1 Introduction

Social media is increasingly being used by organisations to manage business activities and relationships among supply-chain stakeholders (Choi et al., 2020; Huang et al., 2020; Singh et al., 2018) including, new product development (Giannakis et al., 2020), service improvement (Fan & Niu, 2016), daily sales forecasts (Cui et al., 2019), and information dissemination (Kanagarajoo et al., 2019). The prevalence and speed at which social media platforms (e.g., WeChat, Facebook, LinkedIn, Twitter, Flickr, WhatsApp) facilitate the exchange of real-time data (Gorwa & Guilbeault, 2020; Rousidis et al., 2020) has made it a popular tool for effective supply chain management (Choi et al., 2020). For instance, throughout the supply chain, social media is being used to monitor in real-time, risks, transactions, consumer sentiment, and purchasing decisions, as well as interact with consumers to track their behaviour (Brusset & Teller, 2017; Havakhor et al., 2018; Min & Zhou, 2002). Essentially, social media enables stakeholders of a supply chain to communicate and exchange information that can impact the forecasting of disruptions, order fulfilment, and production flow to match supply and demand (Grant, 2016; Rauniar et al., 2014). On one hand social media helps in communicating individuals and organizations about their products, services and developing network of communities that can be helpful to share the information on real time basis and create the awareness on a specific issue such as health during Covid-19 (Mheidly & Fares, 2020). On the other hand, social media is full of possibilities of fake and misleading content such as Covid-19 variants and their cure along with role of vaccination in fighting Covid-19, along with cyber risks (Brummette et al., 2018; Islam et al., 2020).

The universal access to real-time information that social media provides to users (Dwivedi et al., 2018) can lead to supply chain participants becoming highly susceptible to ‘confirmation bias’ and ‘echo chambers’ e.g., situations where individuals “hear their own voice” (Bessi, 2016; Boutyline & Willer, 2017; Brugnoli et al., 2019; Jacobson et al., 2016). The influence of echo chambers includes excluding alternate perspectives (Gillespie et al., 2014), a lack of common ground (Layman et al., 2006) and political, administrative, and social gridlock in many contexts (Hutchens et al., 2021; Kim & Kim, 2019). Therefore, an approach of mindfulness and reliability needs to be developed to avoid echo chambers across the supply chain (Butler & Gray, 2006; Dernbecher & Beck, 2017; Thatcher et al., 2018). Hence, social media induced polarization (SMIP) can be defined as skewed opinions, information/misinformation regarding socio-cultural activities including products or services that can lead to mass destruction of property and other crisis (O’Hara & Stevens, 2015).

The devastating effect of SMIP on communities, businesses, and governments has been identified by the World Economic Forum (2018) as the greatest threat to society due to the speed at which ‘digital wildfires’ spread on a global scale. Discourse on SMIP has reached mainstream media with prominent business leaders including Tim Cook (CEO, Apple) criticising polarisation and misinformation due to social media and the facilitating role that technology companies play by prioritising conspiracy theories and violent incitement because of their high rates of engagement (Reuters, 2021; Weckler, 2021). The providers as well as users are the victims of unreliable information that affect the decision making for most of the organizations.

In the context of SMIP, echo chambers refer to situations where users consume content and engage in discourse that predominantly supports the same point of view that users hold and believe themselves (Boutyline & Willer, 2017; Flaxman et al., 2016). In the context of this study, social media can induce polarisation in the form of echo chambers in global supply chains due to regional differences in terms of selecting, filtering, and sorting of information on social media (Flaxman et al., 2016; Miroudot, 2020). For example, echo chambers can occur as a global event unfolds (i.e., Covid-19, natural disaster) whereby supply chain participants are influenced by their professional and personal experiences (Bhagwat & Sharma, 2007; Muckstadt et al., 2001), as well as the information that is circulated and referred through social media (Gupta et al., 2021; Barberá et al., 2015; Toubiana & Zietsma, 2017). The impact of echo chambers on supply chain decision making can lead to excessive or insufficient stock levels in regional warehouses (Sharma et al., 2020).

The ubiquitous nature and speed of social media has been attributed to its popularity as a tool to coordinate supply chain activities (Choi et al., 2020). The speed of information dissemination on social media is also highly influential as content published first tends to get more attention and it influences the opinion of end users (Roozenbeek & van der Linden, 2019; Wang et al., 2015), as well as decision makers spread across the globe (Arnott, 2006; Wu et al., 2019). Supply chain participants are also increasingly exposed to high volumes of information and over a period they tend to favour certain social media platforms due to their familiarity with the source of information rather than its trustworthiness (Fidel & Green, 2004; Orji et al., 2020). The choice of social media platform can significantly impact supply chains from a risk management perspective (Culp, 2020; Fu et al., 2013; O’Leary, 2011) as supply chain participants are influenced by the sentiment of posts, concerns or tweets that are expressed by customers, retailers, and other stakeholders (Clapham et al., 2019; Wuebben, 2016; Zadeh & Sharda, 2014).

As supply chains become increasingly digitised using social media platforms and other digital technologies (i.e., AI, Big data), supply chain participants are increasingly seeking and believing the information from like-minded individuals (He et al., 2017; Sestino et al., 2020; Tim et al., 2021). This approach can increase the risk to echo chambers as participants may unknowingly seek reinforcement of existing biases and beliefs (Bhardwaj et al., 2016; Flaxman et al., 2016).

The phenomenon of ‘SMIP’ has been studied from different perspectives (e.g., Flaxman et al., 2016; Geschke et al., 2019; O’Hara & Stevens, 2015; Spohr, 2017) and in varied contexts, such as politics (Kim & Kim, 2019; Hyun, & Moon, 2016; Yarchi et al., 2020, Lee, 2016), climate change (e.g., Fisher et al., 2013; Leviston et al., 2013), and news framing (Han & Federico, 2018; Hyun & Moon, 2016). Recent studies (e.g., Del Vicario et al., 2017; Sikder et al., 2020) have begun modelling the combination of bias and polarisation to examine the impact of misinformation in social media networks. There remains however, much scope to advance understanding of SMIP as there is a noticeable absence of research that focuses on the structure of social media polarisation in the context of global events such as Covid-19, where participants are highly susceptible to developing echo chambers through confirmation bias (Boutyline & Willer, 2017; Orji et al., 2020). To this end, the aim of this study is to ‘explore how manifestations of confirmation bias contribute to the development of echo chambers, in the context of supply chains during the Covid-19 pandemic.’ To achieve this aim, we seek to answer the following research question: How do the elements of SMIP lead to the emergence of echo chambers in the context of supply chains?

To answer this research question, this study adopts an interpretive qualitative approach to explore SMIP from the perspective of supply chain participants by using thematic analysis to identify how four elements (e.g., environmental setting, actors, mechanism, and outcome) contribute to the development of echo chambers.

The remainder of the paper is structured as follows. A review of pertinent SMIP literature that informs our conceptualisation of echo chambers through confirmation bias is presented. Next, the research method and data collection and analysis are outlined. Then key findings and emergent themes are presented that form the basis of four cross-cutting propositions. Followed by a discussion, implications, and future research. The paper ends with a conclusion.

2 Literature Review

2.1 Social Media Induced Polarisation

The proliferation of social media has been a catalyst to inducing a dangerous, socio-cultural polarisation of society (Montalvo & Reynal-Querol, 2005; Spohr, 2017; Vishwanath, 2015). This polarisation phenomenon has led researchers to explore SMIP from different theoretical lenses to study specific social media platforms and particular attention given to fake news (Langley et al., 2021; Brummette et al., 2018; Lee et al., 2015a; Shearer & Grieco, 2019; Singh et al., 2020). Manifestations of SMIP include misinformation about presidential election campaigns (Guess et al., 2018; Linvill & Warren, 2020; Schäfer et al., 2017), race (Jamieson, 2020), immigration (Jaramillo-Dent & Pérez-Rodríguez, 2021; Newman et al., 2017), religion (Said, 2008), and pandemics, specifically Covid-19 (Laato et al., 2020). The rapid global adoption of social media technology has significantly influenced many job roles and supply chain activities (Ribarsky et al., 2014; Zamani et al., 2020). Organisations are incorporating social media technology as a core capability to deliver information efficiently and cost-effectively to stakeholders in their supply chain (Osatuyi, 2013; Thatcher et al., 2018; Xie & Lee, 2015; Zhang et al., 2017). For example, increasing its visibility in the supply chain ecosystem, facilitating real-time communication between its customers, and to mitigate risks (Wu & Li, 2018).

Social media tools (e.g., blogs, wikis) and platforms (Facebook, Twitter, and LinkedIn etc.) are increasingly being used by consumers (individuals) and providers (organisations) to (i) post and share information, (ii) develop a network of communities, and to achieve competitive advantage in both developed and developing countries (Batrinca & Treleaven, 2015; Obar et al., 2012; Stieglitz & Dang-Xuan, 2013). Essentially, social media enables individuals and organisations to act as influencers in the dissemination of information in terms of posting, re-sharing, commenting, tagging, and re-tweeting (Lee et al., 2015b; Shin & Thorson, 2017). The dark side of this form of endorsed information sharing is that consumers of the information are exposed to ‘filtered’ information (Fletcher & Nielsen, 2018; Gallaugher & Ransbotham, 2010; Kietzmann et al., 2011; Tran et al., 2020). This exposure increases the risk of misinformation being shared across the social media platform, leading to SMIP, such as echo chambers (Brugnoli et al., 2019; Mondal et al., 2018; Toubiana & Zietsma, 2017; Wu et al., 2019). Therefore, individuals need to rely on mindfulness and adopt reliable technologies consuming and sharing information (Endsley, 2018). Individuals also tend to seek social media platforms that support their beliefs and attitudes, to gain confidence in their biased views (Arnott, 2006; Nikolov et al., 2019; Shin & Thorson, 2017; Susarla et al., 2012). The fake content circulating on social media lead to extreme situations. For example, in 2019, a baby’s bloodied corpse led to the violence and killed at least 10 people in Africa’s Nigeria (LA Times, 2019). In another fake campaign run by secondary school students in Chile against the fare of metro, led to the developments of civil protest, violence and deaths leading to the instability for weeks in the country (Atlantic Council, 2019). Recently, people are posting the content related to Covid-19 and its vaccination impact and people are avoiding taking the vaccine, which is culminating the fear of another variant of Covid-19 and resulting into slow progress of inoculation (India Today, 2021).

The degree of SMIP is however, depends on the type of network (e.g., centralised or egalitarian) that individual consumers have joined (Juris, 2012). In a centralised network there is usually only one or a select number of individuals at the centre who play the role of ‘influencer’ as they can filter or block information before sharing it to a much wider network of its members (Culnan et al., 2010; Gallaugher & Ransbotham, 2010). While in an egalitarian network, all members have equal contact and influence across the network and therefore opinions can be shared by any individual within the network, and information is shared unfiltered to all members (Sikder et al., 2020). The attitude or belief system of a participant can be influenced due to the information processing approaches in both types of networks (Petty et al., 2009). When a participant is motivated and capable, they will carefully engage, reflect, and connect with pre-existing beliefs and then influence the cognitive network (Arnott, 2006). A cognitive network indicates a complex structure, where nodes signify concepts and branches specify one or more type of conceptual relationship, for instance two words describe the same sense in one context or like one another (Islam et al., 2020). The cognitive network is dependent on cognitive ability of individuals those are motivated, not motivated or lack cognitive ability may engage in peripheral processing and rely on the source of information due to its credibility that can lead to a change of attitude (Petty et al., 1987). Compared to central processing, the egalitarian network does not produce long lasting changes in belief system (Kelman, 2006).

SMIP can be witnessed in both types of networks (Centralised and Egalitarian) as partisan bias can become amplified when bias is displayed by a centralised echo chamber, while in the egalitarian echo chambers, the quality of ideas is a basis for their spread as compared to a particular individual (Chou et al., 2018). In the context of supply chains, participants act as ‘partisan’ (e.g., creator and gatekeeper of the shared information) or ‘bipartisan’ (e.g., consumer of the information or reshare/endorse the information) on social media platforms (Anand et al., 2007; Sharif, 2002). For example, due to national and regional lockdowns during the Covid-19 pandemic, people spent more time on social media discussing their views about the virus and vaccines (Sarkis, 2020; Islam et al., 2020; Nabity-Grover, 2020). Using social media in this way is an ‘infodemic’, due to the creation and consumption of misinformation related to the virus and vaccines which created confusion and distrust among people and hampered an effective response to the pandemic (Verma & Gustafsson, 2020) and influencing the supply chains.

2.2 Confirmation Bias

Confirmation bias occurs, when users encounter information that reinforces their pre-existing beliefs and attitudes (Bessi, 2016; Geschke et al., 2019; Zhao et al., 2020). Confirmation bias is defined as the tendency to interpret, search for, recall, and favour the information in a way that confirms one’s pre-existing beliefs or hypotheses resulting into polarised views about an issue or event (Westerwick et al., 2017; Gupta et al., 2021). Confirmation biases indicate why a group of individuals with opposing views on a topic can view the same evidence. The individuals those are victims of confirmation bias give more weightage to evidence to support their beliefs rather than undervaluing the evidence that can disprove it (Westerwick et al., 2017; Huang et al., 2012). Therefore, the tendency of people to favour the information on their prior exposure or hypotheses can be regarded as confirmation bias, which is stronger for the activities and issues those are highly emotive and deeply entrenched (Zhao et al., 2020). The information presented on social media is not only reflective of the designer’s values and beliefs, which can be one-sided and/or sensational (Zhang et al., 2017; Pariser, 2011). Previous studies conclude that users generally tend to search, believe, and share information that conforms to their own beliefs and attitudes, but ignore the information that challenges their beliefs and attitudes (Itzchakov & Van Harreveld, 2018; Knobloch-Westerwick & Kleinman, 2012; van Strien, Kammerer et al., 2016). In the context of SMIP, social media platforms facilitate the configuration and aggregation of ‘like-minded’ people and network that reinforce and further empower polarised groups (Giannoccaro, 2018; Westerwick et al., 2017; Stroud, 2008).

2.3 Echo Chambers

Echo chambers refer to the situation where people ‘hear their own voice’ (Boutyline & Willer, 2017; Flaxman et al., 2016). In the context of social media, it refers to situations where users consume content that expresses the same point of view that users hold themselves (Bessi, 2016; Geschke et al., 2019). As social media platforms largely rely on algorithms to generate large quantity of content, which can lead to the emergence of conspiracy theories, other forms of distorted information, and even extremist groups with a shared ideology, which further lead to the emergence of echo chambers (Boutyline & Willer, 2017; Flaxman et al., 2016; Schilling & Fang, 2014). The evolution of a group or community can be observed through the configuration of communication patterns (Tim et al., 2018). Weak patterns of communication between a group of users can be observed as compared to strong patterns of communication within the group, leading to the formation of echo chambers (Brugnoli et al., 2019). Since it is difficult to further identify the sub-groups by just analyzing the communication pattern on social media platforms, the measurement of echo chambers cannot be identified at group level as it requires the measurement of communication pattern within group as compared to between groups (Carr & Hayes, 2015; Go & You, 2016). However, the level of fragmentation between groups can define the degree of echo in chambers (Brugnoli et al., 2019).

The exponential growth of data creation, data flow and consumption across the spectrum of social media platforms offers many opportunities and challenges for supply chain participants (Legner et al., 2017; Sarimveis et al., 2008). On the one hand, supply chain participants are actively looking for the modes, where they can capture the precise information and improve the visibility, control, and efficiency in their supply chain (Sodhi & Tang, 2019). On the other hand, supply chain activities are governed by participants who have their own pre-existing beliefs and attitudes that can lead to ‘confirmation bias’, which can negatively impact the supply chain activities (Huang et al., 2020; Huang et al., 2012; Zhao et al., 2020). Earlier studies on social media have majorly focused on identity, social power, social capital, and social loafing that led to polarisation based on individuals’ attitudes, actions, and intentions for using social media (Dwivedi et al., 2018; Erkan & Evans, 2016; O’Leary, 2011; Rauniar et al., 2014; Rim et al., 2020). Indeed, such studies have made valuable contributions to knowledge; however, they do not conceptualise how the (i) foundational elements, (ii) role of actors, and (iii) phenomena lead to the emergence of echo chambers through dominance of confirmation bias (Boulianne et al., 2020; Sasahara et al., 2020). In this study, we conceptualize how confirmation bias can illuminate elements that can contribute to the development of ‘echo chambers’ in the context of supply chains (see Fig. 1).

In the context of supply chains, participants include professionals, consulting firms and organisations that play an equal role in creating and sharing the information. It is therefore appropriate to view the system as egalitarian, where partisan and bi-partisan members co-exist. Partisans are the producers and users of the content with one-sided leaning, whereas bi-partisan producers create and consume the content with both side view. The gatekeeper type of users consumes the content with both perspectives but produce it with single sided leaning after some filtration. These consumers and producers represent a smaller group with higher network centrality, without much embedded in their community network. It is relatively easy to identify partisan users as compared to gatekeepers (Garimella et al., 2018). The characteristics of partisans, bi-partisans and gatekeepers leads to the development of confirmation bias, whereby content is related and familiar and posted through new source (Jacobson et al., 2016). When different users on social media consume the information from different sources, they develop the echo chambers on the premise of confirmation of their beliefs resulting into polarisation.

Fig. 1
figure 1

Conceptualisation of echo chambers through confirmation bias on social media

Full size image

3 Research Methods

This exploratory study adopts an interpretive case study approach as it is suited to (i) understanding of the influence of context (Walsham, 1993), (ii) examining a new phenomenon that lacks empirical substantiation (Benbasat et al., 1987; Eisenhardt, 1989; Stake, 2000), and (iii) extending the theoretical underpinning of the study (Benbasat et al., 1987; Pagell & Wu, 2009; Stuart et al., 2002). The research design (see Fig. 2) provides a clear ‘chain of evidence’ (Yin, 2002) with explicit links between the research questions, the data collected, and the conclusions drawn, which helps “follow the derivation of any evidence, ranging from initial research questions to ultimate case study conclusions.” (Yin, 2002, p.83).

Fig. 2
figure 2

Research design

Full size image

3.1 Data Collection

Using snowball sampling (Naderifar et al., 2017; Noy 2008), 35 supply chain practitioners with extensive industry experience and knowledge of AI were identified for the purpose of this study (see Table 1). Interviews were conducted between September and December 2020. Due to Covid-19, interviews were conducted remotely via Microsoft Teams and Zoom. The duration of interviews ranged between 30 and 50 min. Interviews were recorded and transcribed using Otter, an AI-based transcription software, proof-read and annotated. Data saturation (Birks et al., 2013; Nelson, 2017) was achieved after 35 interviews, and therefore, no additional interviews were conducted. For the purpose of research rigour and anoymonity, a code (e.g., R1 to R35) was assigned to each interviewee.

Table 1 Interviewee profile
Full size table

3.2 Data Analysis

Data analysis was conducted using open, axial, and selective coding that further enabled us to derive themes and sub-themes through the process of iteration (Dezdar & Sulaiman, 2009; Strauss & Corbin 1997). Both the Glaserian and Straussian grounded theory methodologies are the most prevalent in extant literature (Matavire & Brown, 2013). The Glaserian approach is rooted in critical and rationalistic thinking, whereas the Straussian approach is pragmatic and can identify the phenomenon through interaction among participants. A thematic analysis of the qualitative data using the Strauss & Corbin (1997) method was adopted as it supports the development of theory based on the stories and views of participants and it provides a pre-defined coding scheme. We conducted primary, secondary, and tertiary analysis (three-layer analysis i.e., open, axial, and selective coding) since it was capable of consolidating and facilitating the raw data to meaningful themes.

Four key propositions emerged out of the data that was analysed through three stages. The first stage of the data analysis involved reading the transcripts, identifying differences and similarities between responses to the interview questions, and observing emerging themes. This stage is about conceptualisation and description of the data at basic level and is commonly known as open coding. In the second stage, we analysed the relationship among open codes leading to develop the axial codes. Second stage of axial coding of the data analysis involved further categorisation, by matching the open codes to the underlying constructs of confirmation bias. The third stage of data analysis known as selective coding employed for identifying the relationship among axial codes to map the phenomena of emerging echo chambers. Through this three-stage process of data analysis, we shed new light on the role of confirmation bias in participants involved in supply chain information processing. The coding scheme is presented in Appendix Table 8. Table 2 presents the triangulation approach, where we found the convergence through industry reports, research articles and data collected from the participants.

Table 2 Triangulation approach
Full size table

4 Key Findings & Emergent Themes

This section presents key findings under four themes that emerged from the data, namely, environmental setting, actors, mechanism, and outcome. Each of these themes are discussed in the context of supply chains and form the basis for the proposed framework of SMIP and the emergence of echo chambers.

4.1 Environmental Setting

The surroundings and fragmented society derive the unequal status and economic displacements in every part of the world (Bessi, 2016). The society represents individuals with democratic and republican mind set from political discourse view. Democrats and republicans develop a ground for opposite views with their own belief system and philosophy (Neiman et al., 2016). The role models, influencers and benchmarks advance the environmental setting. Additionally, the persuasion of social and human ideas, social security, government policies and economic principles lay the foundation in the interaction of diverse stakeholders on different platforms influencing the surrounding factors (Barnett, 2007). These adjoining factors further rise to the diverse culture that has a close relationship towards the active participation and supporting the diverse activities in a supply chain. Table 3 displays the related quotes and codes to environmental setting that represents the formation of echo chambers.

Table 3 Sample of quotes and coding for the environmental setting
Full size table

4.2 Actors

Social media generates large volume of data that is influenced by agents or actors. These agents (e.g., partisan, and bipartisan) work around the specific features and characteristics that focus on certain pain points ranging from individuals to large entities such as supply chain (Shin & Thorson, 2017). For instance, Facebook uses the private data of users to expose them with large amount of content (Bessi, 2016). The companies or product specific characteristics distract the users from the fundamentals through social media and it serves as a society-fracturing tool. The consulting firm’s belief as an actor strengthens when their success or failure stories are witnessed over social media (Anand et al., 2007). The belief system further gets a strong bond when other similar entities succeed or fail that lead to make their belief system more extreme (Teo et al., 2008). Supply chain interaction with social media lead information exposes it to numerous topics and global events. For example, Covid-19 directed practitioners to take certain actions to derive the business operations that define the social identity of the supply chain, which plays a critical and influential role in supply chain decisions. Table 4 displays the related quotes and codes for actors that contribute to the formation of echo chambers.

Table 4 Sample of quotes and coding for actors
Full size table

4.3 Mechanism

In the age of social media not only in individual, but also in business to individual and business to business interactions, the personalisation is pervasive (Neves et al., 2014). The personalisation facilitates the higher user efficiency and greater revenue to the platform by providing relevant content to the user needs. Feeds, group membership, product recommendation and e-advertisement pervade the system towards personalization (Tim et al., 2018). Data and content selection algorithm consider the user’s characteristics, activities on social media and interests based on past behaviour to display the personalised list of content (Stepan et al., 2016). The personalized content and news feed along with other sources can lead to filter bubble in developing a tunnel vision for individuals to groups (Pariser, 2011). This way practitioners as well as systems like supply chain develop their ideological bubble. On the other hand, the choices that individuals make on social platform facilitate group and systemic biases to polarise the belief system (Dubois & Blank, 2018). Content is categorised into diverse groups based on sensitive attributes viz. some stories may have a particular political orientation (liberal or conservative) or a topic that can impact a particular sector of the business or supply chain.

The objective of data selection algorithm is to match the user needs to maximise the positive feedback; hence it must learn and organise about the topics that the user is most interested in. Table 5 displays the related quotes and codes for mechanism that represents the formation of echo chambers.

Table 5 Sample of quotes and coding for mechanism
Full size table

4.4 Outcome

The digital economy and environment have steered the multitudes of information platforms (Tim et al., 2021). The practicing managers in supply chains should examine and analyse large amount of information and by doing so they may favour or develop certain channel of information in their day-to-day operations. This selection of information channels creates the risk of online echo chambers that reinforces the pre-existing biases and beliefs (Flaxman et al., 2016). The choice of social media platforms or information channels have a significant impact on consumer choice and therefore, exposes risk to stakeholders of supply chain. The familiarity and inclination with a social media source of information outplays as compared to its inherent reliability and trustworthiness (Dwivedi et al., 2018; Butler & Gray, 2006). This is observed by the fact that even though witnessing an increase in computerisation and availability of information around 33 % of the companies have not analysed their supply chains for the potential sources of disruption (Gomber et al., 2018). The remaining 67 % of the companies have developed their echo-chambers based on the success and failure stories pre and during Covid-19 to ensure the continuity of their business. Covid-19 has compromised many individual accounts on social as well as personal space that indirectly influence the supply chains and hence consider investing in strengthening cyber security, increasing flexibility, strengthening local supply chains and identify the risk nodes for better response during emergency situations (Spanaki et al., 2019). Table 6 displays the related quotes and codes to outcome that represents the formation of echo chambers.

Table 6 Sample of quotes and coding for outcome
Full size table

Following the thematic analysis, the study captured different open codes from the data of interview. After multiple iterations, back and forth, we recognized the axial codes those were further integrated to develop selective codes. In summary, our study identifies four selective codes presented in Table 7.

Table 7 Summary of findings
Full size table

5 Propositions and Framework for SMIP in Supply Chains

In this section, we developed four propositions based on the thematic analysis of the data. The propositions demonstrate the contextual complexity of supply chain information sharing and the impact of information sharing that is highly nuanced.

  • Proposition 1: The off-line and online experiences of diverse stakeholders and degree of exposure to social media set the diverse expectations from a supply chain and influence the partisan and bi-partisans orientation.

The offline experience of individuals, professionals and organisations acts as base to model their cognitive behaviour in online platforms and social media (Arnott, 2006). Users expect the content in line of their expectation and experience on internet and that further influence their beliefs. In the words of R30 (Logistics coordinator with 3 years of experience from logistics and supply chain), “Earlier it was expected that internet and other social media can provide greater transparency by offering more content, however with varied expectations and experiences of stakeholders has developed a different kind of environment for users and their thought process”. Many social media users consume the information without seeing the source and develop their theories for the topic of their interest or related to it (Westerwick et al., 2017; Zamani & Pouloudi, 2021). Social media is a mix of genuine and fake content posted by related individuals, groups or organisations and it is very difficult to make out and filter the true ones (Tim et al., 2018). In the words of R9 (Analyst with 3 years of experience from consulting), “Many of the supply chain decisions are aligned with the incidents reported on social media to save the image of an organisation. Additionally, many consumers on the experience of their friends or relatives opt social media to complain or connect to company in case of product failure and this has been very frequent during Covid-19 due to face-to-face restrictions. In some cases, the influential people amplify the voice of actual complainer”. This way the exposure and expectation from social media platforms including online and offline experiences act as a basis to develop a certain belief system and influence the activities of value chain.

  • Proposition 2: Supply chain consulting firms create and develop new forms of knowledge and information that help in displaying the social identity. The products and service behaviour are tracked to recognise the consumer choices for personalised content to be offered.

In the age of information many functions of a business are using social media to understand the market trends and reaction to a specific offer (Felix et al., 2017). It has become more obvious for supply chain and related professionals, since facilitating the consumer is one of the key objectives and most of the consumers are present and use social media for many decisions that further influence the activities of a supply chain (Fu et al., 2013). In the words of R26 (Logistics manager with 10 years of experience from construction), “The performance of a supply chain is measured, how the product and organisational efforts are recognized and perceived by real consumers, therefore to track the product performance and resolve the complaints we are connected to our consumers through social platforms such as Twitter, Facebook and LinkedIn and also we consider consulting firms advices and piece of information shared on their websites regarding changing needs of a supply chain”. The expectations of customers are also advanced by some of the consulting firms indirectly and therefore the organisations map their social identity to match the customer expectation and the cluster they belong to (Sharif, 2002). In the words of R31 (Logistics manager with 10 years of experience from logistics and supply chain), “The supply chains design their networks to avoid the role conflict and maintain intergroup relations due to the large players involved in delivering a product and offering aftersales services, therefore the supply chain develops a specific social identity to create the user perception”. The network structure and architecture can act as an actor to influence the user choices on social media. These actors exchange and develop their knowledge based on their feeds.

  • Proposition 3: User choices trigger the algorithmic personalisation that form the filter bubbles and vice-versa. These filter bubbles influence the individual, group, and system biases and vice-versa and it creates the echo-chambers along the supply chain.

The opinion and existing knowledge gained through websites, blogs and other forms of internet help forming the choices of the content a consumer want to view (Zhao et al., 2020). This represents the polarisation cycle from the view of confirmation bias that is pervasive. In the words of R35 (Director with more than 10 years of experience from logistics and supply chain), “We use multiple personalisation approached on the basis of user clusters and their search history to enhance the loyalty of the user to our portal that further advanced the product recommendation through genetic adaptive architecture for profiling”. This personalisation creates the filter bubble for individuals and groups (Garrett, 2009; Kapoor et al., 2018). In the words of R20 (Consultant with 3 years of experience from consulting), “In one of the experiences with my client, I offered applications from different vendors for enterprise resource planning, to my surprise and save on time, they advised me to describe the product offered by Oracle. Two years later, when I visited another organisation from the region, they quoted the earlier organisation success and opted for Oracle”. Therefore, the polarisation occurs due to the individual, group, and system biases.

  • Proposition 4: The user’s choices, experiences and algorithmic personalisation influence the confirmation bias that further develop the echo chambers in supply chains. As a result, some supply chains are focusing more on digitalisation or monitoring real-time information, where some others are investing heavily on cyber security, risk mitigation and regionalisation.

Due to the exposure to large amount of data in supply chain processes, tracking and monitoring of customers and their pain points through social media becomes a complex task (Manavalan & Jayakrishna, 2019). There has been huge wave of data and that has been advanced during Covid-19 due to more time spent by public on internet and social media platforms as well as companies interacting with their customers (Nabity-Grover et al., 2020). In the words of R10 (Engineer with 3 years of experience in logistics and supply chain), “The non- face to face interaction to ensure business continuity has turned many organisations to automation and to strengthen their digital architecture, whereas the players those could have avoided this investment on digitisation have also invested and some of them resulted into low return of investment, thus creating an echo chamber with respect to digitisation”.

Supply chains have been found referring to the websites and reports of consulting companies and sharing few success stories have focused on enhancing the flexibility and approaches for identifying and mitigating the risk at different nodes (Anand et al., 2007). With the digitisation, there are more transactions online, therefore online tracking of real-time information is critical. In the words of R25 (Engineer with 5 years of experience in IT services), “With the increase of digital payments and online tracking the security and safety of the systems is critical and have been advised by a consulting firm to spend a specific amount of the revenue in ensuring the networks are secure and we are proceeding in that direction”. Separately from this, the focus on local supply chains have been emphasised and realised due to time, safety, and cost of sourcing from global supply chains.

Building on the emergent themes and propositions, a conceptual framework of SMIP and the emergence of echo chambers is presented in Fig. 3.

Fig. 3
figure 3

Proposed framework for social media induced polarisation in supply chains

Full size image

6 Discussion

Research has recently addressed the implications of SMIP in the context of supply chains highlighting issues in decision making (Sharma et al., 2020), risk management (Fu et al., 2013; O’Leary, 2011), sentiment (Wuebben, 2016; Zadeh & Sharda, 2014). To date however, limited research exists on understanding how manifestations of confirmation bias contribute to the development of ‘echo chambers’ in the context of supply chains. A context where it has been acknowledged that participants are highly vulnerable to developing echo chambers through confirmation bias (cf. Boutyline & Willer, 2017; Orji et al., 2020).

Drawing on contemporary literature, we frame the theoretical and empirical contributions of this study (Ågerfalk, 2014; Corley & Gioia, 2011; Zamani et al., 2020). This is attained in several ways. The most salient theoretical contribution of this study is to use the confirmation bias and echo chambers as a lens to study SMIP in the context of supply chain during the Covid-19 pandemic. This is an important contribution as studies linking confirmation bias and echo chambers to supply chains remain under studied (Fu et al., 2013; O’Leary, 2011; Orji et al., 2020). This study also contributes to addressing the grand challenges of IS research (Becker et al., 2015; Winter & Butler, 2011), by specifically focusing on important socio-technical challenges such as effective IS for emergency management, and leveraging knowledge from data, with the related management of high data volumes (Becker et al., 2015). As this study builds on extant SMIP literature, it makes an empirical contribution by extending the generalisability of SMIP to provide novel insights that have not been reported in previous studies. This is achieved by identifying four themes (e.g., environmental setting, actors, mechanism, and outcome) and developing a set of propositions in the context of supply chains that may affect the extent to which confirmation bias facilitates creation of echo chamber in supply chains.

6.1 Implications for Research

This research builds on the cumulative body of knowledge related to SMIP, by rigorously identifying four themes (i.e., environmental setting, actors, mechanism, and outcome) and developing a set of propositions that provide rich qualitative insights. These propositions, which were developed through the lens of confirmation bias, can enable researchers from a range of disciplines to assess them through hypothesis testing. The proposed framework conceptualises the complex phenomenon of SMIP, which in turn provides a structure for researchers to examine the four themes collectively or individually, and in different contexts. This study extends previous studies by demonstrating that there is a reciprocal, reinforcing relationship between confirmation bias and echo chambers that accelerates SMIP, as opposed to just the social media platform itself. Our study displays the tendency to search, recall and favour the information with reference to existing beliefs and experiences that professionals have online and offline. Further, our study contributes to SMIP literature by presenting the relationship among different aspects developed based on filter bubbles created and user choices leading to polarized in supply chains. Other studies in the literature indicate the other forms and ways of forming confirmation bias that leads to emergence of echo chambers. For instance, Sasahara et al. (2021) pointed out that minimum social media exposure also influences the thought process and developing a network of choice by un-following people of a different mind-set, leading to homogenous communities and echo chambers. Apart from developing polarization and echo chambers, confirmation bias reveals the personal traits of the individual, group, and system specifically in dynamic times such as Covid-19. Our study has unveiled the structure of elements that lead to the creation of echo chambers across the supply chain.

6.2 Implications for Practice

The research provides further evidence to support supply chain practitioners and members of online communities to reflect and reassess their pre-existing biases (Parra et al., 2021) and how their choice of social medial platforms can positively or negatively reinforce this deep-seated belief system. Before adopting, referring, and developing an opinion about the information on social media that can directly or indirectly influence the supply chain, it is imperative to analyse and understand (i) the source of information (ii) validation of information, and (iii) openness while considering the information. Managers need to increase awareness about highly influential nature of online and offline information in terms of employees’ acceptance of information received via social media and their role as consumer and producer of this information further. Managers need to understand the highly complex, contextualised environment of supply chains due to the large number of actors involved in the ecosystem, perception, orientation of the organisation, type of products and what kind of social identity their supply chain is displaying to online consumers. The managers can identify the nodes of filter bubbles and clusters of echo chambers in their supply networks by employing the proposed framework to solve the issues occurring in their day-to-day operations. The practicing managers, governments and consumers of the information can work together to solve the problems of fake information and developing the perspective on that basis. The technology and social media companies can come together in developing technologies those can identify the fake information and reduce the incentives for the beneficiaries of misinformation on different platforms. The lack of appropriate and precise information can further create information vacuum that offer space for misinformation to circulate. Managers should have a strong focus on the scope of developing echo chambers especially when there is limited information about what is not known. Finally, individuals need to be informed about the implications of dispersed information and how it can lead to a ‘rose tinted’ view of the real-world.

6.3 Limitations and Future Research

We acknowledge this study has two limitations, which also offer directions for future research. First, the study was conducted during a global pandemic that behaved like a ‘once-in-a-century pathogen’ (Gates, 2020) which led to restricted movements due to national lockdowns and heightened emotive discussions around its source and solution. Future research could consider contextual factors such as the role of organisational culture which has not been adequately explored in this context, and there is a need for research linking social media technology to national culture (Gupta et al., 2018; Gupta et al., 2019). Second, the utility of the proposed framework for SMIP in supply chain has not been evaluated. Future studies could adopt a design science approach to assess its utility in the context of supply chains or extend it to other social settings to identify the sources and consequences of echo chambers. Despite these limitations, it provides direction for future multidisciplinary research.

7 Conclusions

This study draws on confirmation bias and echo chambers to advance understanding of how manifestations of confirmation bias contribute to the development of echo chambers in supply chain information sharing during the Covid-19 pandemic. The findings demonstrate that there is a reciprocal, reinforcing relationship between confirmation bias and echo chambers that accelerates SMIP, as opposed to just the social media platforms itself. Concluding, this study advances understanding of the architecture of confirmation bias that leads to echo chambers, not just in the context of global pandemics, but also in national and international events that can have devastating effects on societies.