State of the art of item-level RFID deployments in fashion and apparel retail

Abstract

Fashion and apparel retail have been driving the adoption of RFID for the last 10 years. This paper strives to present an overview of the current adoption of RFID technology in this sector and to improve the understanding of the evolution of use cases (UCs) over time by proposing a framework for the UC lifecycle. This tool allows us to classify UCs with six different labels (core, intermittent, trendy, vintage, innovative or ghost) according to three dimensions. We analysed and classified more than 220 scientific papers, magazine articles, white papers, web news stories and press releases as well as confidential information from system integrators dealing with UHF RFID projects in fashion and apparel retail in recent decades. We are not aware of any other study covering this issue, and the results we obtain represent an updated and thorough overview of UHF RFID adoption in the fashion industry. We believe that, by keeping the framework updated, we could create a barometer of RFID adoption in fashion and apparel retail while obtaining further insights on UC evolution over time.

Keywords

RFID deployment fashion apparel retail item-level tag use cases framework

1 Introduction

Radio frequency identification (RFID) is a pervasive wireless technology that uses a transmitted radio signal to automatically tag, recognise, track and trace the location and movement of an item (Bottani, Montanari, & Romagnoli, 2016). Although several different methods of identifying items using RFID are available (Wu, Nystrom, Lin, & Yu, 2006), all RFID systems are composed of two elements: a reader and a tag (or transponder). The former broadcasts an electromagnetic field and both transmits and receives a radio signal, while the latter transmits its identification, along with additional information, if available.

To do this, a tag consists of at least an integrated circuit and an antenna operating on a specific frequency, and it may also be equipped with a limited capacity to store data, according to the application (Ngai, Moon, Riggins, & Yi, 2008). RFID systems can be considered an evolution of barcodes, and they are currently used for tracking and tracing different items or assets in different sectors (Esposito, Romagnoli, Sandri, & Villani, 2015).

The use of RFID for product tagging has been constantly increasing throughout the supply chains of different sectors due to the many benefits of this technology (Esposito et al., 2015; Hardgrave, Aloysius, & Goyal, 2013; Uckelmann & Romagnoli, 2016). Several researchers suggest that the most important breakthrough in RFID technology has been the jump to item-level tagging (Gaukler, 2010; Hardgrave, 2009), the first implementations of which were seen in the early 2000 s in fashion and apparel retail logistics (Gaukler, Seifert, & Hausman, 2007). Just a few years later, forerunner companies envisioned the value and profit that could be generated by pushing item-level RFID-tagging down to the retail stores (Rizzi, Romagnoli, & Thiesse, 2016). Indeed, fashion and apparel retail has been driving the adoption of UHF RFID for the last 10 years. In 2017, the industry encoded more than 6 billion UHF RFID tags, with a forecasted RFID turnover increase of more than 30% by 2022 according to IdTechEx (Das, 2017). The reason for this is straightforward. No other industry has experienced so many use cases (UCs) for RFID deployments that proved to deliver real value in real environments to manufacturers, third-party logistics service providers and retailers. Moreover, the ability to track item movements from the backroom to the shop floor (and vice versa) can enable greater visibility of store performance and better explain the sell-out values of different stock keeping units (SKUs), by monitoring current stock levels and, possibly, avoiding out of stock situations. In this way, item-level tagging can result in stock-out reduction and better inventory control as well as labour savings, reduced transaction errors and reduced shrinkage (Bottani, Ferretti, Montanari, & Rizzi, 2009). Additionally, RFID is an essential enabler of omnichannel retailing, or ‘omnichanneling’, the goal of which is to provide customers with ‘a consistent and seamless experience whether they’re shopping in a store, on a mobile device, on a home computer or via a catalogue’ (Hardgrave, 2012). According to Bertolini et al. (2017), important factors that explain the widespread adoption of item-level RFID tagging in the fashion and apparel supply chain include the following: (a) the ability to track and trace items from the application of the tag, which may even be at the manufacturing plant, to the point of sale, (b) the high marginality of the fashion and apparel sector, (c) the large number of SKUs that must be managed (i.e. model size, colour, season), (d) ideal environment – absence of liquids and metals, (e) short product lifecycle and (f) the significant developments in information and communication technology that make it possible to manage the massive amount of data generated by item-level tagging.

In their 2016 paper, Rizzi et al. (2016) listed a total of 18 use cases for RFID technology in this sector, grouped into six different categories However, even though it adopted the same methodology that we embrace here, the study by Rizzi et al. reported a static picture of RFID deployments in the fashion retail, which means that it did not comprehend time as a variable; rather, it only considered a given moment in time for its analysis (i.e. July 2015). Moreover, the authors sought to highlight the UC framework in fashion and apparel retail, that is, to understand why end users are adopting RFID technology and what main benefits they are looking for. Conversely, the goal of the current study is not only to present an updated picture of RFID adoption in fashion and apparel retail but also to: (i) provide an exhaustive state-of-the-art review of the technology to benefit researchers and practitioners, (ii) analyse the trend of RFID projects over time by analysing the timeline of UCs and project statuses and (iii) classify RFID UCs using six different labels, namely, core, intermittent, trendy, vintage, innovative or ghost, in a framework that considers three different dimensions: frequency, persistence and eldership (i.e. the state of being older).

The remainder of the paper is organised as follows: In Section 2 we provide a brief overview of the existing literature on RFID adoption in this sector to better position our paper. Section 3 explains the methodology of our literature review and classification procedure (Subsection 3.1) as well as the data collection and classification procedure (3.2) and our proposed UC lifecycle framework (3.3). We present the results of our analyses in Section 4: Subsection 4.1 reports descriptive statistics of the project database (PDB), whereas Subsection 4.2 provides the time-based analyses. Finally, Section 5 concludes the paper and suggests future research trends.

2 A brief overview of the scientific literature

RFID is an automatic identification and data capture technology that consists of a transponder (or tag), a transceiver (or reader), and middleware software. The tag stores identification and other data and sends them to a reader, which could be fixed or handheld, and it collects these data and sends them to the middleware, which is a software layer that works as a bridge between the RFID devices and other applications (e.g. legacy systems). Usually, the middleware pre-processes RFID data and stores them in a back-end database, or sends them directly to another software application (Musa & Dabo, 2016).

RFID technology is used in heterogeneous fields, ranging from healthcare to the logistic and restaurant industries (Ngai et al., 2010; Zhu, Mukhopadhyay, & Kurata, 2012). However, the fashion and apparel sector has been leading the way in RFID item-level tagging and thus in the number of RFID tags encoded by the industry (Das, 2017), accounting for a major part of the journey from the ‘Internet-of-Pallets’ to the Internet-of-Things. The adoption of RFID in the fashion and apparel sector is, in fact, well established, as it helps retailers to solve several key issues (Bottani et al., 2009).

Although RFID is a mature technology, the opportunities for growth are still interesting. While the first deployments of item-level tagging in fashion and apparel retailing were conducted in the early 2000 s (Rizzi et al., 2016), research and applications in this sector are still on-going (Hauser, Günther, Flath, & Thiesse, 2019). Indeed, the scientific literature generally agrees that the benefits provided by RFID overcome its costs, and several frameworks and models were designed and applied to confirm this point (see for example Bertolini, Romagnoli, & Weinhard, 2017; Bhattacharya, 2015; De Marco, Cagliano, Nervo, & Rafele, 2012). Also, many studies aimed at defining the best practices which should be adopted to increase the performance of an RFID system in its deployment phase (Hardgrave et al., 2013; Hardgrave, Goyal, & Aloysius, 2011). More recently, some studies addressed innovative RFID applications, such as the adaptive fitting room (Sjøbakk, Landmark, & Hübert, 2017), which allows different types of customer engagement prior to the point of purchase and opens up new possibilities for integrating product information and recommendations with social media, and automated checkout systems (Hauser et al., 2019), which promise both more sales, due to an improved customer experience, and cost savings, because less store personnel is needed for the same checkout operations.

It is opinion of the authors, however, that there is a lack of research on RFID UCs, particularly in terms of the trend of their usage over time. Interesting contributions from this point of view include the studies by Moon and Ngai (2008) and Rizzi et al. (2016). In their original study, Moon and Ngai (2008) adopted a two-stage methodological approach, involving a multi-case study, and designed a four-proposition framework for examining RFID-generated values for fashion retailers. More recently, Rizzi et al. (2016) extended this framework and built a comprehensive and structured taxonomy for RFID UCs in this sector. To do this, the authors searched different sources of information, such as websites, magazines and newspapers, in addition to scientific papers and conference proceedings.

However, as we stated in the introduction, the study of Rizzi et al. (2016) focused mainly on the description and organisation of UCs, and it depicted the status of RFID implementations at a specific moment in time (July 2015). For these reasons, we believe that more research is needed to (i) depict the current state of the art of RFID technology, (ii) understand the trend of RFID projects over time and (iii) identify the present and the likely future scenarios for RFID UCs in fashion and apparel retail.

3 Methodology

In order to perform a broad review of RFID projects in fashion retailing, a wide-ranging literature study was carried out. Based on this analysis, it was possible to define a frame of reference that allows the categorisation of possible UCs of UHF RFID technology in this field. The approach used for analysing and classifying the literature resources is very similar to that presented by Rizzi et al. (2016):

First, we selected and collected the sources of information relevant to our scope.

Then, we designed and populated two databases: one for the literature (Reference Data Base, RDB) and one for the project classification (Project Data Base, PDB).

3.1 The literature review and classification procedure

The first step of the literature review was the selection of publications relevant to our purpose. We only considered papers related to UHF RFID-technology applications in fashion and apparel retailing, excluding articles that either examined RFID retailing in general or reported on RFID applications in other product categories (e.g. consumer packaged goods, cosmetics, electronics). We started the publication selection by searching the Scopus database for scientific papers and conference proceedings, only considering those that contained ‘RFID’ as a keyword and ‘fashion’ or ‘apparel’ in either the title, abstract or keywords. Since only a small percentage of the companies that deployed an RFID project published the results in scientific literature (Esposito et al., 2015), we also considered other source of information, such as websites, news and confidential information.

To monitor and easily manage every source of information, we designed the RDB, in which each source is stored as a record containing several types of information, including:

Author’s names and paper title

Source name

Source category and subcategory. Possible categories are ‘Conferences and seminars’, ‘Confidential information’, ‘Magazines and newspapers’, ‘Research and scientific reports’, ‘Scientific papers’ and ‘Websites’, whereas the subcategory could include journal, conference or magazine name or title

Source details (volume, issue or date and publication year)

Web link

Company or companies involved in the project/projects

After entering all meaningful information for our analysis in the RDB, we organised another database, called the PDB. In the PDB, the sources are arranged at the ‘project’ level, and they are linked to one or more UCs. As in the work presented by Rizzi et al. (2016), a project could be a feasibility study, a proof of concept, a pilot, a phased deployment or a full deployment of RFID. More specifically:

A feasibility study is the research done before a proof of concept or a pilot. It starts from an as-is scenario and it architects to-be RFID processes. It is used to estimate important parameters—such as ROI (Arain, Campbell, Cooper, & Lancaster, 2010)—that are needed to successfully complete the project development and implementation.

A proof of concept is a prototype of the system, which is useful for demonstrating the feasibility or validity of some of its constituent principles or concepts.

A pilot is a version of the system that is run in miniature to test whether the components of the main study can all work together (Arain et al., 2010). It has an experimental function, and it could be viewed as a model.

A phased deployment is a re-engineering procedure during which the to-be RFID-system gradually replaces some parts (or features) of the as-is system until its eventual complete and final replacement. Typical examples of phased deployments include the ‘store-by-store’ approach, where the RFID time-phased deployment involves all of the retailer’s products in an increasing number of its stores, or the ‘product-by-product’ approach, also known as ‘model-by-model’, where the RFID deployment involves an increasing number of the retailer’s products in all of its stores;

A full deployment is a re-engineering procedure through which the old as-is system is completely replaced with the to-be system (i.e. all products, all stores).

According to this definition of project, it is possible for the same company to be associated with more than one RFID project (e.g. a pilot project followed by the full deployment). A single project can pursue one or more UCs, such as stock visibility, replenishment or inventory accuracy (the complete definition of UCs is available in the paper written by Rizzi et al., 2016).

For the sake of readability, the detailed description of the PDB is presented in Appendix A.

3.2 The data collection and classification

A search in the Scopus database for -KEY (RFID) AND TITLE-ABS-KEY (fashion OR apparel)- was performed on June 2018 and produced 189 documents, although only a small fraction was consistent with our scope due to the lack of project information and company details in most of the studies. We removed these studies, resulting in a total of 38 peer reviewed papers and 18 publications in conference and seminar proceedings. Given that several companies that deployed an RFID project published results in the scientific literature, we also collected information from other sources, such as websites, news and confidential information. At the end of this selection process, our RDB comprised 149 different projects carried out by 97 different companies. The information related to these projects was retrieved from 224 different sources, as shown in Table 1. Due to space constraints, the full list of reviewed references is reported and organised in a table in Appendix B. Please note that we omitted from the above-mentioned table all confidential information (e.g. information protected by non-disclosure agreements).

Table 1
Number and percentage of sources in the Reference Data Base

Type of source # of contributions % of contributions

Websites 147 65.6%

Scientific papers 38 17.0%

Conferences and seminars 18 8.0%

Confidential information 13 5.8%

Research & Scientific reports 5 2.2%

Magazines and newspapers 3 1.4%

Total 224 100.0%

Type of source	# of contributions	% of contributions
Websites	147	65.6%
Scientific papers	38	17.0%
Conferences and seminars	18	8.0%
Confidential information	13	5.8%
Research & Scientific reports	5	2.2%
Magazines and newspapers	3	1.4%
Total	224	100.0%

Most of the information used in this study was obtained from websites: in fact, websites accounted for about the 66% of the reviewed documents (a total of 147 websites were considered). As Table 2 shows, most of those articles were published by RFID Journal (47 articles; http://www.rfidjournal.com) and RFID 24-7 (34 articles; http://rfid24-7.com/). We also included 21 articles from company websites (e.g. fashion firms, RFID technology suppliers and system integrators) and 45 articles from thematic websites (e.g. technology or fashion industry blogs).

Table 2

Number and percentage of websites in the Reference Data Base

Website category/name	# of contributions	% of contributions
RFID Journal	47	32.0%
Thematic sites and others	45	30.6%
RFID24-7	34	23.1%
Company websites	21	14.3%
Total	147	100.0%

The second significant source of information was scientific papers, with a total of 38 papers considered (17% of the reviewed documents). The majority of the papers were from the International Journal of RF Technologies, Research and Application, with a total of nine papers considered (about 24% of contributions in the category of scientific papers). The second most frequent source of scientific papers was the Information Technology Research Institute (3 papers). In addition to these two journals, 21 others accounted for one or two of the scientific papers that were included in the RDB.

The third source of information for our RDB was conference and seminar proceedings, with a total of 18 papers considered (8% of total contributions).

The other sources of information that were considered included:

unpublished data (i.e. information confidentially reported and derived from heterogeneous sources; 5.8% of documents reviewed)

research and scientific reports (2.2%)

magazines and newspapers (1.4%)

After the classification of the information sources by category, we categorised papers based on the year of publication. As shown in Fig. 1, the number of contributions in the last decade has significantly increased, and most of the studies were published between 2012 and 2018. From Fig. 1, it is possible to identify four three-year cycles, with relatively constant values of contributions per year:

Fig.1

Number of contributions per type of source and year of publication.

2006 to 2008, with an average number of 5 contributions per year

2009 to 2011, with exactly 15 contributions per year

2012 to 2014, with an average of 26 contributions per year

2015 to 2017, with more dispersed data and an average of about 19 contributions per year

A total of 18 contributions were reviewed in the first half of 2018 (when we froze the data presented below), encompassing websites, scientific papers, press releases and confidential information, and this number seems to confirm an increasing momentum for RFID in the fashion and apparel industry. Finally, if we exclude the sources that are not linked to a publication year, it is possible to notice that:

94% of the contributions were published in the last decade (2008–2018)

more than 70% were published between 2012 and July 2018

more than 46% were published in the last five years (2014–2018).

3.3 The use case lifecycle framework

With the aim of understanding the lifecycle of RFID UCs, we adapted the study by Fadlalla and Amani (2015) to the adoption of RFID in the fashion and apparel retail sector. The framework is based on an analysis of three dimensions, persistence, frequency and eldership, all of which are time-based measures calculated within our timeframe. The starting year of the analysis is that of the first item-level RFID project in retail (2001), and the last year of the analysis can be selected within the given timeframe (at the time of writing, the data are updated as of mid-2018). As we will see in Section 4.2, we used two different periods for our analyses. Detailed descriptions of our dimensions follow:

Persistence is related to the continuity of the UC over time, and it has been calculated as the number of different years that witnessed the deployment of at least one RFID project tackling a specific UC.

Frequency, on the other hand, gives us information about how many times a specific UC has been investigated in the PDB. Frequency (known as dominance in Fadlalla and Amani, 2015) indicates how much attention the industry has given to a specific UC, while persistence adds the temporal continuity concept.

Eldership refers to the number of years that have passed since the first time a specific UC has been tackled, that is, how old a specific UC is, considering its first deployment.

According to these three dimensions, we identified six different UC labels, as shown in Fig. 2, which we named as follows:

Core – high persistence and high frequency, i.e. UCs that were investigated many times and in several different years;

Intermittent – high persistence but low frequency, i.e. UCs that were investigated with less than the median frequency but in several different years;

Trendy or Vintage – low persistence with high frequency, i.e. UCs investigated above the median in a limited number of different years. According to the value of Eldership, these UCs may be divided into Vintage (i.e. investigated many times in a limited period in the past) or Trendy (i.e. investigated many times in a recent and limited period);

Innovative or Ghosts – low persistence and low frequency. Again, the separation between innovative and ghosts can be made based on eldership. In fact, depending on eldership, a specific UC, falling into the low persistence/low frequency quadrant, can either be identified as Innovative (in case it has low eldership, that is, ‘recent’) or as Ghost (with a high level of eldership, meaning that it has been a long time since the UC was first investigated).

Fig.2

The structure of the UCs lifecycle framework.

4 Results and discussion

As we described in Section 3, all the sources we reviewed, as well as the information they contain, were structured in the RDB and PDB, respectively. Eventually, the PDB comprised:

149 records (i.e. projects)

data collected from 2001 to mid-2018

97 different companies

23,400 stores deployed overall

over 1 billion tags/year overall

89 system integrators, hardware and tag vendors cited overall.

Due to the nature of this study, we chose not to report any company or brand name in any section. Thanks to the PDB, it was possible to analyse data and to report statistics and, consequently, to frame the deployment and use of RFID technology in fashion retail. In this study, we consider time as an important variable, and we identified two different categories of indicators that allow us to underline different aspects of the UC and RFID implementation: a descriptive and a time-based analysis.

As for the descriptive analysis, we mainly focused on the following:

Project status

Tag brand and type

Geography

Benefits achieved with the RFID implementation

With regard to the time-based analysis, we made considered the following:

Project status timeline

UCs analysis (persistence/eldership/frequency)

4.1 Descriptive analyses

4.1.1 Project status

As Table 3 reports, the incidence of pilot projects is predominant in the database, followed by phased deployments and proofs of concept.

Table 3
Distribution of project status

Project status Percentage

Pilot 32.4%

Phased deployment 27.6%

Proof of concept 22.1%

Full deployment 13.8%

Feasibility study 4.1%

Project status	Percentage
Pilot	32.4%
Phased deployment	27.6%
Proof of concept	22.1%
Full deployment	13.8%
Feasibility study	4.1%

As ‘models’ and preliminary tests for RFID implementation, pilots and proofs of concept can identify the benefits of the technology for a retailer, and they are crucial for moving toward adoption. Furthermore, the predominance of phased deployments indicates that the technology is still climbing towards the so-called ‘plateau of productivity’ (O’Leary, 2008).

4.1.2 Tags (tag type and brands)

In our PDB, we specified the types of RFID tags used in each project. That information was stored in the appropriate field, and, in the case of multiple RFID tags types tested within the same project, we saved the information related to each of them. The different tag types identified during this research have been listed in Appendix A (for more details on the types of tags considered in this study, see the description provided by Bertolini, Rizzi, Romagnoli, & Volpi, 2017). Eventually, the PDB contained 94 records with tag type information:

price label or hang tags : The RFID tag is hung on the item price tag.

care label : The RFID tag is included in the item care label. The care label is a woven label that is attached to every textile product.

sewn-in tags : The RFID tag is sewn directly inside the item. Usually these tags are ultra-thin and soft, so they can be easily sewn into different textiles. Microtags can also be considered as sewn-in tags.

hard tag : The RFID tag is inside the EAS (electronic article surveillance) tag and is removed at the point of sale.

The percentage of different tag types in different projects is reported in Table 4. As the table shows, the price label is the most common solution adopted, and it is used in more than 59% of the total projects. Indeed, the price label is the easiest and quickest choice, as it is a cheap, less invasive, ‘one fits all’ type of tag. Moreover, it enables most of the UCs, at least the core ones, as we will see below. The price label tag can be easily removed at the register desk, thus avoiding any privacy issues. On the contrary, price labels do not enable after sales UCs, and they have a negative impact on EAS protection. For these UCs, different tagging solutions may be chosen (e.g. care labels, hard tags or sewn-in tags).

Table 4
Percentage of projects per tag type

Tag type Percentage

Price label / hang tag 59.1%

Care label 6.0%

Hard tag 6.0%

Sewn-in tag 5.4%

Unknown 23.5%

Tag type	Percentage
Price label / hang tag	59.1%
Care label	6.0%
Hard tag	6.0%
Sewn-in tag	5.4%
Unknown	23.5%

Concerning tag brands, 31 records in our PDB indicated the additional detail of the tag brand, including 10 different tag-manufacturing companies.

4.1.3 Geography

The geographical information retrieved from the database shows, as expected, a widespread implementation of RFID technology globally.

We organised geographical clusters as follows (see Figs. 3 and 4):

Fig.3

Percentage of projects per geographical cluster.

Fig.4

Percentage of stores with an RFID deployment per geographical cluster.

Europe

USA

Rest of the world (meaning all countries other than USA and Europe)

Worldwide (concerning projects that involved more than one country across the three categories previously mentioned).

Figure 3 shows that Europe is currently leading in RFID adoption, accounting for 49% of the total RFID implementations, followed by the US, with 22%. This result may be biased by the informal sources of information we received from local end users and integrators. We also introduced ‘stores number’ as an output variable. This information is available for 87% of the projects (i.e. a total of 129 projects out of 149). The results are reported in Fig. 4.

Again, we recognise the predominance of RFID adoption in Europe but, at the same time, there has been significant growth in the ‘Worldwide’ percentage and a corresponding reduction in the ‘Rest of the world’ percentage, as shown in Fig. 3 (which does not include the number of stores). This shift can be explained simply if we consider that most of the retailers included in the worldwide basket are those that are rolling out – or that have already completed – phased or full RFID deployments, which involve multiple stores around the world (e.g. a well-known retailer completed a worldwide full deployment in 2018 in 3,000 stores).

4.1.4 Benefits achieved through RFID implementation

The PDB tracks information about the main benefits achieved with the RFID project. Table 5 reports the number of occurrences of each result in the PDB, partitioned as qualitative and quantitative results. By qualitative results, we mean a project that explicitly reports positive benefits related to one of the results we considered (see the last part of Appendix A for more details) but does not provide numbers to support the claim (e.g. the RFID implementation has allowed us to reduce the total inventory time of the store). If the results are supported by numerical values, we considered them as quantitative results. We are aware that, in some cases, the number of occurrences for a specific result does not provide a wide sample for inferential statistics (e.g. R2 cancellation rate reduction). However, we chose to report descriptive statistics to provide the reader with some figures related to results that could become more consistent in the future. Further, it should be considered that some results are due to emerging UCs, and thus the number of companies that could possibly achieve them is consequently limited.

Table 5
Number of occurrences (occur.) per each result, categorised into qualitative and quantitative ones

Result No. of occur. No. of qualitative occur. (percentages in brackets) No. of quantitative occur. (percentages in brackets)

R1-Process time reduction 46 38 (83%) 8 (17%)

R2-Cancellation rate reduction 3 0 (0%) 3 (100%)

R3-Process accuracy increase 23 21 (91%) 2 (9%)

R4-Turnover / Sales increase 57 35 (61%) 22 (39%)

R5-Inventory time reduction 61 41 (67%) 20 (33%)

R6-Inventory accuracy 74 32 (43%) 38 (57%)

R7-Shrinkage reduction 23 19 (83%) 4 (17%)

Result	No. of occur.	No. of qualitative occur. (percentages in brackets)	No. of quantitative occur. (percentages in brackets)
R1-Process time reduction	46	38 (83%)	8 (17%)
R2-Cancellation rate reduction	3	0 (0%)	3 (100%)
R3-Process accuracy increase	23	21 (91%)	2 (9%)
R4-Turnover / Sales increase	57	35 (61%)	22 (39%)
R5-Inventory time reduction	61	41 (67%)	20 (33%)
R6-Inventory accuracy	74	32 (43%)	38 (57%)
R7-Shrinkage reduction	23	19 (83%)	4 (17%)

Figure 5 shows the box and whiskers plots for the quantitative results of RFID implementations.

Fig.5

Box-and-whiskers plots of consequences of RFID implementations. R1-Process time reduction; R2-Cancellation rate reduction; R3-Process accuracy increase; R4-Turnover / Sales increase; R5-Inventory time reduction; R6a-Inventory accuracy; R6b-Inventory accuracy increase; R7-Shrinkage reduction.

A total of 46 projects reported benefits in process time reduction, and 8 of them specified the exact percentage of reduction, which were extremely dispersed (3–95%).

Only 3 projects reported the percentage reduction of cancellation rates of online orders, which ranged between 30 and 60%.

The (non-inventory) process accuracy was increased in 23 projects, although only 2 of them reported the exact percentage of increase (7–9%).

Moreover, 57 projects reported increased turnover/sales due to RFID implementation, with exact percentages reported by 22 projects and ranging between 1 and 21% (the latter being quite impressive).

Similarly, 61 projects reported a reduction in inventory count times after RFID implementation, with the reductions (provided by 20 projects) ranging from 50 to 100% (i.e. when traditional inventory counts were not carried out any longer after RFID implementation).

The matter is more complex if we consider inventory accuracy. In fact, some projects reported incremental results related to inventory accuracy (e.g. baseline + x%), whereas others reported the absolute values of inventory accuracy after the RFID implementation. Altogether, 74 projects reported an increase in inventory accuracy: 32 of them only indicated it qualitatively, 34 of them provided the resulting inventory accuracy values (95–100%) and 8 reported an increase in inventory accuracy (ranging from 7–50%).

Finally, 23 projects reported a reduction of shrinkage, and 4 of them provided percentages, which ranged between 9 and 55%.

4.2 Time-based analysis

As previously discussed, we based our analysis on the UC framework presented by Rizzi et al. (2016). That study presented a total of 18 specific UCs, corresponding to 18 different objectives pursued through RFID implementation.

4.2.1 UCs and project status timeline

First, we built a timeline that, year by year, highlights the number of feasibility studies, proofs of concept, pilots, phased and full deployments that every UC has had in the 2001–2018 period. This timeline is reported in Fig. 6, which, due to space constraints, only shows three screenshots of the timeline (2008, 2013 and 2018 data, respectively).

Fig.6

UCs and project status timeline: 2008 data (6A), 2013 data (6B), 2018 data (6C).

As Fig. 6 shows, 4 UCs were still ‘unknown’ to the industry in 2008. The data were less dispersed, as the gap between minimum and maximum frequency was just 16, and no full deployments were available at that time but rather only a few phased deployments. Just five years later, a handful of full deployments started to appear, all UCs had at least one project and predominant UCs started to emerge from the average (56 projects with 4.1-Process automation). If we focus the description on 2018 data, the figure shows that UC 5.1-Out of stock/inventory accuracy was the most popular UC, targeted in more than 60% of the projects (95 out of 149). The objective of this UC is to avoid the situation where one or more SKUs are out of stock (OOS), both on the sales floor and in the backroom.

This UC is particularly suited for replenishable products (i.e. products that may be replenished from the distribution centre during the season). For example, consider the case of a new model that has sold out sooner than forecasted during a season. Since demand for that product has been higher than expected, it is likely that the model will sell again if the store is replenished during the same season. RFID deeply impacts inventory inaccuracy, both the overstated and understated inventory, and it can reduce or even avoid the so called ‘frozen OOS’ (Hardgrave, 2009). Frozen OOS is defined as the situation that occurs when an SKU is OOS at the store level, both on the sales floor and in the backroom, but it is not replenished from the distribution centre because of record inaccuracies between real inventory (0) and overstated ‘on hand’ inventory in the information systems (typically above the reorder point). A 2013 study indicated that the average inventory accuracy of traditional stores is in the range of 70% (Hardgrave et al., 2013), potentially increasing to 99.9% (Esposito et al., 2015).

The second most popular UC is UC 4.1-Process automation, which is linked to labour cost reduction. In fact, RFID UCs were classified as ‘Process automation’ only when the technology was employed to automate existing processes (i.e. processes that were also in place before the RFID implementation). Thanks to the introduction of RFID technology, processes can be fully automated and/or carried out at higher speed, cutting down related labour costs. Specifically, packing, shipping and receiving processes typically benefit from RFID deployment.

The third most implemented UC is 1.4-Stock visibility/Replenishment from the backroom. This UC is strictly connected to item availability on the sales floor. Different studies have analysed customer response to stock-outs, both theoretically and practically (see e.g. Aastrup & Kotzab, 2010), highlighting the importance of knowing which SKUs have to be replenished from the backroom as well as which models/colours/sizes have never been taken to the sales floor. In general, ‘results indicate that RFID-based policies have the potential to improve cost efficiency and service levels’ (Condea, Thiesse, & Fleisch, 2012).

On the other side, the least popular UCs (i.e. those investigated the minimum number of times) are 2.1-Social shopping, 6.1-Grey market fighting, 6.2-Anti-counterfeiting and 3.3-Store associate empowerment. All these UCs were pursued in less than 5% of the projects in our PDB. However, the frequency of occurrence of each UC, alone, could not discriminate between emerging UCs (e.g. due to technological improvements, the UC was only made available recently), intermittent UCs (e.g. the UC was tested in at least two different time periods to ascertain its validity) and ghost UCs (e.g. the UC was pursued in the past but then not tested anymore). For this reason, we considered the time period of the projects and developed the following framework.

4.2.2 The UC lifecycle framework

The UC lifecycle framework that we proposed in Section 3.3 makes it possible to answer the following questions:

What are, at present, the core use cases for RFID adoption in fashion and apparel retail?

What changes, if any, occurred in the last three years? (we refer to the study of Rizzi at al. 2016, publishing data up to 2015)

What are the likely future scenarios, due to promising UCs that can become core in the near future?

According to the scaffold reported in Fig. 2, UCs were represented in the UC framework both in Fig. 7 (2001–2015 data) and in Fig. 8 (2001–2018 data). We note that, in addition to the axes (Frequency and Persistence, in the blue part of Fig. 2), the eldership of UCs is highlighted by the saturation of the colours of different units in the graph – the darker the bubble, the older the first project initiation of that specific UC (i.e. eldership).

Fig.7

Representation of UCs in the UC lifecycle framework, 2001–2015 data.

Fig.8

Representation of UCs in the UC lifecycle framework, 2001–2018 data.

The ‘should be’ lifecycle of a UC begins in the third quadrant (e.g. Innovative) and then develops into each of the other three quadrants. For instance, Trendy UCs were pursued in several different projects in a limited period, but they were abandoned with no real continuity over time (thus becoming Vintage) or continuously investigated (thus becoming Core). Core UCs are present over the years and have been pursued many times. If the UCs become intermittent, the continuity over time is a matter of fact, but the lack of implementations suggests that one or more further steps are needed to clarify the results and benefits to be expected from the adoption of that specific UC.

Figure 7 shows the UC lifecycle framework in 2015, in which UC 5.1-Out of stock/inventory accuracy, 4.1-Process automation and 1.4-Stock Visibility/Replenishment from the Backroom were the most investigated UCs. According to Hardgrave (2012), ‘RFID technology is the only efficient way to achieve in-store inventory accuracy. The case for managing inventory with RFID is even stronger when you extend it beyond the store, to the need for visibility in distribution centres and the supply chain’. Fashion products have a limited lifecycle at their full prices; additionally, it is well known that customers only buy what they see. Several studies have explored customer response to stock-outs, both theoretically and practically (see e.g. Campo, Gijsbrechts, & Nisol, 2000; Zinn & Liu, 2008). Therefore, it is no surprise that UCs connected to shop floor management and inventory and supply chain management are the most frequent in our PDB.

In detail, UC 5.1-Out of stock/inventory accuracy, 4.1-Process automation and 1.4-Stock visibility/Replenishment from the backroom are all currently labelled as Core, as they were highlighted in 2015, being continuously and several times investigated UCs. Indeed, OOS and out-of-shelf situations (the latter being the condition where a product is in the backroom but not on the sales floor area) are key problems that retailers face. RFID makes it possible to have items on the shelves, at the customer’s disposal and at the right time, and therefore positively impacts sales. A significant number of studies agree that RFID technology has a significant impact on OOS reduction (see e.g. Bertolini et al., 2017; Hardgrave, Riemenschneider, & Armstrong, 2008; Hardgrave, Waller, & Miller, 2006).

Similarly, the literature also agrees on the fact that one of the main RFID UCs is automating processes; this was indeed the oldest pilot in our PDB for item-level RFID deployments (dating back to 2003).

Table 6 reports full details of UC lifecycle analysis.

Table 6

UCs reported per each label, 2018 data

Ghost	Intermittent	Vintage
2.2-Store associate availability / Customer knowledge	1.1-Locating items	- none -
3.2-Cross selling / Cross promotions
3.3-Store associate empowerment
4.3-After sales / Returns
Innovative	Trendy	Core
2.1-Social shopping	5.2-Omnichanneling	1.2-Loss prevention
6.1-Grey market		1.3-POS transaction / Faster checkout
6.2-Counterfeiting		1.4-Stock visibility /
		Replenishment from the backroom
6.3-Traceability		3.1-Customer experience
		4.1-Process automation
		4.2-Process accuracy
		5.1-Out-of-stock / Inventory accuracy
		5.3-Supply chain visibility

If we compare the two analyses (2015 vs. 2018 data), reported in Figs. 7 and 8, respectively, we acknowledge a change of label for UC 3.2-Cross selling/Cross promotions, which shifted from trendy (2015) to ghost (2018), and UC 5.2-Omnichanneling, which shifted from innovative (2015) to trendy (2018).

The development of UC 5.2-Omnichallening is strictly connected to the increase of the number of projects pursuing that UC, meaning that end users are gaining interest in the topic. This is not a surprise, as many traditional retailers are expanding their online businesses to compete with purely online ones. In addition, consumer behaviour has been rapidly changing in the last years, leading companies to consider combined business-to-consumer retail channels and innovative instruments to enhance the classic shopping experience (Lazaris & Vrechopoulos, 2014). These solutions not only fall into ‘marketing’ strategies but are also connected to the reengineering of processes (e.g. in-store visibility technologies, e-commerce websites, augmented reality). As Caro and Sadr (2019) suggest, the functions of sales channels are connected to bringing customers information and products.

Through omnichannel marketing, consumers face the same customer experience either physically in the store or virtually on a laptop/smartphone. They can choose from the same assortment and inventory, purchase items and decide whether they want the goods to be delivered (either to the same store or to another one or their own front door).

Thanks to RFID, retailers can fully leverage online sales due to increased inventory accuracy levels. When inventory accuracy is close to 100%, all available items can be confidently listed for sale online, minimising the risk of overpromising and offering products that are not available to customers. This is a pivotal issue for online retailing, as unsatisfied customers not only stop buying from the retailer but can also spread negative comments through social networks. This is why many non-RFID retailers adopt high inventory levels as a cutover for online sales. In this way, they hope to have enough inventory to cover orders in case of overestimated values in their legacy systems. Nevertheless, non-RFID omnichannel retailers lose opportunities and are burdened by opportunity costs that RFID retailers do not have.

Roberti (2019) addresses this issue:

There is an extraordinary amount of complexity in an omnichannel supply chain that just doesn’t exist in the old brick-and-mortar model. How will retailers know where items are located and what is in inventory in every area of their business if they don’t use RFID? It’s possible to get by without RFID, as companies can hold more safety stocks and hide inventory from customers. Many retailers that currently offer BOPIS (buy online pickup in store will not show an online customer a product if only two or three units are left at that shopper’s local store. That’s because they know their inventory numbers are not accurate, and it’s possible they might be out of stock. In their view, it’s better to not sell an item than to sell it and have the customer arrive at the store only to find the product has sold out. Not showing consumers items reduces sales, and holding larger safety stocks increases costs and ties up capital. The retailers that will win in a truly omnichannel world are those that use RFID and other technologies to gain an accurate view of their inventory in real time, and can thus make every product available to every customer, no matter how a shopper chooses to buy that item.

5 Conclusions

The retail sector, and particularly fashion and apparel retail, is the main industry leading the adoption of item-level UHF RFID. This is mainly due to the several UCs for RFID in the industry (no other sector has so many) as well as the high product values, complex product flows in the supply chain (i.e. high number of SKUs to manage), limited product lifespans and the optimal characteristics of products from a physical perspective.

In 2016, Rizzi et al. reported a static picture of UCs and deployments of RFID technology in fashion and apparel retail. The present study takes a further step and analyses the trends over time. Moreover, our study depicts a portrait of the current players, the status and characteristics of RFID projects and the benefits achieved by apparel and fashion retailers. We would like to be clear about a potential bias of our study. We based our research upon published sources on UHF RFID implementation in fashion and apparel retail. Of course, this is just a subset of actual implementations of this technology. Even so, we believe that, given the number of studies we reported, our results could be significant for the whole sector.

As far as the project status is concerned, we report significant numbers of both full and phase deployments as well as new pilots, suggesting that the use of RFID technology is still emerging, entering the so-called ‘early majority’ phase. Pilot projects are likely to become phased deployments, and these are the starting point for full deployments, which still account for less than 15% of the projects in our PDB.

Concerning the tag type, we found a clear predominance of hang tags, which is often the solution of choice due to the low-price, one-size-fits-all, nonintrusive features of price tags. Moreover, the solution enables most of the UCs (the only notable exceptions being UC 4.3-After sales/Returns, 6.2-Counterfeiting and 1.2-Loss prevention). Other tag types (e.g. care labels, hard and sewn-in tags) have seen limited use at the item level (5–6% of projects).

From a geographical perspective, Europe (mostly Italy and Germany) is leading in the number of RFID projects, followed by the USA. This gap is reduced if we compare projects in terms of the number of stores with deployments instead of the number of projects in total. Furthermore, we noted a significant increase in terms of the number of worldwide deployments and the number of stores.

The most common benefits of RFID deployments are inventory accuracy (increasing from 60–70% on average to 99–100% following RFID deployment), reduction of inventory count time (–90%) and increased turnover and sales (after the deployment of RFID, fashion and apparel retailers pursuing this UC reported turnover increases from 1–21%).

RFID technology also results in benefits related to process time reduction (up to 95%), increased process accuracy (from 7 to 9%) and reduced inventory shrinkage (from 9 to 55%). As omnichannel retailing has become increasingly popular, some retailers have also reported a reduction in the cancellation rates of online orders (from 30 to 60%).

Moving to the time-based analyses, and focusing on 2018 data, we report that the UCs that have been investigated the most are 5.1-Out of stock/Inventory accuracy, 4.1-Process automation and 1.4-Stock visibility/Replenishment from the backroom, followed by 4.2-Process accuracy and 1.2-Loss prevention. Accordingly, all these UCs were labelled as Core, according to the UC lifecycle framework that we developed, along with 1.3-POS transaction/Faster checkout, 3.1-Customer experience and 5.3-Supply chain visibility.

UC 5.2-Omnichanneling was labelled as Trendy in our framework, as it emerged more recently (persistence is lower than the median value) but has still been pursued in several projects (more than the median). We may thus expect that this UC will become Core in the near future, as more and more firms deploy RFID to enable omnichannel strategies.

UC 1.1-Locating items was labelled as Intermittent, as this UC was continuously pursued during the time period of our analysis, but the number of projects targeting it remained below the median.

A handful of UCs were labelled as Ghost, namely, UC 2.2-Store associate availability/Customer knowledge, 3.2-Cross selling/Cross promotions, 3.3-Store associate empowerment and 4.3-After sales/Returns. Ghost UCs were investigated at a rate lower than the median, both in terms of number and in terms of the different years of the projects. In addition, Ghost UCs showed high values of eldership, meaning that they are older than the median value. On the contrary, Innovative UCs are relatively similar to Ghost UCs in terms of frequency and persistence, but they are younger (eldership below the median), and thus it may be expected that there will be an increasing number of these deployments in the near future. The full list of Innovative UCs is reported in Table 6.

We believe that the main limitation of our study is related to the sample we analysed. Not all of the companies that assessed or implemented RFID disclosed their project activities, the UCs pursued, or the results achieved. However, we believe that the records we collected represent a consistent sample that can provide useful insights both for researchers, by suggesting specific areas for deeper research, and for practitioners, by providing insights about who has been deploying RFID, where, why and how as well as the qualitative and quantitative benefits that can be expected. We also believe that our up-to-date framework of RFID deployments, based on the persistence, frequency and eldership of UCs, can provide a clear picture of UC evolution over time, thereby improving our understanding of which UCs were, are and will be Core, Vintage, Trendy, Intermittent, Innovative and Ghost. It is the intention of the authors to continue to update the current results, in order to appraise the evolution over time of the UCs timeline and lifecycle framework.

Footnotes

The Appendix section is available in the electronic version of this article: .

References

Aastrup,

, Kotzab,

(2010). Forty years of Out-of-Stock research – and shelves are still empty, The International Review of Retail, Distribution and Consumer Research, 20(1), 147–164. DOI: 10.1080/09593960903498284

Arain,

, Campbell,

M. J.

, Cooper,

C. L.

, Lancaster,

G. A.

(2010). What is a pilot or feasibility study? A review of current practice and editorial policy, BMC Medical Research Methodology, 10(67), 1–7.

Bertolini,

, Maggiali,

, Rizzi,

, Romagnoli,

, Volpi,

(2017). Introducing new RFID-enabled indicators to evaluate the performance of fashion retailers. In Proceedings of the Summer School Francesco Turco, Industrial Systems Engineering, Palermo, Italy, pp. 54–60.

Bertolini,

, Rizzi,

, Romagnoli,

, Volpi,

(2017). Testing an RFID receiving gate for improving process accuracy in fashion and apparel retail, In RTSI 2017 - IEEE 3rd International Forum on Research and Technologies for Society and Industry, Conference Proceedings. DOI: 10.1109/RTSI.2017.8065916

Bertolini,

, Romagnoli,

, Weinhard,

(2017). Proposing a Value-Added Indicators framework for the apparel and fashion sector: Design and empirical evaluation, International Journal of RF Technologies: Research and Applications, 8(3), 143–164. DOI: 10.3233/RFT-171674

Bhattacharya,

(2015). A conceptual framework of RFID adoption in retail using Rogers stage model, Business Process Management Journal, 21(3), 517–540. DOI: 10.1108/BPMJ-06-2014-0047

Bottani,

, Ferretti,

, Montanari,

, Rizzi,

(2009). The impact of RFID technology on logistics processes of the fashion industry supply chain, International Journal of RF Technologies: Research and Applications, 1(4), 225–252.

Bottani,

, Montanari,

, Romagnoli,

(2016). Improving sales turnover in fashion retailing by means of an RFID-based replenishment policy, International Journal of RF Technologies: Research and Applications, 7(1), 65–86. DOI: 10.3233/RFT-150072

Campo,

, Gijsbrechts,

, Nisol,

(2000). Towards understanding consumer response to stock-outs, Journal of Retailing, 76(2), 219–242.

10.

Caro,

, Sadr,

(2019). Internet of Things (IoT) in retail: Bridging supply and demand, Business Horizons, 62(1), 47–54. DOI: 10.1016/j.bushor.2018.08.002

11.

Condea,

, Thiesse,

, Fleisch,

(2012). RFID-enabled shelf replenishment with backroom monitoring in retail stores, Decision Support Systems, 52(4), 839–849. DOI: 10.1016/j.dss.2011.11.018

12.

Das,

(2017). RFID Forecasts, Players and Opportunities 2017-2027.

13.

De Marco,

, Cagliano,

A. C.

, Nervo,

M. L.

, Rafele,

(2012). Using System Dynamics to assess the impact of RFID technology on retail operations, International Journal of Production Economics, 135(1), 333–344. DOI: 10.1016/j.ijpe.2011.08.009

14.

Esposito,

, Romagnoli,

, Sandri,

, Villani,

(2015). Deploying RFID in the fashion and apparel sector: An “ in the field ” analysis to understand where the technology is going to. In Proceedings of the Summer School Francesco Turco, Industrial Systems Engineering, Naples, Italy, pp. 54–60.

15.

Fadlalla,

, Amani,

(2015). A keyword-based organizing framework for ERP intellectual contributions, Journal of Enterprise Information Management, 28(5), 637–657. DOI: 10.1108/17410391011008888

16.

Gaukler,

G. M.

(2010). Preventing avoidable stockouts: The impact of item-level RFID in retail, Journal of Business & Industrial Marketing, 25(8), 572–581. DOI: 10.1108/08858621011088301

17.

Gaukler,

G. M.

, Seifert,

R. W.

, Hausman,

W. H.

(2007). Item-level RFID in the retail supply Chain, Production and Operations Management 65–76. DOI: 10.1111/j.1937-5956.2007.tb00166.x

18.

Hardgrave,

B. C.

(2009). Item-Level RFID for Apparel: The Dillard’s RFID Initiative, Information Technology Research Institute, 2009.

19.

Hardgrave,

B. C.

(2012). Omnichannel retailing - You can’t do it without RFID. RFID Journal, Retrieved July 18, 2014, from http://www.rfidjournal.com/articles/view?10247

20.

Hardgrave,

B. C.

, Aloysius,

J. A.

, Goyal,

(2013). RFID-enabled visibility and retail inventory record inaccuracy: Experiments in the field, Production and Operations Management, 22(4), 843–856. DOI: 10.1111/poms.12010

21.

Hardgrave,

B. C.

, Goyal,

, Aloysius,

J. a.

(2011). Improving inventory management in the retail store: The effectiveness of RFID tagging across product categories, Operations Management Research, 4(1–2), 6–13. DOI: 10.1007/s12063-011-0049-3

22.

Hardgrave,

B. C.

, Riemenschneider,

C. K.

, Armstrong,

D. J.

(2008). Making the business case for RFID. In Haasis

H.-D.

(Ed.), Dynamics in Logistics, pp. 25–35. Springer Berlin / Heidelberg. DOI: 10.1007/978-3-540-76862-3

23.

Hardgrave,

B. C.

, Waller,

, Miller,

(2006). RFID’s impact on Out of Stocks: A sales velocity analysis. Information Technology Research Institute, (Vol. Jun 2006).

24.

Hauser,

, Günther,

S. A.

, Flath,

C. M.

, Thiesse,

(2019). Towards digital transformation in fashion retailing: A design-oriented is research study of automated checkout systems, Information Systems Engineering, 61(1), 51–66. DOI: 10.1007/s12599-018-0566-9

25.

Lazaris,

, Vrechopoulos,

(2014). From multi-channel to “omnichannel” retailing: Review of the literature and calls for research. In 2nd International Conference on Contemporary Marketing Issues ICCMI.

26.

Moon,

K. L.

, Ngai,

E. W. T.

(2008). The adoption of RFID in fashion retailing: A business value-added framework, Data Systems, 108(5), 596–612. DOI: 10.1108/02635570810876732

27.

Musa,

, Dabo,

A. A. A.

(2016). A review of RFID in supply chain management: 2000-2015, Global Journal of Flexible Systems Management, 17(2), 189–228. DOI: 10.1007/s40171-016-0136-2

28.

Ngai,

E. W. T.

, Moon,

K. K. L.

, Riggins,

F. J.

, Yi,

C. Y.

(2008). RFID research: An academic literature review (1995–2005) and future research directions, International Journal of Production Economics, 112(2), 510–520. DOI: 10.1016/j.ijpe.2007.05.004

29.

Ngai,

E. W. T.

, To,

C. K. M.

, Moon,

K. K. L.

, Chan,

L. K.

, Yeung,

P. K. W.

, Lee,

M. C. M.

(2010). RFID systems implementation: A comprehensive framework and a case study, International Journal of Production Research, 48(9), 2583–2612. DOI: 10.1080/00207540903564942

30.

O’Leary,

D. E.

(2008). Gartner’s hype cycle and information system research issues, International Journal of Accounting Information Systems, 9(4), 240–252. DOI: 10.1016/j.accinf.2008.09.001

31.

Rizzi,

, Romagnoli,

, Thiesse,

(2016). A new framework for RFID use cases in fashion and apparel retailing, International Journal of RF Technologies: Research and Applications, 7(2–3), 105–129.

32.

Roberti,

(2019). The state of omnichannel retailing. RFID Journal, Retrieved March 21, 2019, from https://www.rfidjournal.com/articles/view?18205

33.

Sjøbakk,

, Landmark,

A. D.

, Hübert,

H. P.

(2017). The adaptive fitting Room, Lecture Notes in Electrical Engineering, 413, 79–88. DOI: 10.1007/978-3-319-48511-9_7

34.

Uckelmann,

, Romagnoli,

(2016). RF-based Locating of Mobile Objects. In 6th International Conference on the Internet of Things (IoT‘16) Stuttgart, Germany, pp. 147–154.

35.

Wu,

N. C.

, Nystrom,

M. A.

, Lin,

T. R.

, Yu,

H. C.

(2006). Challenges to global RFID adoption, Technovation, 26(12), 1317–1323. DOI: 10.1016/j.technovation.2005.08.012

36.

Zhu,

, Mukhopadhyay,

S. K.

, Kurata,

(2012). A review of RFID technology and its managerial applications in different industries, Journal of Engineering and Technology Management, 29(1), 152–167. DOI: 10.1016/j.jengtecman.2011.09.011

37.

Zinn,

, Liu,

P. C.

(2008). A comparison actual and intended consumer behaviour in response to retail stockouts, Journal of Business Logistics, 29(2), 141–159. DOI: 10.1002/j.2158-1592.2008.tb00090.x

State of the art of item-level RFID deployments in fashion and apparel retail

Abstract

Keywords

1 Introduction

2 A brief overview of the scientific literature

3 Methodology

3.1 The literature review and classification procedure

3.2 The data collection and classification

4.1 Descriptive analyses

4.1.1 Project status

Table 3 Distribution of project status Project status Percentage Pilot 32.4% Phased deployment 27.6% Proof of concept 22.1% Full deployment 13.8% Feasibility study 4.1%

Table 4 Percentage of projects per tag type Tag type Percentage Price label / hang tag 59.1% Care label 6.0% Hard tag 6.0% Sewn-in tag 5.4% Unknown 23.5%

4.2.1 UCs and project status timeline

Footnotes

References

Table 3
Distribution of project status

Project status Percentage

Pilot 32.4%

Phased deployment 27.6%

Proof of concept 22.1%

Full deployment 13.8%

Feasibility study 4.1%

Table 4
Percentage of projects per tag type

Tag type Percentage

Price label / hang tag 59.1%

Care label 6.0%

Hard tag 6.0%

Sewn-in tag 5.4%

Unknown 23.5%