Towards Digital Transformation in Fashion Retailing: A Design-Oriented IS Research Study of Automated Checkout Systems

Abstract

Automated checkout systems promise greater sales due to an improved customer experience and cost savings because less store personnel is needed. The present design-oriented IS research study is concerned with an automated checkout solution in fashion retail stores. The implementation of such a cyberphysical system in established retail environments is challenging as architectural constraints, well-established customer processes, and customer expectations regarding privacy and convenience impose limits on system design. To overcome these challenges, the authors design an IT artifact that leverages an RFID sensor infrastructure and software components (data processing and prediction routines) to jointly address the central problems of detecting purchases in a reliable and timely fashion and assigning these purchases to individual shopping baskets. The system is implemented and evaluated in a research laboratory under real-world conditions. The evaluation indicates that shopping baskets can indeed be detected reliably (precision and recall rates greater than 99%) and in an expeditious manner (median detection time of 1.03 s). Moreover, purchase assignment reliability is 100% for most standard scenarios but falls to 42% in the most challenging scenario. Keywords Design-oriented IS research Á Digital innovation Á Internet of things Á Cyberphysical systems Á Retail industry Á Radio frequency identification Á Machine learning Á Automated checkout systems Accepted after two revisions by the editors of the special issue.

Debating Digital Innovation: A Literature Review on Realizing Value from Digital Innovation

“…Looking at the underlying technologies, we can identify which technologies are particularly relevant to create digital innovation value. For instance, technologies such as analytics and business intelligence [30, p. 8], augmented reality [31,32], cloud computing [33], CRM systems [19], digital control systems [34], digital platforms [30,35], digital prototyping technologies [36], project management systems [37], RFID [38], social media [39] and 3D printing [2] are some common technological foundations for digital innovation value. We uncover that digital innovation succeeds effectively in organizations when an orchestrated integration of technology in the work system and a collective consideration of the social and technical systems within an organization take place [40,41].…”

False Positive RFID Detection Using Classification Models

“…Static tags occurred when the tags were located within the configured read range, whereas moved tags are tags that were moved through the reader gate. However, several previous retail store domain studies have shown that certain moved tags (e.g., tag movement close to the gate) were considered as false positives [23,24]. Hauser et al (2015) investigated classification models to detect false-positive RFID readings in retail stores to help prevent shop theft [23].…”

“…RSS based attributes were extracted as features for the considered classification models: DT, LR, random forest (RF), artificial neural network (ANN), and SVM. Another study revealed that integrating machine learning models with RFID could enable automated checkout systems, particularly for fashion retailing [24].…”

“…Previous studies have successfully applied ANN, LR, SVMs, DT, RF, and XGBoost machine learning models to detect false positives with high classification accuracy using features extracted from tag RSSs [21][22][23][24].…”

“…Machine learning models have also been utilized to successfully detect false positives using RSS based features as inputs. However, some retail store-based studies showed that tag movement close to the gate and through the gate were classified as false positive and true positive readings, respectively [23,24]. Therefore, false positive classification must consider other tag movement types, such as tag movement close to the gate, as well as static tags for WMSs.…”