion forecasters were simply defining the fashion trends by deciding which style, design or color would cause the biggest impact on the consumers based on their pure creativity. The images of these predetermined fashion products would be introduced to the public by various fashion magazines, and then the products would be delivered to the marketplace for consumption after 12 to 18 months. In addition, the fashion industry was mostly dependent on companies’ old data including their sales records and in-house inventory details. The fashion brands were only able to compare the different number of sales between each year and collection. The companies were also trying to gather some information about their competitors’ sales from different and sometimes unreliable sources. Eventually, collecting and analyzing the limited amount of data and purely relying on the creativity and direction of the fashion forecasters were not sufficient to correctly determine the actual customer interests and demands.

Using big data in in fashion trend forecasting

With the introduction of internet and continuous developments in the digital technologies and social media, enormous amount of data related to the fashion industry has become available. The retailers, apparel designers and manufacturers could make more accurate evaluations based on the new and comprehensive data sets. Today, companies are embracing the big data to reshape the fashion industry in new and innovative ways. Since the customers gradually expect and demand personalized shopping experiences, only focusing on the historical sales is inadequate in today’s business dynamics. For instance, well known Spanish apparel retailer, Zara is one of the most effective users of big data. Their team daily analyzes the big data collected from various sources to determine their customers’ needs and demands, and they translate their analysis into tangible designs. Some brands go even further in the use of big data like Ralph Lauren with breakthrough innovations. The company recently teamed up with an advanced bio sensing apparel manufacturer, OMsignal to design a new product called “the PoloTech Shirt”. This new product was designed for active sportsmen and equipped with numerous sensors. The aim of this design is to collect real-time data on the wearer’s direction and movement, including biometric data such as heart and breathing rates. The company is currently analyzing the transmitted data by using special algorithms in order to determine the new fashion trends for their world-renowned Polo Shirts.

Today, there are mainly two methods of using big data in fashion trend forecasting: human-based and software-based analysis. In the human-based fashion trend forecasting, the big data consisting of fashion images collected from social media outlets, fashion shows, visual merchandising, blogs, etc. was analyzed by the specialized trend forecasters who are responsible for identifying new trends and predicting how those trends will shape the fashion industry. There are two important factors in the success of human-based fashion trend forecasting. First, the collected data must be sufficient and relevant in reflecting the actual customer interests and demands. Second, the trend forecasters must possess required analytical skills and creativity as well as detailed knowledge about the fashion industry in order to analyze and interpret the related data based on their observations, interviews, and media scans. They must also be media savvy in order to follow up the consumer behavior and trends in the social media. Some fashion companies such as Urban Outfitters have in-house trend forecasting teams and specially developed big data analysis models. On the other hand, there are some independent trend forecasting agencies such as WGSN, providing services to various industries including fashion industry.

In the software based fashion trend forecasting, various data analytics software and artificial intelligence based software programs are used in the analysis of the big data instead of specialized trend forecasters. The aim is to eliminate the time consuming and labor-intensive human-based analysis and to process the large amounts of data faster by learning and memorizing the consumers’ preferred styles and feedbacks. These programs specifically concentrate on the computational representation of garments, modeling human behavior towards different styles and detecting the current fashion trends. A typical software-based fashion trend forecasting program works in three steps. In the first step, the target consumer profile is determined for the specific data collection process. Then, the specific software program consisting of computer vision technology is applied to the images and pictures collected from social media, e-commerce web sites and other sources. The computer vision technology detects and categorizes various components from the shapes to fabrics, prints, and colors. In the final step, the machine learning algorithms are used to predict the fashion trends based on the components detected in the previous step. There are various AI methods currently used for fashion trend forecasting such as Fuzzy logic, genetic algorithms, neural networks, decision trees, Bayesian networks and knowledge based systems and their variations. EDITED is one of the well-known companies providing valuable information to the world’s biggest fashion brands and retailers by using its A.I. and advanced machine learning tool. The web site, Google Trends might be the best example of how big data can be exploited and visualized in order to forecast the fashion trends of consumers. Based on academic studies, it is estimated that almost 70% of luxury purchases were influenced by online interactions. Google Trends simply analyzes the popularity of top search queries in Google Search across various regions and languages by using data analytics software and provides detailed information about its analysis. Eventually, many fashion companies are actively using Google trends for forecasting different fashion variables such as Identifying Seasonal Patterns in Demand and predicting fashion consumer behavior.

In 2018, Mikayla DuBreuil and Sheng Lu from the University of Delaware conducted an empirical study to compare the similarities and differences between human-based and software-based trend analysis, and they concluded that using big data software tools in fashion trend forecasting has a great potential. Their findings indicate that the artificial intelligence based software programs are very effective in forecasting color and apparel pattern. On the other hand, human-based trend analysis seems to be more effective in predicting design details consisting more complex factors.

Conclusion

Even today, there are a lot of speculations and debate over the concept of trend forecasting. While some experts claim that trend forecasting creates new trends, others believe that it only identifies the existing trends. In addition to this long standing debate, there is also a debate among academicians regarding the effectiveness of using big data analytics for fashion trend forecasting. For instance, Ayelet Israeli and Jill Avery from Harvard Business School claim that using the historical purchasing data is more feasible and reliable to predict the future fashion trends since the fashion taste of consumers stay relatively stable over time. On the other hand, Ali Fallah Tehrani and Diane Ahrens from Deggendorf Institute of Technology in Germany claim that the use of big data may significantly improve the accuracy of fashion trend forecasting because it is more factual based rather than relying on designers’ opinionated guesswork. Even though new academic studies will continue to be conducted to explore this issue more in the future, it is a clear fact that the big players in the fashion industry are seriously relying on the use of big data for trend forecasting.

References

1. Silva, Emmanuel Sirimal, Hossein Hassani, and Dag Øivind Madsen. “Big Data in Fashion: Transforming the Retail Sector.” Journal of Business Strategy ahead-of-print, no. ahead-of-print (2019). https://doi.org/10.1108/jbs-04-2019-0062.

2. Wong, Mei Yen, Yilu Zhou, and Heng Xu. “Big Data in Fashion Industry: Color Cycle Mining from Runway Data.” Big Data in Fashion Industry: Color Cycle Mining from Runway Data, 2016, 1–10.

3. Al-Halah, Ziad, Rainer Stiefelhagen, and Kristen Grauman. “Fashion Forward: Forecasting Visual Style in Fashion.” 2017 IEEE International Conference on Computer Vision (ICCV), 2017. https://doi.org/10.1109/iccv.2017.50.

4. Copeland, Lauren, Giovanni Luca Ciampaglia, and Li Zhao. “Fashion Informatics and the Network of Fashion Knockoffs.” First Monday, February 2019. https://doi.org/10.5210/fm.v24i12.9703.

5. Dubreuil, Mikayla, and Sheng Lu. “Traditional vs. Big-Data Fashion Trend Forecasting: an Examination Using WGSN and EDITED.” International Journal of Fashion Design, Technology and Education 13, no. 1 (February 2020): 68–77. https://doi.org/10.1080/17543266.2020.1732482.

6. Silva, Emmanuel, Hossein Hassani, Dag Madsen, and Liz Gee. “Googling Fashion: Forecasting Fashion Consumer Behaviour Using Google Trends.” Social Sciences 8, no. 4 (April 2019): 111. https://doi.org/10.3390/socsci8040111.

7. Shi, Mengyun, and Van Dyk Lewis. “Using Artificial Intelligence to Analyze Fashion Trends.” Using Artificial Intelligence to Analyze Fashion Trends, n.d., 1–39.

8. Rogers, Richard, “Digital Methods for Cross-Platform Analysis” in The SAGE Handbook of Social Media. London: Sage, 2017.

9. Hutchinson, Jonathon, “An introduction to digital media research methods: how to research and the implications of new media data”, Communication Research and Practice, 2:1, 1-6, DOI: 10.1080/22041451.2016.1155307

10. Tehrani, A. F., & Ahrens, D. (2016). Improved forecasting and purchasing of fashion products based on the use of big data techniques. In Supply Management Research (pp. 293-312). Springer Gabler, Wiesbaden.

11. Israeli, A. & Avery, J. (2018). Predicting consumer tastes with big data at Gap. Harvard Business School (pp. 1-27). Retrieved from https://files.transtutors.com/cdn/uploadassignments/2661391_2_gap.pdf

12. “How to Become a Trend Forecaster: Career Advice & Interview Tips: WayUp Guide.” Career Advice & Interview Tips | WayUp Guide, June 29, 2018. https://www.wayup.com/guide/how-to-become-a-trend-forecaster/.

13. Wang, Haosha, Joshua De Haan, and Khaled Rasheed. “Style-Me – An Experimental AI Fashion Stylist.” Trends in Applied Knowledge-Based Systems and Data Science Lecture Notes in Computer Science, 2016, 553–61. https://doi.org/10.1007/978-3-319-42007-3_48.

14. “How Heuritech Forecasts Fashion Trends Thanks to Artificial Intelligence.” Heuritech, March 30, 2020. https://www.heuritech.com/blog/articles/how-heuritech-forecasts-fashion-trends-thanks-to-artificial-intelligence/.

15. “About.” EDITED. Accessed May 27, 2020. https://edited.com/about/.

16. “Google Trends.” Wikipedia. Wikimedia Foundation, May 15, 2020. https://en.wikipedia.org/wiki/Google_Trends.