Prelims

Half Title Page

AI in Fashion industry

Title Page

AI in Fashion industry

By

Dr. Satya Banerjee

National Institute of Fashion Technology, India

Dr. Sanjay Mohapatra

Xavier Institute of Management, India

And

Dr. M. Bharati

Veer Surendra Sai University of Technology, India

United Kingdom – North America – Japan – India – Malaysia – China

Copyright Page

Emerald Publishing Limited

Howard House, Wagon Lane, Bingley BD16 1WA, UK

First edition 2022

Copyright © 2022 Satya Banerjee, Sanjay Mohapatra and M. Bharati. Published under exclusive licence by Emerald Publishing Limited.

Reprints and permissions service

Contact: permissions@emeraldinsight.com

No part of this book may be reproduced, stored in a retrieval system, transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without either the prior written permission of the publisher or a licence permitting restricted copying issued in the UK by The Copyright Licensing Agency and in the USA by The Copyright Clearance Center. Any opinions expressed in the chapters are those of the authors. Whilst Emerald makes every effort to ensure the quality and accuracy of its content, Emerald makes no representation implied or otherwise, as to the chapters' suitability and application and disclaims any warranties, express or implied, to their use.

British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library

ISBN: 978-1-80262-634-6 (Print)

ISBN: 978-1-80262-633-9 (Online)

ISBN: 978-1-80262-635-3 (Epub)

List of Figures

List of Tables

Acknowledgement

Though on the cover of this book, I shall remain thankfully indebted to all those learned souls, known and unknown hands who directly and indirectly motivated me to achieve this goal and enlightened me with the touch of their knowledge and constant encouragement. The words ‘the one who directs the path of progress is angelic’ are inadequate to express my deep sense of thankfulness to my parents, the two wonderful human beings, for their resolute guidance, unwavering encouragement, abiding interest, constructive criticism, tremendous enthusiasm and malicious supervision throughout this project.

With stupendous ecstasy and profundity of complacency, I pronounce my deep sense of gratitude to Prof Rahul Thakurta, Associate Dean, Doctoral Program, Prof Shridhar Kumar Dash, Dean, XIMB, for giving me this opportunity to work on this project and for providing me with unmatched support, infrastructure and resources. A special thanks to Dr Fr. Anthony R. Uvari, S.J., whose charismatic leadership and transformative vision of excellence motivated me to deliver this piece of work.

I am thankful to my colleagues at work for giving me time and space for delivering this. I am also grateful to my fellows at XIMB for supporting me in various ways throughout this period.

A special thanks to my wife, Anupama, for being a pillar of inspiration and motivation in this journey. I also thank Prof Sanjay Mohapatra (my co-author here), for being my partner through this journey.

I thank God, Heaven, Fathers and Forefathers for protecting all of us in these difficult times…

Abstract

Fashion is a fabulous industry. It is perceived with glamour, vibrance, beauty, money, fame and massive profits from the outside. However, from the inside, it is a wounded industry. An industry with a size of 3 billion USD, 150 billion units of products per year, and annual growth of 3–4% CAGR, the fashion industry is characterized by low shelf-life products, wrong forecasts, low inventory turnovers, frequent discounts, low realized margins and operating profits, and ever-increasing competition. Best of the fashion retailers make operating profits of 10–12%, making it extremely difficult for most fashion businesses to sustain. An investigation into the existing literature enabled us to conclude that most of the fashion industry's challenges point out wrong forecasts. On the periphery, technology is rapidly invading fashion industry with the most emerging forms such as Artificial Intelligence, Machine Learning, Deep Learning, Artificial Neural Networks, Human-Robot Interface and a list of others, already making their way into this industry in recent years. The field of fashion forecasting in light of data-driven intelligence does not remain untouched by these new developments in the practitioners' world; however, very little has been documented in this area's academic literature.

In this piece of work, we address some of these issues. We start with an exhaustive literature review in the fashion industry and narrow it down to the fashion forecasting industry. We discuss some recent works in fashion forecasting, thereby developing a ‘framework of AI-based fashion forecasting’ and empirically validate the framework with a qualitative case study of the world's first fashion intelligence company based in Bengaluru, India. We observe that the internet and particularly social media have a lot to offer in terms of data and especially photographs or images of consumers that carry information on what they wear. We attempt to study the relationship between fashion and social media engagement of fashion consumers and reveal that ‘fashion identity’ is the connecting element between fashion motivations and social media motivations. We create a ‘conceptual model of fashion identity’ from existing literature to answer one of our research questions, ‘Why Social-media-based information can reveal fashion forecasts?’. From here, we create a ‘framework of short-range fashion forecasting’ and argue on how internet may assist in fashion forecasting. Subsequently, we move to our primary objective to create a ‘conceptual framework of fashion e-forecasting’. As the name suggests, this framework may create forecasts based on data from social media, e-commerce, and other web data. After conceptually developing this framework based on previous frameworks present in this area and available literature, we validate this empirically through a case study. The case study chosen for this purpose addresses forecasting based business problem of a family owned fashion retail business. We collect data in the form of photographs of consumers on their social media pages using a popular and emerging research method called ‘Netnography’ and convert this into attributes and labels using numeric binary coding. A total of 176 photographs were picked from 634 interested participants for further study. Using hierarchical clustering followed by k-means clustering on software SPSS, 7 clusters or popular combinations of attributes and labels were retrieved, giving rise to 20 popular styles of the chosen product that consumers are wearing now. We finally made illustrations of these 20 popular styles as an output of the research. The case study validates our hypothesis that fashion forecasting or ‘nowcasting’ in the present context may be done by using data from the internet.

The present study is unique in multiple ways. First, it suggests a novel method of fashion product development in the light of data-driven intelligence; second, it documents some of the rapid developments in the field with the onset of technology. It also addresses some of the fundamental questions that are becoming more relevant in the recent years.