Understanding Product Returns: A Systematic Literature Review using

Machine Learning and Bibliometric Analysis

Quang Huy Duong1, Li Zhou1*, Meng Meng2, Truong Van Nguyen3, Petros Ieromonachou1, Duy Tiep
Nguyen1

1Faculty of Business, University of Greenwich, London SE10 9LS, UK

2 School of Management, University of Bath, Bath, BA2 7AY, UK

3 Brunel Business School, Brunel University London, Uxbridge, UB8 3PH, UK

*Corresponding author: [Link]@[Link]

 Understanding Product Returns: A Systematic Literature Review

1 using Machine Learning and Bibliometric Analysis

2
3
4
5
6 Abstract
7
8 Product Returns (PR) are an inevitable yet costly process in business, especially in the online marketplace.
9 How to deal with the conundrums has attracted a great deal of attention from both practitioners and researchers.
10
11 This paper aims to synthesise research developments in the PR domain in order to provide an insightful picture
12
13 of current research and explore future directions for the research community. To ensure research rigour, we

14 adapt a six-step framework - defining the topic, searching databases, cleaning and clustering data, paper
15
16 selection, content analysis, and discussion. A hybrid approach is adopted for clustering and identifying the
17
18 distribution and themes in a large number of publications collected from academic databases. The hybrid
19
20 approach combines machine learning topic modelling and bibliometric analysis. The machine learning results
21 indicate that the overall research can be clustered into three groups: (1) operations management of PR, covering
22
23 (re)manufacturing network design, product recovery, reverse distribution, and quality of cores; (2) retailer and
24
25 (re)manufacturer issues including return policy, channel, inventory, pricing, and information strategies; and

26 (3) customer’s psychology, experience, and perception on marketing-operation interface. Furthermore, from
27
28 the content analysis, five potential future directions are discussed, namely digitalisation in the context of PR;
29
30 globalisation versus localisation in the context of PR; multi-layer (i.e., retailer, manufacturer, logistics
31 provider, online platform) and multi-channel (i.e., online, offline, dual and omni channel) oriented bespoke
32
33 return policy; understanding and predicting customer return behaviour via online footprints; and customer
34
35 return perception across the marketing–operations interface.
36
37 Keywords
38
39 Product returns; literature review; machine learning; topic modelling; bibliometric analysis; framework
40

55
56
57
58
59
60
61
62
63 1
64
65
 1
2
3
4
5
6
7
8
41 1. Introduction
42
43
44 The history of product returns (PR) can be traced back to the 60s-70s, when PR was implicitly embedded
45
46 in business tactics and strategies. For example, one of the first types of PR was refund policy (i.e., warranty

47 claim) introduced by Menke (1969). Years later, other types of return policies were developed to mitigate
48
49 customer purchasing risks and promote sales, such as money-back guarantee (Davis et al., 1995; Patankar &
50
51 Mitra, 1995). It is undeniable that PR is difficult to control and manage. According to the customer transaction

52 analysis, Barclaycard (2018) reveals that their users returned nearly a quarter of their purchased values
53
54 (approximately £7 billions) in both brick-and-mortar and online stores from 2016 to 2018. In addition, according

to GlobalData (2018), in the clothing and footwear industry, it is predicted that monetary values of online
returns will continue increasing to £5.6 billions by 2023, raising a huge concern to the retailers’ and OEMs’
capability for tackling returns. The situation became worse during the pandemic when retailers had to
extend their return window to allow customers more time taking product back to stores or post offices. This return
policy leniency encouraged customers to return even more than usual.

Besides, PR is driven by not only customers, but also environmental impacts, regulations, and
profitability. Many firms have been implementing a take-back program for over last two decades such as
Kodak, Xerox and HP to collect end-of-life products for remanufacturing or recycling (Pishvaee et al., 2010).
This leads to a necessity of integrating “Green” factors into remanufacturing tasks, for instance, production
9
10 and inventory model (Teunter, 2004; Teunter, 2001; Teunter, Van der Laan, & Inderfurth, 2000; Van Der

11 Laan, Dekker, & Salomon, 1996a; Van Der Laan, Dekker, Salomon, & Ridder, 1996b). These incorporations 12
13 with PR are the premise for an appearance of two well-known concepts: Reverse logistics (RL) and closed-
14
15 loop supply chain (CLSC). It is argued that PR can be used as a profit source. One of the typical benefits of

16 returned products is that they could be disassembled into components or recycled to raw materials. It would 17
18 save resources for buying new materials from their suppliers while reducing the amount of unwanted waste.
19

55
56
57
58
59
60
61
62
63 2
64
65
20
21 However, how to manage PR has proven to be challenging. It has attracted enormous attention both
22
23 academically and practically over the last couple of decades. Early studies include, for example, setting up

24 repair facility or return plant location (Daryanani & Miller, 1992; Marín & Pelegrín, 1998), finding optimal 25
26 RL/CLSC solutions to manage used product’s flow and maximise firm’s benefits (Fleischmann et al., 2001;
27
28 Jayaraman et al., 2003; Krikke et al., 1999). Recently, due to the return policy leniency to boost sales, the

29 development of e-commerce and the complication of customer’s psychology in nature, abusive returners have 30
31 been detected and started becoming a considerable concern in business (Ketzenberg et al., 2020). Even worse,
32
33 fraudulent PR started exceeding the organisation’s manageability. As the PR management is a complex

34 multifaceted process and a subject of high uncertainty in terms of quality, quantity, and timing, it is critical to
35
36 properly address PR issues to ensure the success of the whole RL and CLSC system. This motivates us to
37
38 conduct a literature review that systematically synthesises existing findings of PR literature to accelerate
39 progression of the research area.
40
41

42 There were some attempts to review some aspects of PR. While some existing literature reviews focus on 43
44 RL/CLSC/Green supply chain management in which PR was mentioned as one of the elements (Govindan et
45
46 al., 2015; Kazemi et al., 2019; Srivastava, 2007), other reviews focus more on specific angles of PR

47 management. For example, Janakiraman et al. (2016) reviewed return policy using a meta-analysis to find the 48
49 effects of different dimensions of return policy to purchase and return proclivity. Abdulla et al. (2019)
50
51 conducted a review of consumer return policy/management and its effect on consumer return behaviours based
52 on their proposed conceptual framework. Robertson et al. (2020) focussed on the provision of a commentary 53
54 in retail PR to find how PRs are transforming from the customer’s journey to retailer’s perspective. To the best of
our knowledge, the comprehensive review that provides a holistic view of PR management is still largely
understudied. So far, we only find one recent literature review of this topic by Ambilkar et al. (2021) but we
believe that by using a different approach, our systematic review enriches their findings for new explorations
of the subject area.

55
56
57
58
59
60
61
62
63 3
64
65
 1
2
3
4
5
6
7
8
Our main contribution is to map out an overall picture of PR which includes the synergy of current research
and future directions. The key questions to be answered in this review are:

• RQ1: What are the key research problems that have been addressed in the literature of PR
management?
• RQ2: What are the research methodologies used to address these key problems?
• RQ3: What are the emerging future research directions in the PR? 9
10
11 The paper is organised as follows. In section 2, we propose a methodology framework for the literature

12 review, describing data collection, cleaning, and clustering protocol. Section 3, we implement the process
of 13
14 bibliometric analysis on defined clusters from section 2. In section 4, we conduct a co-occurrence keyword
15
16 analysis on the selected papers. Section 5 is for reviewing the contents and answer RQ1 and RQ2. Next,

17 discussion and future directions RQ3 are laid out in Section 6, where related parties can retrieve or seek 18
19 theoretical, practical, and methodological implications for PR. Finally, section 7 is the conclusion and
20
21 limitations.
22
23
24 2. Methodological Framework and Publication Clustering
25
26 2.1. Methodology Framework
27
28 In this section, we propose a methodology framework to conduct the systematic literature review. It is a
29
30 synergy and advancement of Cooper (2010) and Mayring (2010), consisting of six fundamental steps as shown
31
32 in Figure 1: (1) Research background, (2) Defining search terms and collecting data using advanced search

33 function in Scopus, EBSCOhost and Web of Science (WoS) databases, (3) Data cleaning and clustering using
34
35 several cleaning and transformation techniques as well as an unsupervised ML model, (4) Paper selection for
55
56
57
58
59
60
61
62
63 4
64
65
 1
2
3
4
5
6
7
8
36
37 content analysis using Bibliometric Analysis, (5) Content analysis, and (6) Discussion and future directions.
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54

55
56
57
58
59
60
61
62
63 5
64
65
 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40 2.2. Materials
Collection 41
42 Papers were collected
across three large databases -
Scopus, WoS and Figure 1. Methodology framework EBSCOhost
via a two-step process.
43
44 Firstly, the keywords were chosen based on the same approach from Abdulla et al. (2019), Janakiraman et al.
55
56
57
58
59
60
61
62
63 6
64
65
 1
2
3
4
5
6
7
8
45
46 (2016) and Kazemi et al. (2019), in which the main topic became the keywords with different combinations.

47 The chosen keywords were “core acquisition”, “buyback core”, “product return”, “return behaviour”, “return
48
49 policy”, “consumer return”, “customer return” in different forms, for example, “return policy” versus “return
50
51 policies” in the papers. We also set the focus on Business, Management, Operations, Logistics, Supply Chain,

52 and Finance with English being the focused language. Publications were collected until Q2/2021. On that 53
54 account, we collected 973 papers from Scopus, 823 papers from WoS, 1206 papers from EBSCOhost, equating to
3002 papers in total. Secondly, as the data could be duplicated, invalid or omitted to some extent, we
crosschecked different datasets and removed duplications as well as invalid data by utilising Python to check
their titles, authors, years, sources. In the end, 1209 papers were selected for further analysis.

55
56
57
58
59
60
61
62
63 7
64
65
 1
2
3
4
5
6
7
8
2 2.3. Descriptive Analysis

3 The descriptive analysis is shown in Figure 2 indicating the distribution of the papers relating to PR. As
4 can be seen, researchers have been paying substantial attention to PR problem since the last decade, which has
5 gained further momentum in recent years. Table 1 provides a list of the top 20 journals contributing to PR
6 research in the dataset, totalling 603 papers (49.87% of the total publications). International Journal of
7 Production Economics, International Journal of Production Research, and European Journal of Operations
8 Research are the most contributed journals, with 260 papers being contributed (21.5% of the overall). Lastly,
9 the diversity of journal domains (e.g., Marketing Science, International Journal of Physical Distribution and
10 Logistics Management, Computers and Industrial Engineering, Production and Operations Management, etc)
11 shows the breadth of influence that PR has in the field of business and management.

12

13 Figure 2. The distribution of publications across our study period (1969 – Q2/2021)

14 Table 1. Top 20 contributed journals across the period of dataset (1969-Q2/2021)

55
56
57
58
59
60
61
62
63 8
64
65
 1
2
3
4
5
6
7
8
Number Of
Publications
Journals
International Journal of Production Economics 106
International Journal of Production Research 78
European Journal of Operational Research 76
Journal of Cleaner Production 62
Production and Operations Management 34
International Journal of Physical Distribution and Logistics Management 26
Marketing Science 21
Omega 20
15

55
56
57
58
59
60
61
62
63 9
64
65
 1
2
3
4
5
6
7
8
Annals of Operations Research 20
Computers and Industrial Engineering 20
Management Science 19
Journal of Retailing and Consumer Services 19
Journal of Retailing 15
Manufacturing and Service Operations Management 14
Journal of Operations Management 14
9
10 International Journal of Retail and Distribution Management 13
11
12 Transportation Research Part E: Logistics and Transportation Review 13
13
14 Psychology and Marketing 12

15 Journal of Manufacturing Systems 11

16
17 Decision Sciences 10
18
19 2.4. Latent Dirichlet Allocation Clustering and Results
20
21
22 Latent Dirichlet Allocation (LDA) Clustering and Visualisation
23
24 LDA-based topic modelling is a typical clustering technique for grouping documents (i.e., papers) in a

25 collection of documents (i.e., dataset) into topics (i.e., clusters) based on their semantic similarities (Blei
et al.,
26
27 2003). Traditionally, to determine which cluster a paper belongs to, we create a matrix to check the occurrence
28
29 of a word appears in every individual document such as latent semantic indexing (Deerwester et al., 1990;

30 Dumais et al., 1988) and probabilistic latent semantic indexing (Hofmann, 1999). However, these approaches
31

55
56
57
58
59
60
61
62
63 10
64
65
 1
2
3
4
5
6
7
8
32 would result in an exponentially large combination and expensive computation. One of the initiatives of LDA
33
34 is that it reduces the number of dimensions between words and their documents by adding a latent layer, so-

35 called latent clusters. These latent clusters will also be the clusters to which papers are classified. LDA also
36
37 provides the top frequent and representative words per cluster. The mathematical aspect of LDA can be seen
38
39 in Appendix A1.
40
41 LDA requires user to pre-define the number of clusters (K) before clustering. To find the optimal number
42
43 of K, Blei et al. (2003) originally proposed a perplexity score which was used to judge how good a LDA model

44 to derive meaningful cluster given a textual dataset. This metric has been widely used by many authors to find
45
46 the optimal number of topics from textual data such as online employee reviews (Jung and Suh, 2019),
47
48 discussion forum postings (Narang et al., 2021), loan requests (Netzer et al., 2019) and automobile insurance

49 claims (Wang and Xu, 2018). We use Python Gensim package which applies a log of the perplexity score for
50
51 evaluating the model. The higher the log of a perplexity score is, the better performance of the LDA model.
52
53 The LDA model with the K having the highest perplexity score will be selected as the optimal number of 54 clusters

for the dataset. How the log of a perplexity score is calculated can be found in Appendix A2.

The procedure for clustering and labelling our dataset has the following steps:

1) Abstracts cleaning and preparation.

2) Identifying the optimal value of K using log perplexity score.
3) Clustering 1209 abstracts using LDA model and the optimal K.
4) Visualising top keywords in each cluster using LDAvis technique (Sievert & Shirley, 2015) (see
Appendix A3 for mathematical expressions).
5) Cluster labelling.

55
56
57
58
59
60
61
62
63 11
64
65
 1
2
3
4
5
6
7
8
LDA Results
Before clustering the data, a protocol of cleaning noises such as punctuations, stop words (e.g., and, or,
9 they, could, etc) and lemmatising each abstract is necessary to reduce noises and increase the accuracy.

10 Furthermore, bigram and trigram (e.g., two, three words are likely together) models are applied to improve the 11
12 meaningfulness. The choice of cleaning process is synthesised from Chae and Olson, (2021), Netzer et al.,
13
14 (2019) and Wang and Xu (2018).
15
16 To determine the optimal K value, different K values are trained in the range from 2 to 71 clusters,

17 equating to 70 runs in total. From Figure 3, we can see that three clusters are the highest value of log
perplexity.
18
19 Hence, in the next analysis, we select three clusters.
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38 39
40 Figure 3. Log perplexity values by number of clusters
41
42

55
56
57
58
59
60
61
62
63 12
64
65
 1
2
3
4
5
6
7
8
43 Our dataset is clustered with K = 3 clusters using LDA topic model. The outcomes include two main

44 figures: (1) top keywords of each cluster (Table 2) and (2) papers in each cluster (Table 3).
45
46
47 In order to visualise the top keywords of each cluster, we adopt LDAvis (Sievert & Shirley, 2015) to

48 depict the semantic distance map among three clusters and their keywords. The trained LDA model is
input to 49
50 LDAvis for keywords’ visualisation. Figure 4 shows the results from LDAvis. There are two main parts from
51
52 LDAvis results: (1) The left side shows the semantic map over the clusters. Overall, three clusters are highly

53 distinctive in terms of semantic words. (2) The right side shows top keywords in each cluster adjusted by the
54 value 𝜆. The value of 𝜆 is between 0 and 1. When 𝜆 is closer to 1, the model returns common keywords that
may appear in other clusters. In contrast, when 𝜆 is closer to 0, the unique keywords are returned only in their
clusters. Table 2 depicts the top 20 keywords in three clusters with 𝜆 = 1 and 0

55
56
57
58
59
60
61
62
63 13
64
65
1 In order to label each cluster, we based on a basis of top frequent keywords per cluster in Table 2 and the
2 expertise of our research group. Each researcher in the group independently labels the clusters, then all
3 researchers collectively discuss and cross-check to determine the final labels for the three clusters. As a result,
4 labels of the three clusters are displayed on Table 3.

5
(a)

7 (b)

8 Figure 4. Semantic keyword visualisation with 𝜆 = 0 (a) and 𝜆 = 1 (b)

9 Table 2. Top 20 keywords with 𝜆 = 0 and 𝜆 = 1 for each cluster

Cluster 1 Cluster 2 Cluster 3

14
 𝝀=𝟏 𝝀=𝟎 𝝀=𝟏 𝝀=𝟎 𝝀=𝟏 𝝀=𝟎
Product Network Product Contract Product Author
Model Manufacturing Retailer Pricing Customer Survey
Remanufacture Facility Price Decentralize Consumer Intention
Cost Disassembly Manufacturer Valuation Study Trust
Propose Center Consumer Equilibrium Retailer Perception
System Heuristic Policy Game Policy Recall
Use Transportation Model Dealer Service Empirically
Problem Life Strategy Utility Research Regret
Demand Reuse Online Centralize Management Understanding
Consider Constraint Channel Tax Purchase Abuse
Inventory Forecast Optimal Centralize Online Feedback
Study Energy Profit Showroom Result Cue
Paper Computational High Centralized Use Mediate
Policy Horizon Decision Sensitive Increase Device
Optimal Forecasting Increase Wholesale price Effect Leniency
Time Plant Quality Freight Also App
Production Programming Market Buyback Process Questionnaire
Base Planning Customer Decentralized Identify Returner
New Allocation Contract Revenue share High Rating
Solution Portfolio Sell Greenness Firm Employee
10 Table 3. Cluster labels and number of publications from results of LDA

Cluster Label Number of Publications

Cluster 1 Operations Management of PR 478

Cluster 2 Retailer and (Re)Manufacturer 418

Cluster 3 Customer’s Psychology and Experience 313

11 3. Bibliometric Analysis for Paper Selection of Content Analysis

12 After three clusters are identified from LDA cluster modelling, the next stage is to select the most
13 representative papers for each of the clusters for further analysis. Here, bibliometric analysis is adopted for
14 three reasons. Firstly, bibliometric analysis enables users to demonstrate the dynamic knowledge domain based
15 on scientific connection and development. Secondly, it is a quantitative method that map out the relationship
16 between publications mathematically and statistically (Si et al., 2019). Lastly, there has been a widespread use

15
 1
2
3
4
5
6
7
8
of bibliometric analysis in several aspects of operations, logistics and supply chain management to select
appropriate papers for the content analysis (Ciano, Pozzi, Rossi, & Strozzi, 2019; Kazemi et al., 2019; Laengle
et al., 2017; Li, Huang, & Yang, 2020; Merigó & Yang, 2017; Si et al., 2019; Wang, Chen, Rogers, Ellram, &
Grawe, 2017; Xu, Zhang, Feng, & Yang, 2020; Xu et al., 2018). In this paper, we use VOSviewer software
(van Eck & Waltman, 2010) among various available bibliometric software such as Pajek (Dohleman, 2006),
Gephi (Bastian et al., 2009), Sci2 (Börner et al., 2003) to conduct the analysis. VOSviewer is compatible with
9 a range of handy and flexible publication formats such as RIS, Scopus, WoS, etc. Also, advanced bibliometric

10 tools such as co-author, co-occurrence, co-citation, and bibliographic coupling are integrated into VOSviewer.
11
12
13 3.1. Co-citation and Bibliographic Coupling Analysis
14
15 From a scientific perspective, co-citation (Marshakova, 1973; Small, 1973) and bibliographic coupling

16 (Kessler, 1963) analysis are two tools utilised to map the citation relationships among publications in a
17
18 systematic manner. Papers are more likely to represent the same clusters and methodologies if they cite or are
19
20 cited by each other more frequently (Hjørland, 2013). The reason why we do not use the direct citation analysis

21 to find representative papers is because direct citation is argued to require a lengthy yet less accurate time 22
23 connection to fully depict the whole citation picture (Boyack & Klavans, 2010). Instead, co-citation and
24
25 bibliographic coupling connect two papers using a third paper, hence the connection is not influenced by the

26 timeline, but it is complemented by a third paper. Therefore, it is easier to accurately capture a full picture of
27
28 the topic in a limited time period. We simplify the definition of both analyses as illustrated in Figure 5: (a)
29
30 Paper B and C are considered to have a co-citation relationship when there is paper A cites both B and C, and
31
32 (b) in contrast, paper B bibliographically coupled with paper C when they both cite paper A (van Eck &

33 Waltman, 2014). Both techniques have been widely used in different domains to find the well-discussed papers
34
35 within a cluster (Modak et al., 2020; Pilkington & Meredith, 2009; Wang et al., 2016; Xu et al., 2018).
36

55
56
57
58
59
60
61
62
63 16
64
65
37 In this paper, we apply a hybrid approach, running co-citation and bibliographic coupling analysis at the
38
39 same time for each cluster. We input three clusters respectively to VOSviewer and set up a threshold of 3
40 connections to be included in the network. The chosen papers will be the ones which satisfy 2 criteria: (1) got
41
42 at least 3 co-citation connections (2) got at least 3 bibliographic coupling connection. As a result, Cluster 1 (29
43
44 papers), Cluster 2 (59 papers) and Cluster 3 (23 papers) are selected for further analysis.
45
46
47
48
49
50
51
52 53

54 (a) (b)

Figure 5. Example of a co-citation relationship (a) and a bibliographic coupling relationship (b)
3.2. Recent Publication Analysis

1 For the papers which are left outside bibliographic coupling and co-citation networks, there is a concern

2 about overlooking the emergence of new trends in scientific research. Specifically, it is unlikely that many
3
4 recent papers are cited by peers, yet they are very valuable as these papers have contributed to the most up-to-
5
6 date findings. Within each cluster, there are a number of papers without any connection (Cluster 1: 147 papers,

7 Cluster 2: 153 papers, Cluster 3: 178 papers). As such, we scrutinise papers published between 2019 to 8
9 Q2/2021 in order to capture the current trends, with 62, 61, and 76 papers the being selected corresponding to
10
11 Cluster 1, Cluster 2, and Cluster 3 respectively.
12
13 3.3. Journal Quality Analysis
14
15 We conduct content analysis on the papers published in Grade 3, 4, 4* journals based on the Academic 16
17 Journal Guide (AJG) 2018. AJG (2018) is a well-known guide which ranks the quality of journals in Business
18
19 and Management domains. In summary, Cluster 1 (91 papers), Cluster 2 (120 papers) and Cluster 3 (99 papers)

20 equating to 310 papers after co-citation, bibliographic coupling and recent publication analysis are checked 21
22 via the journal quality analysis. The results show 167 out of 310 papers (53.87%) were published in Grade 3,
23
24 4, 4* journals, with the details for each cluster are shown in Table 4.
25
26 Table 4. Number of selected papers for content analysis after the journal quality analysis
27
28

17
 1
2
3
4
5
6
7
8
29 Cluster 1 Cluster 2 Cluster 3
30
31
32
Selected papers with all 21 55 10
33 connections
34
35
36 Recent published papers 25 42 14
37
38 without connections
39
40
41 4. Cluster Analysis using Co-occurrence Keyword Heatmap
42
43 In this section, we deploy a co-occurrence keyword analysis on selected papers. This analysis provides a

44 comprehensive overview of what has been addressed among the publications as well as initialises potential
45
46 sub-clusters in each cluster. We implement the analysis by using VOSviewer software. Having the function of
47
48 density visualisation, under heatmap format, the software provides glances of papers’ content on each cluster.
49
50 Figure 6 illustrates the distributions of the keywords in Cluster 1 (373 keywords), Cluster 2 (554
51
52 keywords) and Cluster 3 (167 keywords). With regards to Cluster 1, a first glance gives an initial sense of how

53 operations management have appeared in PR research (Figure 6a). More specifically, considering keywords’
54
55 magnitude, the most prevalent keywords are “remanufacturing”, “reverse logistics”, “remanufacturing 56
57 system”, “logistics”, “supply chain”, “location” and “inventory”. Furthermore, the density of these keywords 58
locating in the middle relatively shows the connection between papers within the cluster. This indicates that 59
60 the papers in Cluster 1 specialise in manufacturing and structuring of a supply chain as well as dealing with
61
62
63
64
65
problems at the upstream and (re)manufacturing stage. These stages integrate a reverse flow while implicitly
embed a PR flow into a supply chain. In terms of methodology, despite sharing the same clusters, their
solutions, which can be seen at the edges of Figure 6a, vary from “assortment optimisation”, “stochastic
55
56
57
58
59
60
61
62
63 18
64
65
 model”, “data-driven simulation”, etc, illustrative of the diversification in techniques for solving operational
problems in RL and CLSCs.

With respect to Cluster 2, what stands out in Figure 6b is that “product returns” is explicitly discussed.

9 “return policy”, “electronic commerce”, “supply chains”, “costs”, “sales”, “money-back guarantee”, “online
10
11 retailing”, “optimal pricing”, etc. also stand out among the common keywords. This indicates that authors
12
13 mainly focus on solving PR problems at the single/dual echelon level with the participation of manufacturers

14 and retailers. Regarding methodology, there are few methodology keywords such as “artificial intelligence”,
15
16 “field experiment”, “average treatment effect”. It suggests that solutions are more case-oriented than using
17
18 generic solutions/optimisations.
19
20 The final cluster addresses customers’ different perceptions of PR. It is apparent from Figure 6c that the
21
22 distribution is scattered, hence, it could be explained as customer’s elements are heterogeneous in nature.

23 Interestingly, while “product return” is the most common cluster, it has been approached using different themes
24
25 such as “consumer behaviour”, “attractiveness of alternatives”, “return abuse”, “brand equity”, “customer
26
27 satisfaction”, “brand familiarity”, etc. These keywords provide an initial overview of what has been done by

28 previous authors from the customer perspective. Methodologies are also affected by the variety of perceptions.
29
30 In this regard, authors adopted some techniques and data such as “econometric model”, “formalization”,
31
32 “behavioral studies”, “archival data”.
33

19
 1
2
3
4
5
6
7
8
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54

(a)

55
56
57
58
59
60
61
62
63 20
64
65
1 (b)

3 (c)

4 Figure 6. Density graphs with co-occurrence author and indexed keywords analysis for Cluster 1 (a), Cluster
5 2 (b) and Cluster 3 (c)

6 Overall, Cluster 1 covers the whole structure of a closed-loop supply chain, integrating main
7 (re)manufacturing activities and supporting activities such as collecting, acquisition, etc. Cluster 2 targets the
8 way how firms handle and respond to the returns flow focusing on (re)manufacturer and retailer domain.
9 Cluster 3 studies the way customers perceive returns function/process/policy by themselves or under retailers’

21
 1
2
3
4
5
6
7
8
impacts. Therefore, based on the heatmap from Figure 6 and the labels from Table 3, we can draw a map to
integrate the three clusters into a typical CLSC model (Figure 7).

9
10
11
12
13
14
15
16
17 18
19
20 Figure 7. Clusters’ focus to forward and reverse flows in a closed loop supply chain
21
22
23 5. Content Analysis
24
25 This section closely explores the insights of sub-themes in each of the three clusters by examining their
26
27 research issues and methodologies. Compared with previous literature reviews in topics such as Ambilkar et

28 al. (2021), which identified six research themes of PR management (i.e., return policy, product recovery, 29
30 consumer behaviour, forecasting PR, product uncertainty, and technology), our findings unveil some of the
31
32 other key research issues that are worthwhile of investigation. The details of our research themes can be seen
33 as below.
34
35
36 5.1. Cluster 1 – Operations Management of Product Returns
37
38 With 46 papers, Cluster 1 is the second largest cluster. According to LDA results and co-occurrence
39
40 keyword analysis, papers in this cluster prominently discuss about operational aspect of PR such as

55
56
57
58
59
60
61
62
63 22
64
65
41 remanufacturing, disassembly and distribution. In fact, these concepts were initially established to deal with
42
43 various reasons for PR from environmental factors (e.g., end-of-life products (core), waste disposal) to
44
45 mismatch with customer preferences (e.g., defective returns, unused returns). This cluster has gained more

46 attention in recent years (2019-2021) due to the uncertainties of the dynamic world (e.g., disasters, 47
48 digitalisation).
49
50 Because of the homogeneous connections of the papers in this cluster and the hierarchical order via a 51
52 preliminary analysis, the content analysis will be conducted by dividing and reviewing papers according to
53
54 three subsections: (1) Operations issues in (re)manufacturing, (2) Network design and product recovery

management, and (3) Reverse distribution and quality of returned product for reprocessing.

Operations Issues in (Re)Manufacturing

23
 1
2
3
4
5
6
7
8
Overall, there are 27 papers (58.69%) in this subtheme. It is clear that the papers in the earlier years
focussed on various aspects of operations for production and inventory problems in both planning and control in
manufacturing/remanufacturing operations (de Brito & van der Laan, 2009; DeCroix & Zipkin, 2005; Guide Jr, 2000;
Mahadevan, Pyke, & Fleischmann, 2003; Minner & Kleber, 2001; Niknejad & Petrovic, 2014; Van Der Laan, Salomon,
Dekker, & Van Wassenhove, 1999; Zhou, Tao, & Chao, 2011). Until recently, while the stream of inventory and
production are still trending, the research focus has evolved to some contemporary 9 combinations such as trade-in
and refurbished service level (Jiang et al., 2019; Shin et al., 2020), inventory

10 policy via PR forecasting (Chou et al., 2020), location-inventory problems (Guo et al., 2020a, 2020b),
11
12 repairable inventory system (Lin, Leung, Zhang, & Gu, 2020), quality-grading scheme (Ponte et al., 2021),
13
14 and capital investment (Reddy and Kumar, 2021). In terms of production, researchers have prominently

15 addressed production issues along with machinery problems, failures and maintenance (Assid et al., 2020;
16
17 Ndhaief et al., 2020; V. Polotski et al., 2019; Vladmir Polotski et al., 2019) and global production and
18
19 distribution (Mishra & Singh, 2020a, 2020b).
20
21 Beyond what has been discussed, researchers also study some other aspects of operations such as

22 disassembly planning (Kumar & Putnam, 2008; Wang et al., 2020), dynamics in the hybrid system (Aras,
23
24 Boyaci, & Verter, 2004; Zhou, Naim, & Disney, 2017) and accounting-financial tension (Mutha et al., 2021).
25
26 In addition, climate change and environmental issues prompt organisations to extend their single financial

27 objective into multi-objective comprising “Green Performance” e.g. minimising negative environmental 28
29 impacts (Mishra and Singh, 2020b, 2020a). Another common stream in remanufacturing returned products is
30
31 extracting salvageable value of returned products for further inspection, disassembly, cleaning or

32 refurbishment (Chou et al., 2020; Minner & Kleber, 2001; Zhou, Tao & Chao, 2011).
33
55
56
57
58
59
60
61
62
63 24
64
65
 1
2
3
4
5
6
7
8
34
35 The comprehensive summary of the papers in this sub-theme is shown in Appendix B.
36
37 Network Design and Product Recovery Management 38
39 There are 13 papers (28.26%) in this subtheme. According to Appendix C, the addressed clusters and
40
41 methodologies are as premises for remanufacturing cores. There are two main streams discussed: network
42
43 design and product recovery management.
44
45 Remanufacturing cannot be operated without various prerequisites. One of these significant elements is

46 reverse logistics/closed-loop supply chain network design for PR flows (Alumur, Nickel, Saldanha-Da-
Gama, 47
48 & Verter, 2012; Fleischmann et al., 2001; Tosarkani & Amin, 2018; Yıldız & Soylu, 2019; Zarbakhshnia,
49
50 Kannan, Kiani Mavi, & Soleimani, 2020). Despite the intention towards corporate objectives, the practical

51 implications of network design research are prominently to address some daily issues in remanufacturing 52
53 operations such as facilities, collection centres, repair services and cost of CO2 emission management.
54
Product recovery management is another considerable stream in this sub-theme. In fact, handling returned
products from the downstream back to manufacturers is initially challenging and cost inefficiency. This is the
reason for researchers to study comprehensively pre-activities of remanufacturing such as cores acquisition
management (Guide & Jayaraman, 2000), product recovery management (e.g., procurement, disposal)
(Inderfurth, 1997), data-driven simulation for smart remanufacturing system (Goodall et al., 2019), facilities
(e.g., location, repair facilities) (Krug et al., 2021; Min and Ko, 2008a).

Forecasting PR is a crucial aspect of product recovery management and plays an important role in ensuring
remanufacturing stability, inventory control and reducing uncertainty. Accurate predictability can be enhanced
by various attributes (e.g., sales, transactions) (Clottey et al., 2012; Shang et al., 2020) and data mining
techniques (Cui et al., 2020).

55
56
57
58
59
60
61
62
63 25
64
65
 1
2
3
4
5
6
7
8
9
10 The comprehensive summary of papers in this subtheme is in Appendix C.
11
12 Reverse Distribution and Quality of Returned Product for Reprocessing
13
14 Also in this cluster, we review 6 papers (13.05%) addressing reusable products from PR. Varying the
15
16 quality of returned products prompt researchers to find the optimal profitability of reuse activities (Zikopoulos

17 & Tagaras, 2007), dynamic assortment (Rusmevichientong et al., 2020), product lifecycle (Muyldermans et
18
19 al., 2019) and reverse distribution (Jayaraman et al., 2003). Their methodologies are quantitative models such
20
21 as finite lifecycle model, choice model. In this cluster, there are two literature reviews of closed-loop supply

22 chain and potential uncertainties on the remanufacturing (Goltsos et al., 2019; Souza, 2013).
23
24
25 Overall, the papers in this cluster are strongly connected with each other. From a methodological
26 perspective, most of the papers adopt quantitative research (i.e., mathematical modelling or simulation),
27
28 starting from identifying problems, mathematical model development, model solution, numerical experiments,
29
30 and practical and managerial implications. There are some common quantitative models such as mixed-integer

31 linear programming (MILP) (Alumur et al., 2012; Fleischmann et al., 2001; Krug et al., 2021; Min and Ko, 32
33 2008b; Yıldız and Soylu, 2019) or mixed-integer non-linear programming (MINLP) (Guo et al., 2020a, 2020b;
34
35 Mishra and Singh, 2020b, 2020a) or fuzzy mixed-integer programming (FMIP) (Niknejad & Petrovic, 2014),

36 full fuzzy programming (FFP) (Tosarkani & Amin, 2018), continuous-time Markov chain model (Aras, Boyaci,
37
38 & Verter, 2004), finite horizon periodic review backlog models (Chou et al., 2020; Zhou et al., 2011) and data-
39
40 driven simulation (Goodall, Sharpe, & West, 2019).
41
42 5.2. Cluster 2 – Retailer and (Re)Manufacturer
55
56
57
58
59
60
61
62
63 26
64
65
 1
2
3
4
5
6
7
8
43
44 97 papers make Cluster 2 the largest cluster. At the first glance, researchers mainly seek to solve marketing
45
46 related problems at manufacturer/retailer level because most papers (69.07%) addresses return policy issues.

47 We structure this cluster as return policy issues and miscellaneous issues.

48
49
50 Return Policy (RP)
51
52 There are 67 papers (69.07%) reviewed, starting from 1985 to 2020. Even though RP has been presenting
53
54 since 1985, the recent figure shows the highest number of papers in 2020 (13 papers) in comparison with previous

years. It suggests that RP is still a state-of-the-art cluster and retailers still need to enhance their RP to cope
with new dynamic concepts (e.g., e-commerce, AI). Some interesting patterns are recognised. Firstly, because
of the longitudinal existence, RP has been evolving from monetary perspectives to more supply chain related
operations such as inventory, information strategies, ordering quantity, brand competition, return
freight insurance, pricing, etc. Secondly, RP is usually deployed at retailer level where PR occurs from
customers (Akçay, Boyacbox, & Zhang, 2013; Anderson, Hansen, & Simester, 2009; Bonifield, Cole, &
Schultz, 2010; Bower & Maxham, 2012; Che, 1996; Chen & Chen, 2017; Chen & Bell, 2012; Heiman,
McWilliams, Zhao, & Zilberman, 2002; Hess, Chu, & Gerstner, 1996; Hsiao & Chen, 2012, 2014, 2015;
Huang & Zhang, 2020; Janakiraman et al., 2016; Jin, Caliskan-Demirag, Chen, & Huang, 2020; Khouja, Ajjan,
&
Liu, 2019; Lal & Sarvary, 1999; Lee, Choi, & Edwin Cheng, 2021; Lepthien & Clement, 2019; Wei Li, Chen,
9 & Chen, 2018; Li, Xu, & Li, 2013; Lin, Zhang, & Cheng, 2020; McWilliams, 2012; Mukhopadhyay &

10 Setaputra, 2007; Mukhopadhyay & Setoputro, 2005; Nageswaran, Cho, & Scheller-Wolf, 2017; Shang,
11
12 Ferguson, & Galbreth, 2019; Shehu, Papies, & Neslin, 2020; Shulman, Coughlan, & Savaskan, 2009, 2011;
13
14 Su, 2009; Sun, Chen, Tian, & Yan, 2021; Wagner & Martínez-De-Albéniz, 2020; Xing Wan, Li, Chen, & Lei,

55
56
57
58
59
60
61
62
63 27
64
65
 1
2
3
4
5
6
7
8
15 2020; Wood, 2001; Xu & Duan, 2020; Yalabik, Petruzzi, & Chhajed, 2005; Yang, Ji, & Tan, 2020).
16
17 Furthermore, RP in the coordination channels between single/dual manufacturer – single/dual retailer in which
18
19 PR is induced by unsold stocks or consumer PR is also considered (Ai, Chen, Zhao, & Tang, 2012;

20 Bandyopadhyay & Paul, 2010; Chen & Grewal, 2013; Chen, 2011; Choi, Li, & Yan, 2004; Crocker & Letizia,
21
22 2014; Emmons & Gilbert, 1998; Fan & Chen, 2020; Lee, 2001; Li, Li, Sethi, & Guan, 2019; Li, Chen, Liang,
23
24 & Chen, 2018; Liu, Mantin, & Wang, 2014; Ma, Di, & Hsiao, 2020; Matsui, 2010; Padmanabhan & Png, 1997;
25 Pasternack, 1985; Ruiz-Benitez & Muriel, 2014; Shulman, Coughlan, & Savaskan, 2010; Tsay, 2001; Wang,
26
27 2004; Xiao, Shi, & Yang, 2010; Xu, Li, Govindan, & Xu, 2015; Yang, Chen, Chen, & Chen, 2017; Yoo, 2014;
28
29 Yue & Raghunathan, 2007). In terms of RP elements, five factors are commonly discussed namely monetary

30 (e.g., refund policy, money-back guarantee (MBG), restocking and handling fees, shipping fees), time (e.g.,
31
32 return window), effort (e.g., the extent of hassle in return procedures), exchange (e.g., refund in cash or
33
34 discount, vouchers) and scope (e.g., different policies within the same retailer) (Janakiraman et al., 2016). The

35 existence of five factors is fully observed in aforementioned papers, yet it is worthy to point out that a single-
36
37 factor research tendency in the past is being replaced by a joint effect of a dual-factor in recent years, for
38
39 example, time and scope (Ma et al., 2020; Shang et al., 2019). More recently, there is a tendency to focus on

40 RP in the context of e-commerce due to the acceleration of online commerce platforms such as post-purchase
41
42 warranty and return insurance for platform providers (Chen et al., 2021; Fang et al., 2021; Li et al., 2021) and
43
44 service strategy (Gäthke et al., 2021).
45
46 From a methodological perspective, most papers are conducted with similar procedures of problem
47

55
56
57
58
59
60
61
62
63 28
64
65
 1
2
3
4
5
6
7
8
48 identification, mathematical modelling, scenario comparison, and testing/experiments with a diversity of

49 product types (e.g., consumables, durable, build-to-order, etc.). Some common models can be listed such as
50
51 newsvendor models, field experiment, meta-analysis, and stylised model. There are a few papers addressing
52
53 problems by using conceptual frameworks with customer engagement (Bonifield et al., 2010; Bower & 54

Maxham, 2012; Wood, 2001) and only one paper reviewing the literature (Janakiraman et al., 2016).

Appendix D summaries current focusses in PR.

Channel, Inventory, Pricing, and Information Strategies in the Context of Retailers

30 papers (30.93%) concern the other dilemmas of managing and handling PR in the context of retailers.
Such supply chain coordination and reverse channel are focusses which are used to decide whether a
manufacturer collected consumer returned products directly from customers or via a retailer/collector with the
risk of fusing unsold products and used products together (Ferguson et al., 2006; Hosseini-Motlagh et al., 2020;
Kumar et al., 2019; Lee and Rhee, 2021; Liu et al., 2020; Mandal et al., 2021; Ofek et al., 2011; Pun et al.,
2020; Savaskan and Van Wassenhove, 2006; Tang et al., 2020; C. Wang et al., 2020; X. Wang et al., 2021; 9
Zhang and Choi, 2020). Inventory management and the Bullwhip Effect are also issues because unsold

10 products are returned to manufacturers, which may induce overstocking or understocking (Dimitrov and
11
12 Ceryan, 2019; Papanagnou, 2021; D. Wang et al., 2020; Zhang et al., 2020). While some retailers may
13
14 anticipate and take into account RP for pricing strategy (Chen & Bell, 2009; Mahmoudzadeh, 2020; Vorasayan

15 & Ryan, 2006; Wang, Wang, & Chen, 2021; Zou, Zhou, & Jiang, 2020), some authors argue that information
16
17 sharing/ asymmetric information between parties in the CLSC or between customers and retailers are also
18
19 important in mitigating the uncertainties (Casalin & Dia, 2019; Rao et al., 2014; Seeger et al., 2019; Yan &

20 Cao, 2017). Finally, solutions for some special problems of PR in retailers such as optimal resources allocation
21
22 and trade-in value effects are derived (Borenich et al., 2020; Ke & Yan, 2020; Petersen & Kumar, 2015) or the
55
56
57
58
59
60
61
62
63 29
64
65
 1
2
3
4
5
6
7
8
23
24 interpretation and calculation of return rates (El Kihal et al., 2021).
25
26 Regarding the methodological aspect, along similar lines with other papers in this cluster, the majority of

27 papers are conducted by quantitative research. A minor number of papers adopt a conceptual framework
and 28
29 hypothesis testing to find the customer insights (Mahmoudzadeh, 2020; Rao et al., 2014; Seeger et al., 2019;
30
31 Yan & Cao, 2017). A more detailed summary can be found in Appendix E. 32
33 5.3. Cluster 3 – Customer’s Psychology and Experience
34

35 In this cluster, our review captures and expresses the content of 24 selected papers. The prominent 36
37 concentration is in ways that customers perceive various types of returns management strategy/activity from
38
39 sellers. Hence, the majority of papers are conducted in conceptualisation with customer’s participations as a

40 distinct feature. Furthermore, it is undeniable that customer perception is a multi-faceted concept (e.g., 41
42 consumer behaviour, cognitive dissonance, satisfaction, risk, etc) under many business problems (e.g., sales,
43
44 environment, etc), that is also not exceptional for PR. Hence, in customer perspectives prevails, various aspects

45 of what have been addressed in the literature, comprising of (i) customer’s psychology, (ii) marketing and 46
47 operation interface, (iii) customer perception on information and technologies, and (iv) customer perception
48
49 on RL and CLSC Management will be navigated from the selected papers.
50
51 Customer’s Psychology
52
53 There are 5 papers (20.83%) which solely studied the dynamic perceptions of customers, but their focuses
54 varied in diverse circumstances of PR. To be more specific, Powers & Jack (2013) and Ketzenberg et al.
(2020) addressed the motivations to return a product, including cognitive dissonance and return abuse
when a customer is uncertain about their decisions or just taking it for granted under the presence of liberal

55
56
57
58
59
60
61
62
63 30
64
65
 1
2
3
4
5
6
7
8
return policies. From another aspect, in 2005, a study’s focus was on the perception of consumers in the
pre-choice
process (i.e., between choice and consumption) which may be reversible because of disconfirming information
at the post-purchase stage (i.e., a customer recognises an inappropriateness in their purchase) (Bechwati &
Siegal, 2005). In contrast, Griffis et al. (2012) examined customer returns experience towards repurchase
behaviours. The final paper concentrated on customer viewpoints under two key psychological rewards (e.g.,
financial and emotional) toward product retention and disposal decision when a product still possesses reusable
value (Simpson et al., 2019).

9 Regarding to research methods, most papers use conceptual modelling (4 papers), except one paper
10
11 (Ketzenberg et al., 2020) conducts a large data-driven approach in finding customer return behavioural
12 patterns.
13
14
15 Marketing and Operations Interface
16
17 In this subset, a total of 6 papers (25%) are explored. In terms of marketing aspect, customer loyalty
18
19 towards firms with or without the existence of returns management system is studied via different determinants
20 such as customer satisfaction, perceived value offerings and previous service experience in internet retailing
21
22 (Mollenkopf et al., 2007). An interesting trend that is being developed recently is brand awareness in PR.
23
24 Ertekin et al. (2019) and Lee & Yi (2019) prominently focused on promotion/stacked discount framing on PR,

25 they stated that the effects may differentiate by different levels of brand recognition. Recently, Moreau (2020)
26
27 raised an issue of first brand impression through improving their package for delivery, called “doorstep
28
29 branding” strategy that sharpens their brand equity and decreases return probabilities. Yet, the elements of

30 branding and promotion in e-commerce have not been discussed thoroughly and reveals some potentials for
31
55
56
57
58
59
60
61
62
63 31
64
65
 1
2
3
4
5
6
7
8
32 future contribution.
33
34 However, due to the explosion of e-commerce, the line between operations and marketing is gradually
35
36 blurred, which prompts a well-discussed concept called marketing-operations interface (M-OI). Bijmolt et al.

37 (2021) and Mollenkopf et al. (2011) examined the role of M-OI in driving the way customers perceive returns
38
39 management process and how PR engages into a M-OI process in the omni-channel.
40
41 In addition, there is a wide range of methodologies that have been adopted ranging from conceptual 42
43 modelling (5 papers) and literature review (1 paper). 44
45 Customer Perception on Information and Technologies
46
47
48 There are 8 papers (33.34%) in this subset. With the shift from brick-and-mortar retailer to online retailer,

49 the role of information and technology in terms of sufficiency and accuracy to supplement the lack of
physical
50
51 inspection is deemed to receive much attention. Because customers cannot directly experience in-store
52
53 products to mitigate their uncertainties and perceived risk about the new products, they are inclined to rely on

54 the only information available –website information (WI). On one hand, some articles address aspects of WI
generating by novel technologies (e.g., information search in mobile, zoom, alternative photos, social media,
and colour swatch) on PR and customer returns perception (e.g., attractiveness of alternatives, perceived of
risk) (De et al., 2013; Huang et al., 2020; Li and Choudhury, 2021; Zhang et al., 2021). These technologies are
to dictate factual and impression-based information which matters at both pre-purchase and post-purchase
stages to consolidate for purchasing and varying returns. On the other hand, Maaya, Meulders & Vandebroek
(2020) raised quality of information of provision of WI such as the arbitrary WI of delivery time, price, discount, returns,
and especially the reliability. In addition to WI, online customer ratings are considered as a valuable source for exploring
customer behaviour (e.g., customer loyalty) towards some risk characteristics of products (Ramanathan, 2011). Another
55
56
57
58
59
60
61
62
63 32
64
65
 1
2
3
4
5
6
7
8
noteworthy phenomenon is retail return abuse which could be due to the leniency of RP. Akturk et al. (2021) studied
two technologies which are customer profiling and product 9 tracking to potentially address this issue. Another stream
of information is commonly derived from website

10 data can be studied the short-term and long-term impact of promotion to PR after sale (Liu et al., 2021).
11

12 In summary, online information and technologies have effects on consumer returns cognition, but they 13
14 are undeveloped and fragmented with rare connections to one another. The methodologies of what have been
15
16 conducted to address the research problems in this subset are mathematical modelling (1 paper), conceptual

17 modelling (6 papers) and a systematic review (1 paper).

18
19
20 Customer Perception on RL and CLSC Management
21
22 There are 5 papers (20.83%) in this subset, mainly focusing on consumer perception in product returns
23
24 operations. Different with what have been discussed in Cluster 1, the concepts of RL and CLSC are considered

25 as a marketing strategy of firms which affects customer’s perception when they make purchase and returns 26
27 decision. In addition, there is no overlap with Cluster 1 in terms of both theories and methodologies. In fact, 3
28
29 papers discuss about the role of designing RL program and process such as return policy, capabilities,

30 innovation onto the business performance (Autry, 2005; Richey et al., 2005; Stock et al., 2006). Yet, their 31
32 studies consider a business-to-business (B2B) model under the perspective of buyers as customers. Another
33
34 aspect of RL and CLSC is raised recently by Stauffer & Kumar (2021) based on the emergence of disasters

35 (e.g., hurricane, flood, pandemic, etc.) in the context of donation that could be managed by incorporating 36
37 humanitarian PR and disposal in a preparation stage. Frei et al. (2020) conducted a general view of the scale
38
39 and significance of returns systems towards achieving simultaneously economic and ecological sustainability.
40
41 The dominated methodologies in this subset are conceptual modelling (3 papers), interdisciplinary
55
56
57
58
59
60
61
62
63 33
64
65
 1
2
3
4
5
6
7
8
42
43 research (1 paper) and stochastic optimisation modelling (1 paper). Overall, clusters in this cluster stretch in a

44 wide range and contribute significantly to PR literature, yet each focus potentially proposes some prospective
45
46 directions for researchers that will be discussed in section 6. 47
48 5.4. Summary of Theoretical, Managerial/Practical and Methodological Contributions
49

50 To summarise, the identified theoretical, practical, and methodological contributions provide a rationale 51
52 behind the future directions for researchers.
53

54 To answer RQ1, a summary is derived based on a subject-wise approach for three clusters. Generally, Cluster 1
covers operations management of PR, considering the manufacturing/remanufacturing from product recovery
and resale management when finding solutions for improvements. This cluster can be best treated under three
level headings in business management: (1) operations issues in (re)manufacturing, (2) network design and
product recovery management, and (3) reverse distribution and quality of returned products for reprocessing.
The quantitative models cover network design, product recovery and disposal management (e.g., production
and inventory in both planning and control), facility location, remanufacturing, recycling, financial and quality
aspects in the context of PR. These models enable managers/practitioners to design their business
remanufacturing structure and address their specific dilemma. Furthermore, these initiatives are credible as a
guide to potentially map out the procedures and outcomes for practitioners to follow. It helps them save
resources and reduce the probability of failure in deploying remanufacturing.

9 Turning now to the discussed evidence in Cluster 2, at the retailer and (re)manufacturer level, there is a
10
11 shift in PR from manufacturing/remanufacturing operations to marketing aspects, especially manufacturer-

12 retailer-customer relationships. These relationships can vary depending on the degree of coordination or 13
14 competition and the number of players in a CLSC. In terms of theoretical aspects, papers have been broadened
15
16 to mainly addressing RP issues (both for manufacturer and retailer) in many dimensions such as monetary,

55
56
57
58
59
60
61
62
63 34
64
65
 1
2
3
4
5
6
7
8
17 time and process of RP. These RP issues usually incorporate with channelling, pricing, information, and 18
19 inventory strategies. Regarding managerial/practical aspects, with the wide range of RP issues, practitioners
20
21 decide what is the optimal RP in terms of terms, conditions, and implementation in their customised context.
22
23 Cluster 3 considers customer perspective. In terms of theoretical aspect, this cluster prominently
24
25 concentrates on how customers perceive PR as well as returns management system from retailer. These

26 psychological aspects can be listed as customer loyalty, behaviour, purchase and returns intention in the context
27
28 of PR. By understanding customer’s psychology, practitioners can adjust their operational and marketing
29
30 strategies to mitigate the effect of PR and enhance business performance.
31
32 To answer RQ2, a summary is drawn from the methodological aspect. A wide range of methodologies
33
34 has been used to deal with PR issues. Cluster 1 prominently applies quantitative models (e.g., mixed-integer

35 programming models) and procedures (i.e., mathematical modelling, formulating/optimising, testing, and 36
37 inference). In Cluster 2, the methodologies are more diversified, listing newsvendor models, field experiment,
38
39 meta-analysis, stylized model, and conceptual model. In Cluster 3, the main methodology is

40 theoretical/conceptual framework development using both quantitative and qualitative for finding the 41
42 implications.
43
44 A full detailed summary is in Appendix F. 45
46 6. Discussion and Future Directions
47
48
49 In order to answer RQ3, the suggestions for future directions are structured based on three discussed

50 clusters with a total of five future streams as follow:

51
52
53 (1) Digitialisation in the Context of Product Returns
55
56
57
58
59
60
61
62
63 35
64
65
 1
2
3
4
5
6
7
8
54
There are potential research areas in the context of digitalisation such as digital twin (DT); physical
internet (PI); artificial intelligence; big-data driven simulation and optimisation; Internet of Things (IoT),
Blockchain, Industry 4.0 for strategically implementing and manging a (re)manufacturing system in a
RL/CLSC. So far, most of the research have been on logistics/supply chains but none of them specifically
focuses on RL, not to mention PR. For example, among the aforementioned technologies, a digital twin is the
one that integrates most of the technologies into its operations. It constitutes a digital model for a real-world
object by using sensors and network technologies. Some studies found the extended application to digitalise a
logistics/supply chain network, called the digital logistics/supply chain twin (Ivanov et al., 2019; Ivanov &
Dolgui, 2020) but no one has studied a digital twin on remanufacturing system. In fact, the integration of
technologies in a digital twin would possibly visualise, simulate, and diagnose the status of real-time PR,
9 remanufacturing and RL/CLSC systems.
10
11 The application of digitalisation on remanufacturing returned products can be divided into two potential
12 parts:
13
14
15 Firstly, digitalisation would be applicable for RL/CLSC network design and monitor. A RL/CLSC is

16 increasingly complicated to manage because firms cannot control completely the whole operations of their
17
18 CLSC if they are outsourcing or running multiple suppliers/retailers. Successfully implementing a RL/CLSC
19
20 digitalisation for the whole network would help firms keep up to date with the status of their reverse flow of

21 PR (e.g., Blockchain, IoT). Moreover, it helps obtain a diagnostic and predictable system to detect probable
22
23 root causes in the RL/CLSC. However, the way and areas to which a digital RL/CLSC can be applied are still
24
25 big dilemmas. In the future, researchers could develop a digital system for RL/CLSC in reverse shipment,

26 collection centres, RL/CLSC infrastructure and global RL/CLSC networks.

27
55
56
57
58
59
60
61
62
63 36
64
65
 1
2
3
4
5
6
7
8
28
29 Secondly, digitalisation would be applicable for managing product lifecycle. Conventionally, sellers

30 usually lose track of their products after sales, which induces the uncertainties in returns management
such as
31
32 quality, quantity and time of PR. Digitalisation turns out to be a potential solution to access the product
33
34 footprints that used to be impossible to measure at post-purchase stage. Digitalisation would enable real-time

35 product’s footprints after sales to be visual and measurable for forecasting returns likelihood. By doing this,
36
37 remanufacturing could effectively and efficiently prepare their resources and capabilities as well as reduce
38
39 variations. Nevertheless, the traceability after sales could controversially cause a negative implication for

40 customer’s privacy and transparency. In the future, how to apply digitalisation to obtain product and customer’s
41
42 information after sales need to be considered carefully. 43
44 (2) Globalisation versus Localisation in the Context of Product Returns 45

46 Making the decision between globalisation and localisation modes for a logistics and supply chain system 47
48 is always challenging, especially in the context of PR. Wang et al. (2021) initialised three cross-borders e-
49
50 commerce modes, namely overseas-to-overseas (O2O), overseas-to-domestic (O2D) and domestic-to-
51
52 domestic (D2D) to reflect the trade-off between globalisation and localisation. Although this paper discussed

53 the impact of three cross-borders e-commerce modes in delivery time at pre-purchase stage to returns 54
likelihood of customers, there is no attempt to develop a cross-borders e-commerce in handling the reverse
flow of PR. It raises an issue of how an international corporate resolves a dilemma of reverse flows, for
example, choosing the optimal mode between O2O, O2D and D2D. Furthermore, research on identifying and
tackling the extra challenges related to the events such as pandemic and Brexit in the context of
remanufacturing is noteworthy to dig into. Hence, handling product returns in cross-boarder business would
be a considerable challenge in the future.

55
56
57
58
59
60
61
62
63 37
64
65
 1
2
3
4
5
6
7
8
Some PRs with low salvageable value would be disposed and ended up at landfill as wastes. Global waste
importation is increasingly urgent and controversial in many countries. Large multinational companies abuse
the lenient policies to export their wastes (e.g., plastic recyclables, electronic and paper products) into some
countries (e.g., China, Malaysia, Thailand and African countries). For example, China was the largest landfill
9 of the world, importing nearly 50% of the plastic of the world for three decades (McNaughton and
10
11 Nowakowski, 2019). The low-cost labours and domestic landfill reduction attracted companies to turn these

12 countries into waste hubs in which they are beneficial to set up remanufacturing/recycle facilities nearby. 13
14 However, a growing number of countries are fighting against environmental risks and public health, prompting
15
16 these countries to ban waste importers (BBC News, 2019). This would reshape and challenge the whole global
17 remanufacturing systems and facilities, strictly transforming from globalisation to localisation in dealing with
18
19 end-of-life PRs. In the future, researchers/practitioners need to address the challenges of managing waste
20
21 resources for recycling and remanufacturing returned products domestically.
22
23 (3) ‘Multi-X’ Oriented Bespoke Return Policy (RP)
24

25 Research in RP has been evolving in decades. There exists various forms of RP among manufacturers and 26
27 retailers. However, these proposed RP models are generic which possibly misalign to individual customers.
28
29 With the development of ML/Deep Learning, manufacturers/retailers could develop big data-driven bespoke

30 RPs for individual customers. This type of adaptive RP system can automatically adjust return terms and 31
32 conditions based on the likelihood of each customer return prediction. In addition, the presences of
33
34 digitalisation and globalisation also leads to the complex structure of RL/CLSC which prompts research to

35 extend their focus into a multiple structure. We call this a ‘multi-X’ structure for RP where X is a variable 36
37 refers to:
38
39 The multi-layer structure where multi-manufacturer, multi-retailer, and multi-third-party logistics (3PL)
55
56
57
58
59
60
61
62
63 38
64
65
 1
2
3
4
5
6
7
8
40
41 service providers are attached to RPs. As observed from Appendix D, literatures prominently address RP as a
42
43 single manufacturer – a single retailer; or a single manufacturer – multiple retailers; or multiple manufacturers

44 – a single retailer with both supply chain coordination and competition. Nevertheless, in reality, this type of
45
46 single x vs. multiple y relationship rarely exists. Furthermore, it is not uncommon that many
47
48 manufacturers/retailers are partnering with multi-3PL service providers at the same time due to geographical

49 advantages, packaging, and handling specialties. Taken together, it leads us to propose a potential future 50
51 direction for developing optimal RPs dealing with multi-players in the manner of either coordination or/and
52
53 competition.
54
The above multi-layer for RPs can extend to online platform provider. According to the results from
Appendix D, researchers focus mainly on manufacturers and retailers. Yet, the presence of platform providers
is increasing in quantity (e.g., Amazon, Alibaba, eBay, etc). They supply an infrastructure to display and sell
products directly to consumers, which is especially beneficial for small and medium enterprises who want to
spread their brand awareness globally. However, this business model raises questions about who is responsible
for return policy and handling returns as well as how customers perceive the new RP model. In the future, it is
worth to explore further the effect of the multi-layer RPs in various dimensions towards PR in the context of
sellers and platform providers.

Multi-channel is also an interesting aspect in designing RP. Conventionally, RP research in different

channelling strategies is not a new cluster. In fact, some channelling strategies have been well-discussed such
9 as a single online/offline channel, dual channel and omni channel. While these brick-and-mortar and click-
10
11 and-mortar are still the main channels for trading, the evolution of Industry 4.0 and advanced technologies

12 have reshaped the way retailers operate these channels in many aspects. For example, the diversity of payment
13
14 methods such as pay later, instalments, and mobile/watch payment may dictate differently the monetary aspect
55
56
57
58
59
60
61
62
63 39
64
65
 1
2
3
4
5
6
7
8
15
16 of RP in different channels. Furthermore, operating a multi-channel system including mobile channel is

17 increasingly popular due to its portability and convenience. It raises a question of how to redesign the PR 18
19 process and the level of incurred hassle accordingly in their RP. To summarise, exploring a multi-channel with
20
21 advanced technology in the context of RPs is proposed as a potential approach in the future.
22
23 (4) Understanding and Predicting Customer Returns Behaviour via Online Product Reviews (OPRs) and
24
25 Customer’s Footprint
26
27 Customers are one of the significant roots causing returned products. It is undeniable that customer
28
29 product reviews are valuable sources to expose customer experience and behaviour, especially in the context

30 of PR. However, in researching OPR effects on PR, many papers extracted the sentiment of reviews by using
31
32 customer ratings (Minnema et al., 2016; Sahoo et al., 2018) which is subjective and insufficient to explain
33
34 customer return behaviours. Natural language processing (NLP) techniques (e.g., text mining) which is

35 developing significantly would potentially help researchers embed customer reviews quantitatively to explain
36
37 their behaviour toward returns decision in the future.
38

39 Most literatures studied customer’s perspective on the post-purchase stage. There are little efforts on the 40
41 pre-purchase stage. A possible explanation would be the lack of customer data at this stage. However, as
42
43 customers switch to the online channel, not only they would leave OPRs, but also, they unconsciously leave

44 their footprints. These valuable data can be used to analyse pre-purchase customer behaviour. Researchers and
45
46 practitioners may seek to understand customer behaviour at pre-purchase stage as a signal to predict purchase
47
48 and returns behaviour, thus mitigate negative impacts of PR, especially in the context of e-commerce.
49
55
56
57
58
59
60
61
62
63 40
64
65
 1
2
3
4
5
6
7
8
50 (5) Customer Returns Perception on Marketing – Operations Interface (M-OI) 51
52 M-OI has been recognised as a bridge between marketing and operations activities. The coordination
53
54 between marketing and operations may resolve the mismatch between customer demand stemming from marketing
strategies and supply from operations capabilities. There are some discussions of M-OI such as inventory,
product lead-time information disclosure at the pre-return stage and returns management system at the post-
return stage. They mainly focus on how M-OI is deployed to prevent/manage PR. There is no effort to
understand customer perception on M-OI aspects. In fact, customers may selectively prioritise marketing
information which is necessary for them to make decisions regardless purchasing or returning. Understanding
which marketing-operations collaboration is the most contributed to perceptions of customers helps firms
allocate their resources in a more strategic way. In addition, M-OI could be extended to marketing and
RL/CLSC interface where customers can specify their perceptions on “Green” approaches. It enables firms to
better understand the customer’s attitude and behaviour toward their RL/CLSC operations when they use it as
a marketing strategy. Hence, it is noteworthy to dig into customer’s perceptions toward M-OI and Marketing
9 – RL/CLSC interface.
10
11 Overall, from the results of the content analysis and five future directions, our paper proposes some

12 potential research questions deriving from research gaps for each direction in Table 5. In summary,
Figure 8 13
14 advances Figure 7, showing the overall picture of PR future directions and their applicable areas.
15
16 Table 5. Research Gaps and Potential Research Questions
17

Future Direction Potential Research Gap Some Potential Research Questions

55
56
57
58
59
60
61
62
63 41
64
65
 1
2
3
4
5
6
7
8
18
- What are appropriate tools and technologies for
19 There might be gaps in detecting process and product faults in the
20 adopting cutting-edge context of smart remanufacturing? (e.g., sensors,
technologies (Industry 4.0)
21 block-chain, ML/DL)
for digitalisation to
22 managing PR process and - How computer vision (i.e., virtual reality) can
23 product flows, especially remotely visualise disruptions in reverse
24 under the uncertainties such shipment and collection?
25 Network Design and as disruptions (e.g.,
Monitoring - What extent robots can alternate human at
26 pandemic)
collection points under cost perspectives?
27
28 Digitialisation in Limited papers - How can PI be integrated into a RL using big
29 the Context of acknowledge the role of big data-driven simulation?
30 Product Returns data and digital - What are architectures of DT to build a Smart
31 technologies to design a Remanufacturing System using data-driven
RL/CLSC network
32 simulation?
33
Most papers in this cluster
34 did not consider customers - What and how DT/PI/Blockchain infrastructures
35 as an important element for track and trace the status of products without
36 Managing Product managing remanufacturing violating customer privacy?
37 Lifecycle and operations activities, - Which customer and product information impacts
38 while customers induce the PR timing, quality, and quantity as a basis for a
39 uncertainties of timing, digital remanufacturing system?
40 quality, and quantity of PR.
41
42
43
44
45
46
47
48
49
50
51
52
53
54

55
56
57
58
59
60
61
62
63 42
64
65
 Many papers addressed the
impact of the international
1 or domestic channelling - What are cost benefits of O2O, O2D and D2D?
2 choice on PR decision and
3 E-commerce for RL - What are architectures of O2O, O2D and D2D for
intention. No one has
4 addressed how to design an RL?
5 e-commerce channel for
6 Globalisation reverse flows.
7 versus Localisation Some papers addressed
8 in the Context of environmental aspects of
9 Product Returns emission such as CO2
10 (Mishra and Singh, 2020a, - Where is the strategic location for setting up
disposal and remanufacturing centres?
11 Environmental 2020b). There might be a
12 Sustainability gap in considering - What are social and economic considerations for
13 economic, social, and setting up local and global facilities?
geographical factors for
14
dealing with wastes after
15 returns.
16
17 - What is the optimal RP for multi-layer structure
18 A structure of that optimises restocking fees?
multimanufacturer-3PL-
19 - What is the optimal buy-back contract for
Multi-layer structure retailer RP is still
20 understudied due to its
multilayer structure?
21 complexity. - What are the optimal return deadlines for each
22 party in a multi-layer structure?
23
24 In modern world, a
coordination between - Should sellers adopt platform provider’s RP in
25
sellers and platform terms of cost perspectives?
26 ‘Multi-X’ Oriented Online platform
providers is increasingly
27 Bespoke Return provider
common. However, a study - Which factors of RP are more beneficial for
28 Policy (RP) about RP in this context is sellers and platform providers?
29 still missing.
30
Single, dual and
31
omnichannel have been well
32 discussed. However, with - What is the optimal coordination of channels
33 the development of new under current RP of sellers?
34 Multi-channel technologies (i.e., smart
35 phone, tablet, smart watch), - What extent the rigorousness and leniency of RP
36 RP studies with these under multi-channel of a seller?
37 channels are still
38 understudied.
39 Several attempts tried to
40 understand online customer
Understanding return behaviour (OCRB)
41 and Predicting
42 using valence, volume of - By using text mining and customer footprints,
Customer Returns
customer reviews what drivers associate with high return
43 Behaviour via OPRs and Online
(Minnema et al., 2016; intentions?
44 Online Product Footprints
Sahoo et al., 2018). There
45 Reviews (OPRs) - Can these drivers be sufficient to predict OCRB?
and Customer’s might be a gap in analysing
46 textual information and
Footprint
47 customer footprints (e.g.,
48 website logs) using NLP.
55
56
57
58
59
60
61
62
63 43
64
65
 1
2
3
4
5
There may be a link
6
between pre-purchase and
7 post-purchase behaviours
8 (Dowling et al., 2020).
49 - Which signals from customers at the pre-purchase
Pre-purchase OCRB is a part of
stage associate with OCRB at the post-purchase
50 behaviour postpurchase behaviour. A
stage?
51 study for understanding
52 prepurchase behaviours to
53 navigate OCRB is still
54 missing.
9 Designing M-OI system
10 was mainly focussed by
11 researchers (Bijmolt et al.,
12 2021; Mollenkopf et al.
Customer Returns Perception on 2011). How customer return
13 Marketing – Operations Interface (MOI) perspective is influenced by
- Which informational drivers of M-OI leading to
14 information given by a MOI the variation of returns intention?
15 (or extending to
16 Marketing-RL/CLSC
17 Interface) is still
understudied.
18
19

55
56
57
58
59
60
61
62
63 44
64
65
 1
2
3
4
5
6
7
8
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49 50
51
52 Figure 8. Applicable areas for PR future directions
53
54
7. Conclusion

The aim of this paper is to review the state-of-art research in product returns and provide future research
directions. A total of 1209 publications were collected from three large scientific databases, Scopus,
55
56
57
58
59
60
61
62
63 45
64
65
 1
2
3
4
5
6
7
8
EBSCOhost and WoS. Following the review framework, machine learning clustering model and bibliometric
analysis are undertaken. Three broad categories are labelled as Cluster 1 – operations management of PR
(39.53%), Cluster 2 – retailer and (re)manufacturer (34.57%), Cluster 3 – customer’s psychology and
experience (25.9%), and 167 papers are selected for content analysis. The results from these analyses are used
to derive the answer to the key research question stated at the beginning.

This study has provided the overall picture of PR themes. The contributions can be summarised as
following. Firstly, the investigation of current PR literatures helps researchers locate their PR research in our
9
10 map of PR literatures. They will be able to access to relevant research to supplement their own research area.

11 In terms of practical implications, practitioners could position their circumstances into the map to strategically
12
13 apply these initiatives for their business problems. Secondly, a wide range of identified methodologies will
14
15 serve as an advice for researchers to appropriately inherit or develop their own methodology. Finally, the

16 carefully proposed five research directions lay the groundwork for future research on PR.
17

18 This study also presents some limitations. Firstly, choosing keywords for search terms cannot be 19
20 completely exhaustive. Secondly, our chosen cluster of PR has many synonyms, for example, return behaviour
21
22 or consumer return also means as return on investment in finance which is irrelevant to our study. Even though,

23 we made significant efforts to remove them. We believe there might be still a certain number of irrelevant 24
25 papers available in our dataset which is inevitable if you obtain a large dataset.
26
27
28
29
30
31
32
33
34
55
56
57
58
59
60
61
62
63 46
64
65
 1
2
3
4
5
6
7
8
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54

References

Abdulla, H., Ketzenberg, M., Abbey, J.D., 2019. Taking stock of consumer returns: A review and classification
of the literature. J. Oper. Manag. 65, 560–605. [Link]
Ai, X., Chen, J., Zhao, H., Tang, X., 2012. Competition among supply chains: Implications of full returns
policy. Int. J. Prod. Econ. 139, 257–265. [Link]
AJG, 2018. Academic Journal Guide (2018) by Chartered ABS [WWW Document]. [Link]. URL
9 [Link] (accessed 1.20.21).
10
11 Akçay, Y., Boyacbox, T., Zhang, D., 2013. Selling with money-back guarantees: The impact on prices,

55
56
57
58
59
60
61
62
63 47
64
65
 1
2
3
4
5
6
7
8
12 quantities, and retail profitability. Prod. Oper. Manag. 22, 777–791. [Link]
14 5956.2012.01394.x
15
16 Akturk, M.S., Ketzenberg, M., Yıldız, B., 2021. Managing consumer returns with technology-enabled
17 countermeasures. Omega (United Kingdom) 102. [Link] 18
19 Alumur, S.A., Nickel, S., Saldanha-Da-Gama, F., Verter, V., 2012. Multi-period reverse logistics network
20
21 design. Eur. J. Oper. Res. 220, 67–78. [Link]
22 Ambilkar, P., Dohale, V., Gunasekaran, A., Bilolikar, V., 2021. Product returns management: a
comprehensive 23
24 review and future research agenda. Int. J. Prod. Res. [Link]
25
26 Anderson, E.T., Hansen, K., Simester, D., 2009. The option value of returns: Theory and
empirical evidence.

27 Mark. Sci. 28, 405–423. [Link] 28

29 Aras, N., Boyaci, T., Verter, V., 2004. The effect of categorizing returned products in remanufacturing. IIE
30
31 Trans. (Institute Ind. Eng. 36, 319–331. [Link]

32 Assid, M., Gharbi, A., Hajji, A., 2020. Production control of failure-prone manufacturing-
remanufacturing 33
34 systems using mixed dedicated and shared facilities. Int. J. Prod. Econ. 224, 107549.
35
36 [Link]
37
38 Autry, C.W., 2005. Formalization of reverse logistics programs: A strategy for managing liberalized
returns.
39 Ind. Mark. Manag. 34, 749–757. [Link]
40
41 Bandyopadhyay, S., Paul, A.A., 2010. Equilibrium returns policies in the presence of supplier competition.
42
43 Mark. Sci. 29, 846–857. [Link]

55
56
57
58
59
60
61
62
63 48
64
65
 1
2
3
4
5
6
7
8
44 Barclaycard, 2018. Return to Sender: Retailers face a ‘Phantom Economy’ of £7bn each year as shopper
returns 45
46 continue to rise [WWW Document]. Barclay. URL [Link]
47
48 releases/2018/06/return-to-sender/
49 Bastian, M., Heymann, S., Jacomy, M., 2009. Gephi: An open source software for exploring and
manipulating
50
51 networks, in: International AAAI Conference on Weblogs and Social Media.
52
53 BBC News, 2019. Why some countries are shipping back plastic waste [WWW Document]. [Link]. URL 54

[Link] (accessed 2.15.21).

Bechwati, N.N., Siegal, W.S., 2005. The impact of the prechoice process on product returns. J. Mark. Res. 42,
358–367. [Link]
Bijmolt, T.H.A., Broekhuis, M., de Leeuw, S., Hirche, C., Rooderkerk, R.P., Sousa, R., Zhu, S.X., 2021.
Challenges at the marketing–operations interface in omni-channel retail environments. J. Bus. Res. 122,
864–874. [Link]

Blei, D.M., Ng, A.Y., Jordan, M.I., 2003. Latent Dirichlet allocation. J. Mach. Learn. Res.
[Link]
Bonifield, C., Cole, C., Schultz, R.L., 2010. Product returns on the Internet: A case of mixed signals? J. Bus.
Res. 63, 1058–1065. [Link]
Borenich, A., Dickbauer, Y., Reimann, M., Souza, G.C., 2020. Should a manufacturer sell refurbished returns
9 on the secondary market to incentivize retailers to reduce consumer returns? Eur. J. Oper. Res. 282, 569–
10 579. [Link]
11
12 Börner, K., Chen, C., Boyack, K.W., 2003. Visualizing knowledge domains. Annu. Rev. Inf. Sci. Technol.
13
14 [Link]

55
56
57
58
59
60
61
62
63 49
64
65
 1
2
3
4
5
6
7
8
15 Bower, A.B., Maxham, J.G., 2012. Return shipping policies of online retailers: Normative assumptions
and
16
17 the long-term consequences of fee and free returns. J. Mark. 76, 110–124.
18
19 [Link]
20 Boyack, K.W., Klavans, R., 2010. Co-citation analysis, bibliographic coupling, and direct citation: Which
21
22 citation approach represents the research front most accurately? J. Am. Soc. Inf. Sci. Technol.
23
24 [Link]
25 Casalin, F., Dia, E., 2019. Information and reputation mechanisms in auctions of remanufactured goods.
Int.
26
27 J. Ind. Organ. 63, 185–212. [Link]
28
29 Chae, B., Olson, D., 2021. A Topical Exploration of the Intellectual Development of Decision Sciences
1975–
30 2016, in: Decision Sciences. pp. 543–566. [Link]
31
32 Che, Y.-K., 1996. Customer Return Policies for Experience Goods. J. Ind. Econ. 44, 17.
33
34 [Link]
35 Chen, B., Chen, J., 2017. When to introduce an online channel, and offer money back guarantees and 36
37 personalized pricing? Eur. J. Oper. Res. 257, 614–624. [Link]
38
39 Chen, J., 2011. The impact of sharing customer returns information in a supply chain with and without a
40 buyback policy. Eur. J. Oper. Res. 213, 478–488. [Link] 41
42 Chen, J., Bell, P.C., 2012. Implementing market segmentation using full-refund and no-refund customer
43
44 returns policies in a dual-channel supply chain structure. Int. J. Prod. Econ.
136, 56–66.
45 [Link] 46
55
56
57
58
59
60
61
62
63 50
64
65
 1
2
3
4
5
6
7
8
47 Chen, J., Bell, P.C., 2009. The impact of customer returns on pricing and order decisions. Eur. J. Oper. Res.
48
49 195, 280–295. [Link]

50 Chen, J., Grewal, R., 2013. Competing in a supply chain via full-refund and no-refund customer returns
51
52 policies. Int. J. Prod. Econ. 146, 246–258. [Link]
53
54 Chen, Z., Fan, Z.P., Zhao, X., 2021. Offering return-freight insurance or not: Strategic analysis of an e-seller’s

decisions. Omega (United Kingdom) 103. [Link]

Choi, T.M., Li, D., Yan, H., 2004. Optimal returns policy for supply chain with e-marketplace. Int. J. Prod.
Econ. 88, 205–227. [Link]
Chou, M.C., Sim, C.K., Yuan, X.M., 2020. Policies for inventory models with product returns forecast from
past demands and past sales. Ann. Oper. Res. 288, 137–180. [Link]
Ciano, M.P., Pozzi, R., Rossi, T., Strozzi, F., 2019. How IJPR has addressed ‘lean’: a literature review using
bibliometric tools. Int. J. Prod. Res. [Link]
Clottey, T., Benton, W.C., Srivastava, R., 2012. Forecasting Product Returns for Remanufacturing Operations.
Decis. Sci. 43, 589–614. [Link]
9 Crocker, K.J., Letizia, P., 2014. Optimal policies for recovering the value of consumer returns. Prod.
Oper.
10 Manag. 23, 1667–1680. [Link]
11
12 Cui, H., Rajagopalan, S., Ward, A.R., 2020. Predicting product return volume using machine learning methods.
13
14 Eur. J. Oper. Res. 281, 612–627. [Link]

15 Daryanani, S., Miller, D.R., 1992. Calculation of Steady-State Probabilities for Repair Facilities with
Multiple
16
17 Sources and Dynamic Return Priorities. Oper. Res. 40, S248–S256.
18
19 [Link]

55
56
57
58
59
60
61
62
63 51
64
65
 1
2
3
4
5
6
7
8
20 Davis, S., Gerstner, E., Hagerty, M., 1995. Money back guarantees in retailing: matching products to
consumer
21
22 tastes. J. Retail. 71, 7–22. [Link]
23
24 de Brito, M.P., van der Laan, E.A., 2009. Inventory control with product returns: The impact of imperfect

25 information. Eur. J. Oper. Res. 194, 85–101. [Link]

26
27 De, P., Hu, Y.J., Rahman, M.S., 2013. Product-oriented web technologies and product returns: An exploratory
28
29 study. Inf. Syst. Res. 24, 998–1010. [Link]
30 DeCroix, G.A., Zipkin, P.H., 2005. Inventory management for an assembly system with product or
component
31
32 returns. Manage. Sci. 51, 1250–1265. [Link]
33
34 Deerwester, S., Dumais, S.T., Furnas, G.W., Landauer, T.K., Harshman, R., 1990. Indexing by latent semantic
35 analysis. J. Am. Soc. Inf. Sci. [Link]
37 ASI1>[Link];2-9
38
39 Dimitrov, S., Ceryan, O., 2019. Optimal inventory decisions when offering layaway. Int.
J. Prod. Res. 57,
40 1161–1175. [Link] 41
42 Dohleman, B.S., 2006. Exploratory social network analysis with Pajek. Psychometrika 71, 605–606.
43
44 [Link]
45 Dowling, K., Guhl, D., Klapper, D., Spann, M., Stich, L., Yegoryan, N., 2020. Behavioral biases in
marketing.
46
47 J. Acad. Mark. Sci. [Link]
48
49 Dumais, S.T., Furnas, G.W., Landauer, T.K., Deerwester, S., Harshman, R., 1988. Using
latent semantic

55
56
57
58
59
60
61
62
63 52
64
65
 1
2
3
4
5
6
7
8
50 analysis to improve information retrieval, in: Proceedings of CHI. 51
52 El Kihal, S., Nurullayev, N., Schulze, C., Skiera, B., 2021. A Comparison of Return Rate Calculation Methods:
53
54 Evidence from 16 Retailers. J. Retail. [Link]
Emmons, H., Gilbert, S.M., 1998. Note. The role of returns policies in pricing and inventory decisions for
catalogue goods. Manage. Sci. 44, 276–283. [Link]
Ertekin, N., Shulman, J.D., Chen, H.A., 2019. On the profitability of stacked discounts: Identifying revenue
and cost effects of discount framing. Mark. Sci. 38, 317–342. [Link]
Fan, Z.P., Chen, Z., 2020. When should the e-tailer offer complimentary return-freight insurance ? Int. J. Prod.
Econ. 230, 107890. [Link]
Fang, B., Fu, X., Liu, S., Cai, S., 2021. Post-purchase warranty and knowledge monetization: Evidence from
a paid-knowledge platform. Inf. Manag. 58. [Link]
Ferguson, M., Guide, V.D.R., Souza, G.C., 2006. Supply chain coordination for false failure returns. Manuf.
Serv. Oper. Manag. [Link]
9 Fleischmann, M., Beullens, P., Bloemhof-Ruwaard, J.M., Van Wassenhove, L.N., 2001. The impact of
product

10 recovery on logistics network design. Prod. Oper. Manag. 10, 156–173. [Link]
11
12 5956.2001.tb00076.x
13
14 Frei, R., Jack, L., Brown, S., 2020. Product returns: a growing problem for business, society and
environment.
15 Int. J. Oper. Prod. Manag. 40, 1613–1621. [Link]
16
17 Gäthke, J., Gelbrich, K., Chen, S., 2021. A Cross-National Service Strategy to Manage Product Returns: E-
18
19 Tailers’ Return Policies and the Legitimating Role of the Institutional Environment. J. Serv. Res.
20 [Link]
21
22 GlobalData, 2018. Value of online returns to rise to £5.6bn by 2023, driven by clothing & footwear [WWW
55
56
57
58
59
60
61
62
63 53
64
65
 1
2
3
4
5
6
7
8
23
24 Document]. GlobalData.
25 Goltsos, T.E., Ponte, B., Wang, S., Liu, Y., Naim, M.M., Syntetos, A.A., 2019. The boomerang returns?
26
27 Accounting for the impact of uncertainties on the dynamics of remanufacturing systems. Int. J. Prod. Res.
28
29 57, 7361–7394. [Link]

30 Goodall, P., Sharpe, R., West, A., 2019. A data-driven simulation to support remanufacturing operations.
31
32 Comput. Ind. 105, 48–60. [Link]
33
34 Govindan, K., Soleimani, H., Kannan, D., 2015. Reverse logistics and closed-loop supply chain: A
35 comprehensive review to explore the future. Eur. J. Oper. Res. [Link] 36
37 Griffis, S.E., Rao, S., Goldsby, T.J., Niranjan, T.T., 2012. The customer consequences of returns in online
38
39 retailing: An empirical analysis. J. Oper. Manag. 30, 282–294. [Link]
40 Guide, V.D.R., 2000. Production planning and control for remanufacturing: Industry practice and
research 41
42 needs. J. Oper. Manag. 18, 467–483. [Link]
43
44 Guide, V.D.R., Jayaraman, V., 2000. Product acquisition management: Current industry practice and a
45 proposed framework. Int. J. Prod. Res. 38, 3779–3800. [Link] 46
47 Guo, H., Zhang, Y., Zhang, C., Liu, Y., Zhou, Y., 2020a. Location-inventory decisions for closed-loop supply
48
49 chain management in the presence of the secondary market. Ann. Oper. Res. 291, 361–386.

50 [Link]
51
52 Guo, H., Zhang, Ying, Zhang, C., Zhang, Yidong, Han, Z., 2020b. A multi-commodity location-inventory
53
54 problem in a closed-loop supply chain with commercial product returns. Int. J. Prod. Res. 58, 6899–6916.
[Link]

55
56
57
58
59
60
61
62
63 54
64
65
 1
2
3
4
5
6
7
8
Heiman, A., McWilliams, B., Zhao, J., Zilberman, D., 2002. Valuation and management of money-back
guarantee options. J. Retail. 78, 193–205. [Link]
Hess, J.D., Chu, W., Gerstner, E., 1996. Controlling product returns in direct marketing. Mark. Lett. 7, 307–
317. [Link]
Hjørland, B., 2013. Citation analysis: A social and dynamic approach to knowledge organization. Inf. Process.
Manag. 49, 1313–1325. [Link]
Hofmann, T., 1999. Probabilistic latent semantic indexing, in: Proceedings of the 22nd Annual International
ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 1999. pp. 50–
57. [Link]
9 Hosseini-Motlagh, S.M., Nematollahi, M., Ebrahimi, S., 2020. Tri-party reverse supply chain
coordination

10 with competitive product acquisition process. J. Oper. Res.

Soc.
11
12 [Link]
13
14 Hsiao, L., Chen, Y.J., 2015. Retailer’s rationale to refuse consumer returns in supply chains. Nav. Res.
Logist.
15 62, 686–701. [Link]
16
17 Hsiao, L., Chen, Y.J., 2014. Return policy: Hassle-free or your money-back guarantee? Nav. Res. Logist. 61,
18
19 403–417. [Link]

20 Hsiao, L., Chen, Y.J., 2012. Returns policy and quality risk in E-business. Prod. Oper. Manag. 21, 489–
503.
21
22 [Link]
23
24 Huang, S., Potter, A., Eyers, D., 2020. Social media in operations and supply chain management: State-
of-the-

25 Art and research directions. Int. J. Prod. Res. 58, 1893–1925.

55
56
57
58
59
60
61
62
63 55
64
65
 1
2
3
4
5
6
7
8
26
27 [Link]
28
29 Huang, X., Zhang, D., 2020. Service product design and consumer refund policies. Mark. Sci. 39, 366–
381.
30 [Link]
31
32 Inderfurth, K., 1997. Simple optimal replenishment and disposal policies for a product recovery system with
33
34 leadtimes. OR Spectr. 19, 111–122. [Link]
35 Ivanov, D., Dolgui, A., 2021. A digital supply chain twin for managing the disruption risks and resilience
in 36
37 the era of Industry 4.0. Prod. Plan. Control 32, 775–788.
38
39 [Link]
40 Ivanov, D., Dolgui, A., Das, A., Sokolov, B., 2019. Digital Supply Chain Twins: Managing the Ripple
Effect, 41
42 Resilience, and Disruption Risks by Data-Driven Optimization, Simulation, and Visibility, in:
43
44 International Series in Operations Research and Management Science.
pp. 309–332.

45 [Link]
46
47 Janakiraman, N., Syrdal, H.A., Freling, R., 2016. The Effect of Return Policy Leniency on Consumer Purchase
48
49 and Return Decisions: A Meta-analytic Review. J. Retail. 92,
226–235.
50 [Link]
51
52 Jayaraman, V., Patterson, R.A., Rolland, E., 2003. The design of reverse distribution networks: Models and
53
54 solution procedures. Eur. J. Oper. Res. 150, 128–149. [Link]

55
56
57
58
59
60
61
62
63 56
64
65
 1
2
3
4
5
6
7
8
Jiang, Y., Shi, C., Shen, S., 2019. Service Level Constrained Inventory Systems. Prod. Oper. Manag. 28, 2365–
2389. [Link]
Jin, D., Caliskan-Demirag, O., Chen, F. (Youhua), Huang, M., 2020. Omnichannel retailers’ return policy
strategies in the presence of competition. Int. J. Prod. Econ. 225.
[Link]

Jung, Y., Suh, Y., 2019. Mining the voice of employees: A text mining approach to identifying and analyzing
job satisfaction factors from online employee reviews. Decis. Support
Syst. 123. [Link]
Kazemi, N., Modak, N.M., Govindan, K., 2019. A review of reverse logistics and closed loop supply chain
management studies published in IJPR: a bibliometric and content analysis. Int. J. Prod. Res. 9
[Link]
10 Ke, C., Yan, B., 2020. Trade-in value effects of used products in remanufacturing with considering consumer
11
12 purchase behavior. J. Oper. Res. Soc. [Link]
13
14 Kessler, M.M., 1963. Bibliographic coupling between scientific papers.
Am. Doc.

15 [Link]
16
17 Ketzenberg, M.E., Abbey, J.D., Heim, G.R., Kumar, S., 2020. Assessing customer return behaviors through
18
19 data analytics. J. Oper. Manag. 66, 622–645. [Link]

20 Khouja, M., Ajjan, H., Liu, X., 2019. The effect of return and price adjustment policies on a retailer’s 21
22 performance. Eur. J. Oper. Res. 276, 466–482. [Link]
23
24 Krikke, H.R., Van Harten, A., Schuur, P.C., 1999. Business case Océ: Reverse logistic network re-design
for

25 copiers. OR Spektrum 21, 381–409. [Link]

26

55
56
57
58
59
60
61
62
63 57
64
65
 1
2
3
4
5
6
7
8
27 Krug, Z., Guillaume, R., Battaïa, O., 2021. Exploring the opportunities in establishing a closed-loop supply
28
29 chain under uncertainty. Int. J. Prod. Res. 59, 1606–1625.

30 [Link]
31
32 Kumar, A., Mehra, A., Kumar, S., 2019. Why do stores drive online sales? Evidence of underlying mechanisms
33
34 from a multichannel retailer. Inf. Syst. Res. 30, 319–338. [Link]
35 Kumar, S., Putnam, V., 2008. Cradle to cradle: Reverse logistics strategies and opportunities across three
36
37 industry sectors. Int. J. Prod. Econ. 115, 305–315. [Link]
38
39 Laengle, S., Merigó, J.M., Miranda, J., Słowiński, R., Bomze, I., Borgonovo, E., Dyson, R.G., Oliveira, J.F., 40

Teunter, R., 2017. Forty years of the European Journal of Operational Research: A bibliometric overview.
41
42 Eur. J. Oper. Res. [Link]
43
44 Lal, R., Sarvary, M., 1999. When and how is the internet likely to decrease price
competition? Mark. Sci. 18,
45 485–503. [Link] 46
47 Lee, C.H., 2001. Coordinated stocking, clearance sales, and return policies for a supply chain. Eur. J. Oper.
48
49 Res. 131, 491–513. [Link]
50 Lee, C.H., Choi, T.M., Edwin Cheng, T.C., 2021. Operations strategies with snobbish and strategic
consumers.
51
52 Nav. Res. Logist. 68, 327–343. [Link]
53
54 Lee, C.H., Rhee, B.D., 2021. Retailer-run resale market and supply chain coordination. Int. J. Prod. Econ.
235. [Link]
Lee, S., Yi, Y., 2019. “Retail is detail! Give consumers a gift rather than a bundle”: Promotion framing and
consumer product returns. Psychol. Mark. 36, 15–27. [Link]
55
56
57
58
59
60
61
62
63 58
64
65
 1
2
3
4
5
6
7
8
Lepthien, A., Clement, M., 2019. Shipping fee schedules and return behavior. Mark. Lett. 30, 151–165.
[Link]
Li, G., Li, L., Sethi, S.P., Guan, X., 2019. Return strategy and pricing in a dual-channel supply chain. Int. J.
Prod. Econ. 215, 153–164. [Link]
Li, M., Choudhury, A.H., 2021. Using website information to reduce postpurchase dissonance: A mediated
moderating role of perceived risk. Psychol. Mark. 38, 56–69. [Link]
Li, W., Chen, J., Chen, B., 2018a. Supply chain coordination with customer returns and retailer’s store brand
9 product. Int. J. Prod. Econ. 203, 69–82. [Link]

10 Li, W., Chen, J., Liang, G., Chen, B., 2018b. Money-back guarantee and personalized pricing in a Stackelberg
11
12 manufacturer’s dual-channel supply chain. Int. J. Prod. Econ. 197, 84–98.
13
14 [Link]
15 Li, Y., Li, G., Tayi, G.K., Cheng, T.C.E., 2021. Return shipping insurance: Free versus for-a-fee? Int. J.
Prod.
16
17 Econ. 235. [Link]
18
19 Li, Y., Xu, L., Li, D., 2013. Examining relationships between the return policy, product quality, and
pricing

20 strategy in online direct selling. Int. J. Prod. Econ. 144, 451–460.

21
22 [Link]
23
24 Li, Z.C., Huang, H.J., Yang, H., 2020. Fifty years of the bottleneck model: A bibliometric review and
future

25 research directions. Transp. Res. Part B Methodol. 139, 311–342.

26
27 [Link]
28
29 Lin, J., Zhang, J., Cheng, T.C.E., 2020. Optimal pricing and return policy and the value of freight
insurance
55
56
57
58
59
60
61
62
63 59
64
65
 1
2
3
4
5
6
7
8
30 for a retailer facing heterogeneous consumers with uncertain product values. Int. J. Prod. Econ. 229,
31
32 107767. [Link]
33
34 Lin, Y., Leung, J.M.Y., Zhang, L., Gu, J.W., 2020. Single-item repairable inventory system with
stochastic

35 new and warranty demands. Transp. Res. Part E Logist.

Transp. Rev. 142, 17.
36
37 [Link]
38
39 Liu, J., Mantin, B., Wang, H., 2014. Supply chain coordination with customer returns and refund-
dependent

40 demand. Int. J. Prod. Econ. 148, 81–89. [Link] 41

42 Liu, X., Zhou, Y.W., Shen, Y., Ge, C., Jiang, J., 2021. Zooming in the impacts of merchants’ participation in
43
44 transformation from online flash sale to mixed sale e-commerce platform. Inf. Manag. 58.
45 [Link]
46
47 Liu, Z., Lu, L., Qi, X., 2020. The showrooming effect on integrated dual channels. J. Oper. Res. Soc. 71, 1347–
48
49 1356. [Link]
50 Ma, B., Di, C., Hsiao, L., 2020. Return Window Decision in A Distribution Channel. Prod. Oper. Manag.
29, 51
52 2121–2137. [Link]
53
54 Maaya, L., Meulders, M., Vandebroek, M., 2020. Online Consumers’ Attribute Non-Attendance
Behavior: Effects of Information Provision. Int. J. Electron. Commer.
24, 338–365. [Link]
Mahadevan, B., Pyke, D.F., Fleischmann, M., 2003. Periodic review, push inventory policies for
remanufacturing. Eur. J. Oper. Res. 151, 536–551. [Link]
Mahmoudzadeh, M., 2020. On the Non-Equivalence of Trade-ins and Upgrades in the Presence of
55
56
57
58
59
60
61
62
63 60
64
65
 1
2
3
4
5
6
7
8
Framing Effect: Experimental Evidence and Implications for Theory. Prod. Oper. Manag. 29, 330–352.
[Link]
Mandal, P., Basu, P., Saha, K., 2021. Forays into omnichannel: An online retailer’s strategies for managing
product returns. Eur. J. Oper. Res. 292, 633–651. [Link]
Marín, A., Pelegrín, B., 1998. The return plant location problem: Modelling and resolution. Eur. J. Oper. Res.
9 104, 375–392. [Link]

10 Marshakova, I., 1973. System of Document Connections Based on References. [Link].

11
12 Matsui, K., 2010. Returns policy, new model introduction, and consumer welfare. Int. J. Prod. Econ. 124, 299–
13
14 309. [Link]
15 Mayring, P., 2010. Qualitative Inhaltsanalyse, in: Handbuch Qualitative Forschung in Der Psychologie.
pp.
16
17 601–613. [Link]
18
19 McNaughton, S., Nowakowski, K., 2019. How China’s plastic waste ban forced a global recycling
reckoning
20 [WWW Document]. Natl. Geogr. Mag. URL
21
22 [Link]
23
24 worldwide/ (accessed 2.15.21).
25 McWilliams, B., 2012. Money-back guarantees: Helping the low-quality retailer. Manage. Sci. 58, 1521–
1524.
26
27 [Link]
28
29 Menke, W.W., 1969. Determination of Warranty Reserves. Manage. Sci. 15, B-
542-B-549.
30 [Link] 31

55
56
57
58
59
60
61
62
63 61
64
65
 1
2
3
4
5
6
7
8
32 Merigó, J.M., Yang, J.B., 2017. A bibliometric analysis of operations research and management science.
33
34 Omega (United Kingdom). [Link]
35 Min, H., Ko, H.J., 2008a. The dynamic design of a reverse logistics network from the perspective of
third36
37 party logistics service providers. Int. J. Prod. Econ. 113, 176–192.
38
39 [Link]
40 Min, H., Ko, H.J., 2008b. The dynamic design of a reverse logistics network from the perspective of
third41
42 party logistics service providers. Int. J. Prod. Econ. 113, 176–192.
43
44 [Link]
45 Minnema, A., Bijmolt, T.H.A., Gensler, S., Wiesel, T., 2016. To Keep or Not to Keep: Effects of Online
46
47 Customer Reviews on Product Returns. J. Retail. 92, 253–267.
48
49 [Link]
50 Minner, S., Kleber, R., 2001. Optimal control of production and remanufacturing in a simple recovery
model 51
52 with linear cost functions. OR Spektrum 23, 3–24. [Link]
53
54 Mishra, S., Singh, S.P., 2020a. A stochastic disaster-resilient and sustainable reverse logistics model in big data

environment. Ann. Oper. Res. [Link]

Mishra, S., Singh, S.P., 2020b. Designing dynamic reverse logistics network for post-sale service. Ann. Oper.
Res. 30. [Link]
Modak, N.M., Sinha, S., Raj, A., Panda, S., Merigó, J.M., Lopes de Sousa Jabbour, A.B., 2020. Corporate
social responsibility and supply chain management: Framing and pushing forward the debate. J. Clean.
Prod. [Link]
Mollenkopf, D.A., Frankel, R., Russo, I., 2011. Creating value through returns management: Exploring the
marketing-operations interface. J. Oper. Manag. 29, 391–403. [Link]
55
56
57
58
59
60
61
62
63 62
64
65
 1
2
3
4
5
6
7
8
Mollenkopf, D.A., Rabinovich, E., Laseter, T.M., Boyer, K.K., 2007. Managing internet product returns: A
focus on effective service operations. Decis. Sci. 38, 215–250. [Link]
9 5915.2007.00157.x
10 Moreau, C.P., 2020. Brand Building on the Doorstep: The Importance of the First (Physical) Impression.
J.
11
12 Retail. 96, 155–167. [Link]
13
14 Mukhopadhyay, S.K., Setaputra, R., 2007. A dynamic model for optimal design quality and return
policies.
15 Eur. J. Oper. Res. 180, 1144–1154. [Link]
16
17 Mukhopadhyay, S.K., Setoputro, R., 2005. Optimal return policy and modular design for build-to-order
18
19 products. J. Oper. Manag. 23, 496–506. [Link]

20 Mutha, A., Bansal, S., Guide, V.D.R., 2021. Managing the Inter-Functional Tension between
Accounting- and 21
22 Financial-Profits in Remanufacturing Multiple-Usecycle Products. Prod. Oper. Manag.
23
24 [Link]
25 Muyldermans, L., Van Wassenhove, L.N., Guide, V.D.R., 2019. Managing high-end ex-demonstration
product
26
27 returns. Eur. J. Oper. Res. 277, 195–214. [Link]
28
29 Nageswaran, L., Cho, S.-H., Scheller-Wolf, A.A., 2017. Consumer Return Policies in
Omnichannel
30 Operations. SSRN Electron. J. [Link] 31
32 Narang, U., Yadav, M.S., Rindfleisch, A., 2021. EXPRESS: The “Idea Advantage”: How Content Sharing
33
34 Strategies Impact Engagement in Online Learning Platforms. J. Mark. Res.
002224372110178.

55
56
57
58
59
60
61
62
63 63
64
65
 1
2
3
4
5
6
7
8
35 [Link] 36
37 Narlikar, L., Ovcharenko, I., 2009. Identifying regulatory elements in eukaryotic genomes. Briefings Funct.
38
39 Genomics Proteomics. [Link]
40 Ndhaief, N., Nidhal, R., Hajji, A., Bistorin, O., 2020. Environmental issue in an integrated production
and 41
42 maintenance control of unreliable manufacturing/remanufacturing systems. Int. J. Prod. Res. 58, 4182–
43
44 4200. [Link]
45 Netzer, O., Lemaire, A., Herzenstein, M., 2019. When Words Sweat: Identifying Signals for Loan Default
in 46
47 the Text of Loan Applications. J. Mark. Res. 56, 960–980. [Link]
48
49 Niknejad, A., Petrovic, D., 2014. Optimisation of integrated reverse logistics networks with different
product

50 recovery routes. Eur. J. Oper. Res. 238, 143–154. [Link] 51

52 Ofek, E., Katona, Z., Sarvary, M., 2011. “Bricks and clicks”: The impact of product returns on the strategies
53
54 of multichannel retailers. Mark. Sci. 30, 42–60. [Link]
Padmanabhan, V., Png, I.P.L., 1997. Manufacturer’s returns policies and retail competition. Mark. Sci. 16, 81–
94. [Link]
Papanagnou, C.I., 2021. Measuring and eliminating the bullwhip in closed loop supply chains using control
theory and Internet of Things. Ann. Oper. Res. [Link]
Pasternack, B.A., 1985. Optimal Pricing and Return Policies for Perishable Commodities. Mark. Sci. 4, 166–
176. [Link]
Patankar, J.G., Mitra, A., 1995. Effects of Warranty Execution on Warranty Reserve Costs. Manage. Sci. 41,
395–400. [Link]
Petersen, J.A., Kumar, V., 2015. Perceived risk, product returns, and optimal resource allocation: Evidence
from a field experiment. J. Mark. Res. 52, 268–285. [Link]

55
56
57
58
59
60
61
62
63 64
64
65
 1
2
3
4
5
6
7
8
9 Pilkington, A., Meredith, J., 2009. The evolution of the intellectual structure of operations management-
1980-

10 2006: A citation/co-citation analysis. J. Oper. Manag. 27,

185–202.
11
12 [Link]
13
14 Pishvaee, M.S., Farahani, R.Z., Dullaert, W., 2010. A memetic algorithm for bi-objective integrated

15 forward/reverse logistics network design. Comput. Oper. Res. 37,

1100–1112.
16
17 [Link]
18
19 Polotski, V., Kenne, J.P., Gharbi, A., 2019. Joint production and maintenance optimization in flexible
hybrid
20 Manufacturing–Remanufacturing systems under age-dependent deterioration. Int. J. Prod. Econ. 216,
21
22 239–254. [Link]
23
24 Polotski, Vladmir, Kenné, J.P., Gharbi, A., 2019. Production control of hybrid manufacturing–
25 remanufacturing systems under demand and return variations. Int. J. Prod. Res. 57, 100–123.
26
27 [Link]
28
29 Ponte, B., Cannella, S., Dominguez, R., Naim, M.M., Syntetos, A.A., 2021. Quality grading of returns
and the
30 dynamics of remanufacturing. Int. J. Prod. Econ. 236. [Link]
31
32 Powers, T.L., Jack, E.P., 2013. The influence of cognitive dissonance on retail product returns. Psychol. Mark.
33
34 30, 724–735. [Link]

35 Pun, H., Chen, J., Li, W., 2020. Channel strategy for manufacturers in the presence of service freeriders.
Eur.
55
56
57
58
59
60
61
62
63 65
64
65
 1
2
3
4
5
6
7
8
36
37 J. Oper. Res. 287, 460–479. [Link]
38
39 Ramanathan, R., 2011. An empirical analysis on the influence of risk on relationships between handling
of

40 product returns and customer loyalty in E-commerce. Int. J. Prod. Econ. 130, 255–261.
41
42 [Link]
43
44 Rao, S., Rabinovich, E., Raju, D., 2014. The role of physical distribution services as determinants of product
45 returns in Internet retailing. J. Oper. Manag. 32, 295–312. [Link] 46
47 Reddy, K.N., Kumar, A., 2021. Capacity investment and inventory planning for a hybrid manufacturing–
48
49 remanufacturing system in the circular economy. Int. J. Prod.
Res. 59, 2450–2478.
50 [Link]
51
52 Richey, R.G., Chen, H., Genchev, S.E., Daugherty, P.J., 2005. Developing effective reverse logistics
53
54 programs. Ind. Mark. Manag. 34, 830–840. [Link]
Robertson, T.S., Hamilton, R., Jap, S.D., 2020. Many (Un)happy Returns? The Changing Nature of Retail
Product Returns and Future Research Directions. J. Retail. [Link]
Ruiz-Benitez, R., Muriel, A., 2014. Consumer returns in a decentralized supply chain. Int. J. Prod. Econ. 147,
573–592. [Link]
Rusmevichientong, P., Sumida, M., Topaloglu, H., 2020. Dynamic assortment optimization for reusable
products with random usage durations. Manage. Sci. 66, 2820–2844.
[Link]
Sahoo, N., Dellarocas, C., Srinivasan, S., 2018. The impact of online product reviews on product returns. Inf.
Syst. Res. 29, 723–738. [Link]

55
56
57
58
59
60
61
62
63 66
64
65
 1
2
3
4
5
6
7
8
Savaskan, R.C., Van Wassenhove, L.N., 2006. Reverse channel design: The case of competing retailers. 9

Manage. Sci. 52, 1–14. [Link]

10 Seeger, M.K., Kemper, J., Brettel, M., 2019. How information processing and mobile channel choice influence
11
12 product returns: An empirical analysis. Psychol. Mark. 36, 198–213. [Link]
13
14 Shang, G., Ferguson, M.E., Galbreth, M.R., 2019. Where Should I Focus My Return Reduction Efforts?
15 Empirical Guidance for Retailers. Decis. Sci. 50, 877–909. [Link]
16
17 Shang, G., McKie, E.C., Ferguson, M.E., Galbreth, M.R., 2020. Using transactions data to improve consumer
18
19 returns forecasting. J. Oper. Manag. 66, 326–348. [Link]
20 Shehu, E., Papies, D., Neslin, S.A., 2020. Free Shipping Promotions and Product Returns. J. Mark. Res.
57,
21
22 640–658. [Link]
23
24 Shin, Y., Lee, S., Moon, I., 2020. Robust multiperiod inventory model considering trade-in program and
25 refurbishment service: Implications to emerging markets. Transp. Res. Part E Logist. Transp. Rev. 138,
26
27 25. [Link]
28
29 Shulman, J.D., Coughlan, A.T., Savaskan, R.C., 2011. Managing consumer returns in a competitive
30 environment. Manage. Sci. 57, 347–362. [Link]
31
32 Shulman, J.D., Coughlan, A.T., Savaskan, R.C., 2010. Optimal reverse channel structure for consumer product
33
34 returns. Mark. Sci. 29, 1071–1085. [Link]

35 Shulman, J.D., Coughlan, A.T., Savaskan, R.C., 2009. Optimal restocking fees and information provision
in 36
37 an integrated demand-supply model of product returns. Manuf. Serv. Oper. Manag. 11, 577–594.
38
39 [Link]
55
56
57
58
59
60
61
62
63 67
64
65
 1
2
3
4
5
6
7
8
40 Si, H., Shi, J. gang, Wu, G., Chen, J., Zhao, X., 2019. Mapping the bike sharing research published from
2010 41
42 to 2018: A scientometric review. J. Clean. Prod. [Link]
43
44 Sievert, C., Shirley, K., 2015. LDAvis: A method for visualizing and interpreting topics. pp. 63–70.

45 [Link]
46
47 Simpson, D., Power, D., Riach, K., Tsarenko, Y., 2019. Consumer motivation for product disposal and its role
48
49 in acquiring products for reuse. J. Oper. Manag. 65, 612–635. [Link]

50 Small, H., 1973. Co citation in the scientific literature: A new measure of the relationship between two
51
52 documents. J. Am. Soc. Inf. Sci. [Link]
53
54 Souza, G.C., 2013. Closed-Loop Supply Chains: A Critical Review, and Future Research*. Decis. Sci. 44, 7–
38. [Link]
Srivastava, S.K., 2007. Green supply-chain management: A state-of-the-art literature review. Int. J. Manag.
Rev. [Link]
Stauffer, J.M., Kumar, S., 2021. Impact of Incorporating Returns into Pre-Disaster Deployments for Rapid-
Onset Predictable Disasters. Prod. Oper. Manag. 30, 451–474. [Link]
Stock, J., Speh, T., Shear, H., 2006. Managing product returns for competitive advantage. MIT Sloan Manag.
Rev. 48, 57–62.
Su, X., 2009. Consumer returns policies and supply chain performance. Manuf. Serv. Oper. Manag. 11, 595–
612. [Link]
Sun, M., Chen, J., Tian, Y., Yan, Y., 2021. The impact of online reviews in the presence of customer returns.
9 Int. J. Prod. Econ. 232, 107929. [Link]

10 Tang, S., Wang, W., Zhou, G., 2020. Remanufacturing in a competitive market: A closed-loop supply chain
11
12 in a Stackelberg game framework. Expert Syst. Appl. 161, 113655.
13
55
56
57
58
59
60
61
62
63 68
64
65
 1
2
3
4
5
6
7
8
14 [Link]
15 Teunter, R., 2004. Lot-sizing for inventory systems with product recovery. Comput. Ind. Eng. 46, 431–
441.
16
17 [Link]
18
19 Teunter, R.H., 2001. Economic ordering quantities for recoverable item inventory systems. Nav. Res.
Logist.
20 48, 484–495. [Link]
21
22 Teunter, R.H., Van der Laan, E., Inderfurth, K., 2000. How to set the holding cost rates in average cost
23
24 inventory models with reverse logistics? Omega 28, 409–415. [Link]

25 0483(99)00070-5
26
27 Tosarkani, B.M., Amin, S.H., 2018. A possibilistic solution to configure a battery closed-loop supply chain:
28
29 Multi-objective approach. Expert Syst. Appl. 92, 12–26. [Link]
30 Tsay, A.A., 2001. Managing retail channel overstock: Markdown money and return policies. J. Retail.
77,
31
32 457–492. [Link]
33
34 Van Der Laan, E., Dekker, R., Salomon, M., 1996a. Product remanufacturing and disposal: A numerical

35 comparison of alternative control strategies. Int. J. Prod. Econ. 45,

489–498.
36
37 [Link]
38
39 Van Der Laan, E., Dekker, R., Salomona, M., Ridder, A., 1996b. An (s, Q) inventory model with

40 remanufacturing and disposal. Int. J. Prod. Econ. 46–47, 339–350. [Link]

42 5273(95)00020-8
43
55
56
57
58
59
60
61
62
63 69
64
65
 1
2
3
4
5
6
7
8
44 Van Der Laan, E., Salomon, M., Dekker, R., Van Wassenhove, L., 1999. Inventory control in hybrid
systems
45 with remanufacturing. Manage. Sci. 45, 733–747. [Link] 46
47 van Eck, N.J., Waltman, L., 2014. Visualizing Bibliometric Networks, in: Measuring Scholarly Impact. pp.
48
49 285–320. [Link]

50 van Eck, N.J., Waltman, L., 2010. Software survey: VOSviewer, a computer program for bibliometric
51
52 mapping. Scientometrics. [Link]
53
54 Vorasayan, J., Ryan, S.M., 2006. Optimal price and quantity of refurbished products. Prod. Oper. Manag.
15, 369–383. [Link]
Wagner, L., Martínez-De-Albéniz, V., 2020. Pricing and assortment strategies with product exchanges. Oper.
Res. 68, 453–466. [Link]
Wan, X., Li, D., Chen, J., Lei, Y., 2020. Managing customer returns strategy with the option of selling returned
products. Int. J. Prod. Econ. 230. [Link]
Wang, C., Chen, J., Wang, L., Luo, J., 2020. Supply chain coordination with put option contracts and customer
returns. J. Oper. Res. Soc. 71, 1003–1019. [Link]
Wang, D., Dimitrov, S., Jian, L., 2020. Optimal inventory decisions for a risk-averse retailer when offering
layaway. Eur. J. Oper. Res. 284, 108–120. [Link]
Wang, H., 2004. Do returns policies intensify retail competition? Mark. Sci. 23, 611-613+631.
9 [Link]
10 Wang, J., Wang, Z., Chen, Y., 2021. Equilibrium strategies and optimal pricing in an online retailing
queueing
11
12 system. Nav. Res. Logist. 68, 556–576. [Link]
13
14 Wang, J.J., Chen, H., Rogers, D.S., Ellram, L.M., Grawe, S.J., 2017. A bibliometric analysis of reverse
15 logistics research (1992-2015) and opportunities for future research. Int. J. Phys. Distrib. Logist. Manag.
16
55
56
57
58
59
60
61
62
63 70
64
65
 1
2
3
4
5
6
7
8
17 [Link]
18
19 Wang, N., Liang, H., Jia, Y., Ge, S., Xue, Y., Wang, Z., 2016. Cloud computing research in the IS discipline:
20 A citation/co-citation analysis. Decis. Support Syst. 86, 35–47. [Link] 21
22 Wang, X., Xie, J., Fan, Z.P., 2021. B2C cross-border E-commerce logistics mode selection considering product
23
24 returns. Int. J. Prod. Res. 59, 3841–3860. [Link]
25 Wang, Y., Lan, F., Liu, J., Huang, J., Su, S., Ji, C., Pham, D.T., Xu, W., Liu, Q., Zhou, Z., 2021.
Interlocking
26
27 problems in disassembly sequence planning. Int. J. Prod. Res. 59, 4723–4735.
28
29 [Link]
30 Wang, Y., Xu, W., 2018. Leveraging deep learning with LDA-based text analytics to detect automobile
31
32 insurance fraud. Decis. Support Syst. 105, 87–95. [Link]
33
34 Wood, S.L., 2001. Remote purchase environments: The influence of return policy leniency on two-stage

35 decision processes. J. Mark. Res. 38, 157–169. [Link] 36

37 Xiao, T., Shi, K., Yang, D., 2010. Coordination of a supply chain with consumer return under demand
38
39 uncertainty. Int. J. Prod. Econ. 124, 171–180. [Link]
40 Xu, J., Duan, Y., 2020. Pricing, ordering, and quick response for online sellers in the presence of
consumer 41
42 disappointment aversion. Transp. Res. Part E Logist. Transp. Rev. 137, 20.
43
44 [Link]
45 Xu, L., Li, Y., Govindan, K., Xu, X., 2015. Consumer returns policies with endogenous deadline and
supply 46
47 chain coordination. Eur. J. Oper. Res. 242, 88–99. [Link]
48
49 Xu, S., Zhang, X., Feng, L., Yang, W., 2020. Disruption risks in supply chain management: a literature
review
55
56
57
58
59
60
61
62
63 71
64
65
 1
2
3
4
5
6
7
8
50 based on bibliometric analysis. Int. J. Prod. Res. [Link] 51
52 Xu, X., Chen, X., Jia, F., Brown, S., Gong, Y., Xu, Y., 2018. Supply chain finance: A systematic literature
53
54 review and bibliometric analysis. Int. J. Prod. Econ. [Link]
Yalabik, B., Petruzzi, N.C., Chhajed, D., 2005. An integrated product returns model with logistics and
marketing coordination. Eur. J. Oper. Res. 161, 162–182. [Link]
Yan, R., Cao, Z., 2017. Product returns, asymmetric information, and firm performance. Int. J. Prod. Econ.
185, 211–222. [Link]
Yang, G., Ji, G., Tan, K.H., 2020. Impact of artificial intelligence adoption on online returns policies. Ann.
Oper. Res. [Link]
Yang, H., Chen, J., Chen, X., Chen, B., 2017. The impact of customer returns in a supply chain with a common
retailer. Eur. J. Oper. Res. 256, 139–150. [Link]
Yıldız, G.B., Soylu, B., 2019. A multiobjective post-sales guarantee and repair services network design
problem. Int. J. Prod. Econ. 216, 305–320. [Link]
9 Yoo, S.H., 2014. Product quality and return policy in a supply chain under risk aversion of a supplier.
Int. J.
10 Prod. Econ. 154, 146–155. [Link]
11
12 Yue, X., Raghunathan, S., 2007. The impacts of the full returns policy on a supply chain with information
13
14 asymmetry. Eur. J. Oper. Res. 180, 630–647. [Link]

15 Zarbakhshnia, N., Kannan, D., Kiani Mavi, R., Soleimani, H., 2020. A novel sustainable multi-objective
16
17 optimization model for forward and reverse logistics system under demand uncertainty. Ann. Oper. Res.
18
19 295, 843–880. [Link]

20 Zhang, P., Pang, K.W., Yan, H., 2020. Coordinating inventory sharing with retailer’s return in the
consignment 21
22 contracts. Int. J. Prod. Res. [Link]
23
55
56
57
58
59
60
61
62
63 72
64
65
 1
2
3
4
5
6
7
8
24 Zhang, T., Choi, T.M., 2020. Optimal consumer sales tax policies for online-offline retail operations with
25 consumer returns. Nav. Res. Logist. 20. [Link]
26
27 Zhang, Y., Voorhees, C.M., Lin, C., Chiang, J., Hult, G.T.M., Calantone, R.J., 2021. Information Search and
28
29 Product Returns Across Mobile and Traditional Online Channels. J.
Retail.

30 [Link]
31
32 Zhou, L., Naim, M.M., Disney, S.M., 2017. The impact of product returns and remanufacturing uncertainties
33
34 on the dynamic performance of a multi-echelon closed-loop supply chain. Int. J. Prod. Econ. 183, 487–

35 502. [Link]
36
37 Zhou, S.X., Tao, Z.T., Chao, X.C., 2011. Optimal control of inventory systems with multiple types of
38
39 remanufacturable products. Manuf. Serv. Oper. Manag. 13, 20–34.
40 [Link]
41
42 Zikopoulos, C., Tagaras, G., 2007. Impact of uncertainty in the quality of returns on the profitability of a
43
44 single-period refurbishing operation. Eur. J. Oper.
Res. 182, 205–225.
45 [Link] 46
47 Zou, T., Zhou, B., Jiang, B., 2020. Product-line design in the presence of consumers’ anticipated regret.
48
49 Manage. Sci. 66, 5665–5682. [Link]
50
51
52
53
54

55
56
57
58
59
60
61
62
63 73
64
65
Method Details (MethodX)

Click here to access/download

Method Details (MethodX)
Supplement [Link]