A systematic review of customer behavior in business-to-business markets and agenda for future research Open Access

The practice of business-to-business (B2B) dates back several thousand years, and it is possible to find evidence of it as far as ancient Greece (Michell, 1940). However, it was not until the past four decades that significant studies on B2B marketing emerged, and we are now in a stage of fast-paced development (Mora Cortez and Johnston, 2017; Zhang et al., 2019). Although still underrepresented within the broader scope of marketing, B2B research is becoming increasingly relevant and an exciting field of study, leading to a richer body of literature (LaPlaca and Katrichis, 2009; Mora Cortez et al., 2021).

B2B research presents unique challenges for researchers, often requiring approaches different from those used in business-to-customer (B2C) domains (Wiersema, 2013). The buying process in B2C deals with more complex and emotional behaviors involving households of few customers (Fetscherin and Heinrich, 2015), whereas B2B buying involves organizations with dozens of individuals from different backgrounds and motivations in the purchase process (Lilien, 2016). The data for research is also far scarcer and more difficult to collect than in B2C settings, requiring the involvement of several cooperating organizations (Gould et al., 2016; Kumar and Pansari, 2016). Additionally, the B2B buying process is complex and heterogeneous (Aarikka-Stenroos et al., 2018). B2B transactions often involve products requiring significant expertise for the purchase decision, such as titanium dioxide for the paper industry or polyvinylchloride for the plastic industry. Finally, the B2B marketing field has undergone fundamental changes in recent years, and we can observe an evolutionary process in the B2B marketing efforts, with an evolutionary shift from the exchange philosophy (transaction-based marketing) to a behavioral philosophy (relationship-based marketing) (Kaski et al., 2017; Xu and Hao, 2021). For this research, we will indistinctively use the terms “B2B buyer” and “B2B customer.”

Because of these differences, understanding the behavior of industrial customers and the B2B buying process remains a leading research priority with high potential for academics and practitioners (Lilien, 2016; Xu and Hao, 2021). Therefore, research on B2B customer behavior is essential to organize and systematize existing knowledge (Grewal et al., 2015; Lilien, 2016) and combine it with new perspectives, steering future researchers to improve the understanding of B2C customer behavior. This study offers a valuable and original contribution to the field by summarizing the literature’s main topics and discussing future research avenues, proposing a comprehensive research agenda that can potentially unlock new theoretical and managerial knowledge about B2B customer behavior and benefiting researchers and practitioners in the B2B marketing domains.

This paper provides a systematic review of the existing literature on B2B customer behavior. To the best of the authors’ knowledge, previous studies have explored various aspects of B2B advertising (Swani et al., 2020), B2B market segmentation (Mora Cortez et al., 2021) or B2B branding (Leek and Christodoulides, 2011). However, there is still no similar systematic review on the topic, which evidences this paper’s timely and relevant contribution. This study seeks to address this gap by mapping the existing knowledge about B2B customer behavior, identifying and summarizing the main topics arising from the existing body of knowledge, contributing with an integrative framework and, finally, identifying future research avenues and offering a research agenda. The analysis is based on a systematic literature review of 219 papers, which were analyzed using a text-mining approach based on the Latent Dirichlet Allocation (LDA) algorithm. The results of this review reveal ten major research topics on B2B customer behavior, and we apply the Theory, Context, Characteristics and Method (TCCM) framework to summarize the main topics (Paul and Rosado-Serrano, 2019).

This study is of practical relevance for academics and managers, offering a descriptive overview of the core topics in B2B customer behavior. It also proposes an integrative framework for those seeking to deepen their knowledge in this field and a future research agenda, contributing to the evolution of research in this area and the advancement of new knowledge in B2B customer behavior.

This paper offers a systematic literature review and future research agenda for B2B customer behavior. We apply a systematic literature review methodology, as it allows researchers to identify, select, critically evaluate and synthesize the literature in a rigorous, transparent and replicable way, leading to solid outcomes in a specific research domain (Christofi et al., 2017; Tranfield et al., 2003). This review method has several advantages when compared to traditional reviews: improves the review process and outcome quality (Leonidou et al., 2018); reduces bias and errors (Tranfield et al., 2003); increases the process validity because of its process replicability (Wang and Chugh, 2014); allows the information synthesis and mapping of a specific research topic (Paul and Criado, 2020); and offers frameworks that researchers and practitioners may use (Kumar et al., 2020; Paul, 2015).

Systematic literature reviews are common in several exact sciences, such as medicine, chemistry and others (Harris et al., 2006; Moher et al., 2009), and they are an increasing trend in the management and marketing fields of study, applied in recent studies published in the premier and high-impact management journals (Cartwright et al., 2021; Hayes and Kelliher, 2022; Kumar et al., 2020; Rosado-serrano et al., 2018). The systematic literature review is the appropriate method for this research, as it provides a comprehensive and high-quality state-of-the-art review of the research focusing on B2B customer behavior.

Review papers can have a variety of forms, such as a structured review focusing on widely used methods, theories and constructs (Kahiya, 2018), a framework-based review (Paul and Benito, 2018), a hybrid narrative with a framework for setting a future research agenda (Bilro and Loureiro, 2020), a theory-based review (Gilal et al., 2019), a meta-analysis review (Knoll and Matthes, 2017), a bibliometric review (Randhawa et al., 2016) and a review aiming for model/framework development (Paul and Mas, 2020). For this paper, the authors adopt a hybrid narrative with a framework review comprising a structured review followed by a TCCM framework.

Researchers conducted an extensive search on the “Web of Science” (WOS) and SCOPUS electronic databases using the six Ws of the literature review method (Callahan, 2014) and the well-established guidelines for review articles search method found in previous reviews (Altuntas Vural, 2017; Paul and Criado, 2020). WOS and SCOPUS are renowned electronic databases; the content of their collections is selective and consistent; and independent detailed editorial processes ensure journal quality (Clarivate, 2021). The use of the journal as the criterion to assess the research quality is widely adopted (Chavarro et al., 2018; Loureiro et al., 2020).

Researchers developed a list of search terms with broad coverage to minimize the possibility of excluding a search term that could generate relevant studies (Leonidou et al., 2018; Müller-Seitz, 2012). The search was only limited to the research process timeline. So it was possible to capture all relevant literature irrespective of the publication date, including all papers published in scholarly journals until July 2022. The keyword selection was based on its relevance to the topic, and the search focused on variables endeavoring to explain customers’ behavior in both spellings “behavior” and “behaviour.” The words “customers” and the “buyers” were used to incorporate the different decision-makers, as we face transactions involving organizations. The research results were restricted to B2B definitions commonly found in the literature, such as “industrial,” “B2B,” “Business-to-Business,” “b-to-b” and “BTB.” The search was conducted for keywords in the title, abstract and keywords (Paul and Criado, 2020). The final query for our search is:

The results were limited to articles and reviews, and the chosen categories focus on business and management: business; management; economics; operations research; management sciences; and business finance. The search resulted in 8,631 articles at SCOPUS and 7,206 at WOS, of which 711 remained after successive filters were applied and duplications were removed. The sequential reading of titles, abstracts and articles allowed the identification of 219 articles independently reviewed by two researchers (Macpherson and Holt, 2007), assuring the focus on this review topic. Only studies that meet all the inclusion criteria specified in the review were included ( Appendix). The strict criteria specified in the systematic review are linked to the need to base the review on the best-quality papers available. Our final pool of papers is the outcome of this process (Figure 1).

The literature about B2B customer behavior dates back to 1971. The Journal of Marketing Research published a paper by Cardozo and Cagley (1971), which undertook an experimental study of industrial buyer behavior. However, most of our final pool of papers were published only after the millennium (n = 185; 84.47%), and slightly half of the studies were published in the past ten years (n = 106; 48.4%), reinforcing the arising relevance of this topic among the marketing literature (Figure 2).

Most of the studies are empirical (n = 185; 84.47%), and most are quantitative (n = 145, 66.21%). We can see flourishing empirical studies after 2007 and some peaks in conceptual research (Figure 2). From our final pool of papers, 158 (72.15%) used a theoretical context to support and expand their findings. The analysis reveals that the social exchange theory (SET) (Blau, 1964; Thibaut and Kelley, 1959) is the most frequently used theory among the papers (n = 27; 19%), followed by the transaction cost economic theory (TCE) (Williamson, 1993) (n = 13; 9%) and the relational exchange theory (RET) (Macneil, 1980) (n = 12; 8%). Table 1 shows the most used theories for at least three papers.

Almost half of our final pool of papers were published in the sectorial journals of “Industrial Marketing Management” (n = 70; 32%) and “Journal of Business & Industrial Marketing” (n = 37; 17%) (Table 2). The analysis also shows that most studies are published in top-tier marketing journals. Tier journals ranking AJG – Academic Journal Guide – 4*, 4 and 3 ratings (former ABS –chartered Academic Business School ranking) have demonstrated interest in this topic, such as the “Journal of Marketing,” “Journal of Marketing Research,” “Journal of The Academy of Marketing Science,” “International Journal of Research in Marketing” and “Journal of Business Research” (Table 2). This analysis suggests that these journals have been highly receptive to publishing articles on this topic and that marketing scholars are positioning their work and articulating its importance to mainstream marketing and business theory and practice.

The topic analysis of the final pool of papers explores the complete paper’s text to capture the full available information and highlight the latent discussions. Full papers were downloaded and transformed into ASCII text (a common encoding format), and researchers conducted the topic analysis using the R software, an open-source statistical tool used for data analysis (Breuer, 2017). We use the packages tm and topic modeling to transform the text into a corpus, producing the document-term matrix and computing the topics through LDA algorithm (Blei et al., 2003), which has been successfully applied in recent research (El Akrouchi et al., 2021; Xiang et al., 2017).

Distinct text-mining tasks are applied to the textual content of the papers. The data cleaning and stemming started by converting the text into lowercase, and numbers, punctuation and whitespaces were removed. Next, we removed common stop words in each sentence, as those words do not have any analytic value. Finally, stemming was applied to reduce all words to their root to avoid related words being considered different (Wu et al., 2017). The remaining text was computed into a document-term matrix (DTM), a matrix-format structure where each row represents a paper, each column a word and within each cell appears the number of times a word occurs within a paper. The number of topics in LDA is an input parameter that must be set previously, so we resort to existing measures (Cao et al., 2009; Griffiths and Steyvers, 2004) to compute the ideal number of topics (Figure 3), with the set of possible topics ranging from K = 2 to K = 40. The log-likelihood and perplexity start establishing around K = 8, reaching their optimal values around K = 15 (minimize K = 6 and maximize K = 10). The strategy for obtaining the ideal topic number is given by the proximity score showing a clear peak, and the nearest neighbor score flattens (Grant et al., 2013). Uncertainties about the point of flattening can be solved by comparing the measures in use (Figure 3). Therefore, for the current analysis, K = 8 was selected.

The topic models were conducted using LDA with a Gibbs sampling technique (a Markov Chain Monte Carlo algorithm), used in this research because of its convergence and performance capabilities. LDA is a mixed-membership algorithm widely used for clustering text into latent topics (Blei et al., 2003). LDA is based on a hierarchical Bayesian analysis and calculates the posterior probability of each word found in the text and each paper belonging to a latent topic. Because of its mixed-membership model feature, each paper may belong to multiple topics (several discussions being addressed in the text).

The profiling of each topic was delineated by analyzing the document-topic classification probabilities using the package tidytext. To know which papers are associated with each topic, we can examine the per-document-per-topic probabilities called γ (gamma). Besides estimating each topic as a mixture of words, LDA also models each document as a mixture of topics. The more words in a document are assigned to that topic, the more weight (gamma) will go on that document-topic classification. In the analysis, gammas present high values, which may be because of lower correlations between the topics. Table 3 shows the top three articles per topic. The content of each topic is discussed and analyzed in the next section.

The buyer–supplier relationship is a prevailing concept in the B2B literature. Effective exchanges between buying and supplying firms are crucial, yet the conflicting goals within the relationship can often lead to conflicts, which pose a significant managerial challenge (Ellegaard and Andersen, 2015). These conflicts stem from differences in behaviors and expectations between the exchange partners, resulting in uncertainty and a breakdown in the relationship, even when parties behave better than expected (Wang et al., 2010). Common sources of conflict include disparities in projected supply/demand, product quality and service performance (Ellegaard and Andersen, 2015). Conflict resolution processes can lead to common behavior patterns such as avoidance or lack of communication, which gives awareness of why exchange relationships that hit a downward spiral can be difficult to secure (Wang et al., 2010).

To overcome these issues, buyers and sellers should aim to achieve joint competitive advantages through inter-organizational goals, congruence and trusting relationships, leading to improvements in profitability, future expectations and relationship functioning (Jap, 2001). Partners should also find ways to collaborate and avoid opportunist behaviors, achieving a trusting relationship (Zhang et al., 2019). The positive effects of collaboration on the relationship can be seen in the restoration of trust, tolerance and avoidance of opportunism (Zhang et al., 2019), making it critical for such relationships to last. Superior buyer–supplier relationships enhance the potential to yield solid outcomes for both parties, reinforcing the partners’ attractiveness in the selection, formation and choice of B2B partnerships, leading to successful outcomes in competition (Gould et al., 2016).

B2B customer behavior and relationships are typically formed by a contract between two or more legally independent parties. The ability of each party to achieve its objectives is contingent upon its relative bargaining power (Porter, 1980). Most aspects influencing bargaining power are often challenging to change, depending on characteristics of the production process, industry characteristics or volume of purchases (Dampérat and Jolibert, 2009). Buyers and sellers seek to exploit asymmetries in their relationship during the negotiations in distinct ways to gain strategic advantages. In certain markets, buyers have acquired advanced procurement techniques and established considerably stronger negotiation positions through control of the procurement process and powerful price negotiation tools (Gadde and Wynstra, 2018).

On the contrary, suppliers concentrate on initiating, signaling and disclosing behaviors to enhance their relationships with buyers, with deliberate efforts to understand their customers’ business conditions, adjust to market changes and disclose information about themselves that reinforce the buyer’s trust (Vieira and Brito, 2015). The influence of suppliers on buyers’ purchasing behavior is evident in their ability to shape more discerning buyers, supporting the notion that favorable and well-formed beliefs about a manufacturer can positively impact its customers’ purchasing decisions (Bonner and Calantone, 2005). The bargaining power is also affected by other factors such as uncertainty, risk and business partners’ strengths. In scenarios where both parties possess significant sources of power, power is not used in a confrontational manner but rather as a means of fortifying the collaborative aspects of the business relationship (Gadde and Wynstra, 2018).

Partnership commitment refers to a firm’s dedication to maintaining a close and lasting relationship with another firm (Kim and Frazier, 1997). This commitment enables independent partners to work together, better serve customer needs and achieve higher performance levels (Morgan and Hunt, 1994). The measurement of commitment varies across partnerships and can take the form of an intention to continue the relationship, the willingness to make short-term sacrifices, confidence in the relationship’s stability, the relationship’s relevance or the internalization of the partner firm’s norms and values (Kim and Frazier, 1997; Kim et al., 2011).

Different dimensions of commitment elicit unique behaviors from partners. Affective commitment promotes extra-role behaviors, while calculative commitment undermines them, whereas normative commitment induces little change in extra-role behaviors (Gruen et al., 2000; Kim et al., 2011). Each effect on partners’ behaviors is because of distinct psychological responses associated with each type of commitment involved (Kim et al., 2011). Firms can also generate relational commitment by fostering alliances and promoting collaborative learning (Cheng et al., 2022). It can maximize the knowledge gained from partners and protect their business from being appropriately (Li et al., 2017).

Interpersonal relationships have recently increasingly interested B2B marketing researchers and practitioners (Aarikka-Stenroos et al., 2018; Wiatr Borg and Vagn Freytag, 2012). The essence of any interpersonal relationship lies in interaction (Kelley, 1979). However, interactions occur on various levels, in different contexts and for different reasons, with existing research suggesting that there is no best way to understand and manage them (Aarikka-Stenroos et al., 2018). Interpersonal relationships often relate to establishing those interactions, specifically in the B2B sales process setting. Research offers different perspectives to understand this dyad better, namely, the firms’ environment and strategies, the firm’s relationships and information gathering, the sales cycle and development or the sales characteristics and selling behavior (Méndez-Picazo et al., 2021; Wiatr Borg and Vagn Freytag, 2012). However, not all methods of understanding and managing interpersonal relationships are equally effective, and dealing with interpersonal relationships in a sales process depends on the level at which the analysis occurs (Crosby et al., 1990; Wiatr Borg and Vagn Freytag, 2012). Adopting an integrated perspective that encompasses all levels of analysis offers valuable insights into effectively addressing interpersonal relationships in B2B sales processes (Dampérat and Jolibert, 2009; Ellegaard and Andersen, 2015; Wiatr Borg and Vagn Freytag, 2012).

Brand sensitivity is a key concept in B2B customer behavior. It refers to the likelihood of choosing a well-known brand over a generic or unknown brand (Hutton, 1997). From a B2B perspective, it also represents the extent to which brand information and business associations are positively evaluated in organizational buying decisions (Brown et al., 2012). However, the prevailing view of the B2B buying process posits that firms are primarily rational decision-makers (Brown et al., 2011), differing in the importance they allocate to brands, which is a source of distress in designing B2B branding strategies (Sharma and Sengupta, 2020). To fully comprehend B2B customer behavior, it is crucial to examine brand sensitivity and understand under which conditions brands become more relevant in B2B contexts. Understanding the conditions under which the brand increases its relevance in B2B environments can help managers adapt their sales and marketing strategies and make them appealing to a specific set of target segments, helping to achieve the profitability goals more successfully (Zablah et al., 2010).

Brand sensitivity is a multi-faceted construct encompassing three dimensions: brand-related information acquisition, brand information processing and buying center memory. The nature of buyer–seller relationships and the number of supplier brands are also known to impact brand sensitivity (Sharma and Sengupta, 2020). Moreover, research suggests that brand sensitivity can exhibit a non-linear relationship with the importance of the purchase and its complexity (Brown et al., 2012). Thus, B2B marketing efforts should aim to establish strong brands and communicate their values in a market with multiple competing brands. This may require rethinking conventional wisdom and emphasizing the brand even when it may appear less relevant to business customers. Doing so is likely to impact performance positively (Brown et al., 2012; Sharma and Sengupta, 2020).

Procurement refers to the process of sourcing and acquiring goods or services from an external source, typically for business purposes (Laffont and Tirole, 1993). It is influenced by the buyer’s perception of their business strategies that affect their priorities, decisions and actions, serving as a mental model for the buyer to achieve a specific task (Strandvik et al., 2012). This process is becoming an essential component of firms’ acquisitions of external resources, with suppliers increasingly offering differentiated value propositions by incorporating services into their offerings (van der Valk, 2008). However, procurement is not a static or standard activity, and its context is constantly redefined by social and economic changes (Torvinen and Ulkuniemi, 2016).

In response to the changing competitive landscape, B2B organizations have adapted to innovative sales processes that align with new buying behaviors of B2B decision-makers, moving beyond traditional seller-oriented models (Strandvik et al., 2012). One of the emerging trends in procurement is the incentive to abandon the traditional practices of doing business and move forward to focus on relationship quality (Rauyruen and Miller, 2007), partnerships, networks and/or strategic alliances (Torvinen and Ulkuniemi, 2016). However, the sales process is not always seamless, as the seller’s value proposition may not always match the buyer’s value requirements. To be genuinely customer-oriented, firms must find ways to bridge this gap and create value for customers (Strandvik et al., 2012). Sellers aim to deliver value for customers, primarily from the solutions they sell and their skills and behaviors, while buyers have expectations about innovativeness, future orientation, long-term relationships and responsiveness to their specific needs (Kaski et al., 2017). Understanding the gaps between buyers and sellers can help recognize the significance of the sales process and value co-creation in B2B environments (Kaski et al., 2017; Loureiro et al., 2020).

Cultural differences result from variations in cultural values. They influence perceptions and play a relevant role in people’s behaviors (Hofstede, 1991). A company’s country of origin and market significantly impact customer behavior (Bilro and Cunha, 2021). Various researchers have extensively studied the relationship between business and culture (Armuña et al., 2020; Belchior and Lyons, 2021; Canestrino et al., 2020). Scholars agree on a positive correlation between business relationships and cultural similarity (Keep et al., 1998; Steward et al., 2010). The cultural analysis can be conducted from a single-country perspective, examining cultural factors, such as self-construal, individualism/collectivism and uncertainty avoidance (Laufer et al., 2005) or from a multi-country comparison, evaluating cultural belonging (i.e. collectivistic vs individualistic) effects on customers’ behavioral intentions (Baghi and Gabrielli, 2019). Research has demonstrated that individuals in a collectivistic culture perceive a prompt resolution of product failure as fairer, leading to higher customer satisfaction, than those in individualistic cultures (Muralidharan et al., 2019). Additionally, the international B2B relationships connect a firm’s national culture to behavior predispositions according to their cultural dimensions (Xu and Hao, 2021). The tension between culturally different partners when cooperating for the common benefit has significant consequences for both parties, endangering the relationship’s stability and contributing significantly to relationship failures (Bilro and Cunha, 2021; Gould et al., 2016; Xu and Hao, 2021).

Salespeople refer to the trade-in occupation within a firm, selling goods or services directly to customers or other businesses or organizations for monetary compensation. Sales can be conducted in-person (e.g. in retail stores or dealerships) or using online communication tools. A successful salesperson is perceived as someone skilled enough to persuade other people, especially in a business or professional setting, to buy their products (Delpechitre et al., 2019), which highlights the relevance of suppliers’ behavior in customer behavioral intentions, such as supplier loyalty or customer satisfaction (Blaese et al., 2021; Selnes and Gønhaug, 2000). While it is expected that salespeople have a good understanding of customer needs, research has shown that this is often not the case, with salespeople failing to provide an adequate value proposition to customers (Homburg et al., 2009; Rapp et al., 2014). These failures can result in less buyer satisfaction and commitment to the supplier (Kumar et al., 2013; Palmatier et al., 2007). For salespeople to be effective, they must have a clear understanding of customer expectations and act in a manner that satisfies those expectations, reducing failures and increasing positive outcomes in relational exchanges (Haas et al., 2012). Salespeople that can provide a proper value proposition to customers transform themselves into a valuable point of differentiation (Kaski et al., 2017).

Firms of all sizes and from all industry sectors are active buyers, and the selection of their supply chain is of foremost importance (Kim et al., 2010). The literature emphasizes the importance of quality, cost, delivery and flexibility attributes when choosing a supplier (Voss et al., 2009). Noteworthy, there seems to be a difference between the perceived value of these attributes and the actual practice, as the operational practices may not align with buyers’ strategic priorities (Verma and Pullman, 1998). Research suggests that managers responsible for supplier selection may prioritize cost and delivery capability over quality, an issue that deserves thoughtful attention (Alikhani et al., 2019).

Supplier selection also needs to be understood under the process stages, as buyers ground their purchasing behavior in several steps or stages before the supplier selection is made. Research highlights that it is essential to understand the choice phase (which is the most visible part), comprising the buyer problem acknowledgment, the criteria definition and the supplier qualification and the quality of the steps that precede it (de Boer et al., 2001). Additionally, the differences in supplier selection criteria and buyer behavior across various industry sectors should also be considered (Ghymn et al., 1999). However, it is essential to note that these differences do not dictate the suitability of a specific decision process – as more suitable for a specific sector – neither the specific industry nor the criteria used to determine the correctness of the buyer decision. Overall, situational characteristics, such as the number of suppliers available, the availability of historical information or the importance of the purchase, are more determinative of the suitability of that decision (Alikhani et al., 2019; de Boer et al., 2001).

Developing joint solutions through buyer–seller interaction requires meticulous attention from all involved parties. By fostering interaction and facilitating joint solution development and co-creation, firms can significantly increase their chances of success (Caruana et al., 2020; Vargo and Lusch, 2011). However, interaction can be frustrating if both parties adopt and promote transactional views of solutions instead of relational views (Tuli et al., 2007). Research shows that firms have been pursuing several approaches to improve the success of their interactions. One such approach is adaptive selling, which entails adjusting sales behaviors to enhance customer-oriented selling during interactions (Franke and Park, 2006). Another is interfirm adaptation and perspective-taking, which allows for a deep understanding of mutual needs and motivations to co-develop solutions that strengthen cooperation (Xu and Hao, 2021).

Other researchers have emphasized the relevance of a more interactive approach to inter-organizational relationships, such as inter-organizational cooperation based on the development of trust or commitment between the parties as precursors to cooperation (Heide and Miner, 1992). When executed effectively, such approaches result in cooperative relationships that are profitable and valuable for both parties involved (Kim et al., 2010).

Following prior systematic literature reviews (Cartwright et al., 2021; Hayes and Kelliher, 2022; Vrontis and Christofi, 2021), our research aims to review customer behavior in the B2B context and proposes a research agenda for future studies. Our study consolidates knowledge in this area and highlights several ways to improve its understanding. This research discusses its findings and future research agenda resorting to the well-known TCCM (Loureiro et al., 2021; Paul and Rosado-Serrano, 2019; Terjesen et al., 2016). Additionally, an integrative framework is presented to enable future researchers to formulate novel conceptual models (as depicted in Figure 4).

Three core theories are the most used as foundation support of the analyzed articles: SET, TCE and RET. Others are the resource-based view of firms and resource dependence. The SET establishes that social behavior results from an exchange process that maximizes benefits and minimizes costs (Anaza and Rutherford, 2014; Ellegaard and Andersen, 2015; Sales Baptista, 2014). Therefore, business partners tend to weigh social relationships’ potential benefits and risks. Business relationships require a long-term process, mutual respect and the acceptance of the other as a partner and co-producer of value, not just a passive element (Li et al., 2017). TCE focuses on cost and efficiency to stipulate a relationship and uses relationships as management structures to reduce hazards (Lui et al., 2009; Steinle et al., 2014). The resource-based view theory refers that the competitive advantage results from accumulated resources and capabilities that are unusual, valuable, non-substitutable and difficult to imitate by the firm’s competitors (Corsaro, 2015). This theory regards the firm as the primary unit of analysis (Mols, 2019). The RD theory analyses how the external resources, the internal resources and the organization’s capabilities affect the organization’s behavior (Bonner and Calantone, 2005).

Other theories less used can be suggested to support further development of this theme, such as Attribution Theory (Mir et al., 2017; Selnes and Gønhaug, 2000), Service-Dominant Logic (Aitken and Paton, 2016) and Cognitive Dissonance Theory (Kim et al., 2011). The power dependence theory is becoming more relevant in studies since 2001 to reflect the power of enduring relationships (Skarmeas and Katsikeas, 2001). This theory treats power as inherent in the relationship rather than the partners involved (Prior and Keränen, 2020). Although the theories of power can be considered (Meehan and Wright, 2012; Narayandas and Rangan, 2004), this one should be further explored in the future. New theories should be sought in different fields of knowledge and brought to the B2B relationships. The combination of different theories is also highlighted and can further add to the explanation of B2B.

Prior studies tend to rely more on buyers than sellers, and a small group (n = 29.14%) is devoted to the dyadic relationship (Lussier et al., 2017; Mir et al., 2017; Narayandas and Rangan, 2004). Therefore, we recommend more studies designed to capture the B2B relationship instead of focusing on a single partner or both independently. Prior studies tend to focus mainly on multi-manufactory industries, leading us to recommend more effort to understand the service sector. North America and Europe are the regions where most studies were conducted. Hence, new opportunities are open to study firms in other regions, particularly developing countries.

Many articles try to explore the drivers of B2B customer behavior (Figure 4). We grouped them into tangibility, environmental, organizational culture and relational behavior. Tangibility represents the quality of goods, services and distribution offered by one partner to another (Davis-Sramek et al., 2009; Voss et al., 2009). Switching costs, price and technical support are also analyzed, as long as the corporate reputation (both the firm and its brands) is considered (Dax et al., 2019; Hunter et al., 2006; Russo et al., 2017). Environmental category means the factors associated with not only the risk, uncertainty, pressure (Corsaro, 2015), cultural differences (Brush and Rexha, 2007) but also sustainability concerns (Prior and Keränen, 2020). Organizational culture focuses on internal factors of the organization, such as norms, values, organizational characteristics (Aitken and Paton, 2016; Kaski et al., 2017), how top managers deal with other employees and the structure of the organization (Valtakoski, 2015). The influence of rituals, norms, artifacts and the complete factors associated with organizational culture should be better analyzed for a more holistic understanding of how they influence inter- and intra-organizational relationships and behavior (Itani et al., 2020) and business purchase decisions, explore value congruency in the two partner organizations (Anwer et al., 2020) and multiple partner relationships (multi-dyadic relationships).

Relational behavior brings together the factors that influence B2B relationships. Thus, the concept of dyadic market-oriented relationships describes how the relationship evolves between partners (Aitken and Paton, 2016; Kaski et al., 2017) and how they create bonds and favorable emotional states to cooperate (Wong et al., 2010). Individual and social characteristics express the individual partner traits and the social skills to interact in dyadic relationships (Lichtenthal and Shani, 2000; Meehan and Wright, 2012). Particularly, organizational customers’ perceptions of supplier employees’ empathy (cognitive and affective) are still not well studied (Delpechitre et al., 2019; Selnes and Gønhaug, 2000).

The quality of communication in B2B relationships has been widely recognized as a critical factor that can influence the longevity of the partnership. It encompasses both the intrinsic qualities of the individuals involved and the methods by which firms disseminate institutional information both internally and externally (Doney et al., 2007; Graça and Kharé, 2020; Sinčić Ćorić et al., 2017). Relational switching costs can avoid the end of a B2B relationship, as they influence the partners’ share-of-wallet, cross-buying behavior and actual switching behavior (Blut et al., 2016). Incentives, such as monetary rewards or appreciation, can encourage partners to maintain their relationships (Tanner, 1996). Additionally, the power dynamic between partners also conditions the relationship’s longevity (Hunter et al., 2006; Narayandas and Rangan, 2004).

A proliferation of outcomes is analyzed in previous research, which we aggregate as firm and relational. Firm outcomes are associated with maintaining the relationship by continuing to purchase, recommend to others or the willingness to pay price premium (Brown et al., 2012; Dong et al., 2017), as well as the performance achieved by the firm or the salesperson (Briggs and Grisaffe, 2009; Chaithanapat et al., 2022; Ng, 2010). Brand sensitivity is an essential factor in inducing behavior in the partner and represents a primary emotion felt about a brand (Brown et al., 2012). Brand sensitivity is still in the early stage of their knowledge. More research must be devoted to creating a proper measurement tool and exploring drivers as individual factors (e.g. stakeholders and personal preferences) or information quality (Sharma and Sengupta, 2020).

Brand equity gives a brand position in terms of value and can generate more revenue when its equity is higher (Bonner and Calantone, 2005). From the relational perspective, we can point out the relationship quality (trust, commitment and satisfaction), which can also act as a mediator between the drivers and other outcomes shown in Figure 4 (Homburg et al., 2005), the power intensity (Meehan and Wright, 2012) and the switching behavior (Blut et al., 2016; Wathne et al., 2001). Other outcomes are suggested, like cost, quality and flexibility of production of the manufacturers and sales performance of the distributors (Li et al., 2017), to refine the specific marketing elements that lead suppliers to make decisions. Researchers should be concerned with design studies that can observe actual behaviors in interaction situations (Kemp et al., 2020) to complement the data collected through interviews and cross-sectional approaches.

The engagement process among all stakeholders can act as a mediator in understanding the decision process between drivers and outcomes of the B2B relationship (Kim et al., 2011; Loureiro et al., 2020; Prior and Keränen, 2020), but this concept and process have not been appropriately studied until now. Although past research addresses the concepts of cooperation and interdependence, more research is welcome to show how these concepts occur and influence relationships and decision-making. The dark side of relationships has been somehow ignored. We highly recommend understanding how to handle and restore non-trust, non-commitment or non-satisfaction situations. When attraction shifts to avoidance, what to do? The concept of sustainability and its influence on B2B relationships and decision-making needs an in-depth study.

Concerning moderators, we categorize them as relationship strength and market. The former is devoted to the length, longevity, socialization, frequency of contact or relationship history (Gould et al., 2016). The latter deals with the type of market, the firm and product characteristics and the competitive intensity (Bode et al., 2011; Brown et al., 2011). Some control variables somehow overlap with the moderators, for instance, the characteristics of the firm (firm size, age, type and power), relationship length and duration, market uncertainty or even the product characteristics (size, complexity and type) (Li et al., 2017; Lussier et al., 2017). However, control variables also consider the country of origin (Reardon et al., 2017) or employee socio-demographics (Kemp et al., 2020). How product innovativeness strengthens or weakens the relationship between organizations and buyer purchase behavior is a moderator not yet explored (Bonner and Calantone, 2005). Other moderators can be profit and non-profit organizations, public and private organizations or even levels of technology incorporated (Lakshmi and Bahli, 2020; Nedjah et al., 2022).

The evolution of technologies and their incorporation into organizations are changing the relationships, particularly with artificial intelligence (AI) agents (Guaita Martínez et al., 2022; Liu, 2020). Robots have been used in industry to develop repetitive tasks rapidly, and AI systems are used to treat a large amount of information (big data). In addition, they are also being incorporated into the context of the service, operating as a virtual assistant (using voice or text) and embedded in a human-like robot and interacting with other employees. Thus, new avenues are open to exploring the multiple interactions between humans and AI agents and between two AI agents. Saura et al. (2021) present diverse research questions to be considered in the future, organized by sender’s cues, training and recruiting, organizational strategy and structure, suitability of digital interaction and dark side of digital sales interactions.

Structural equation model, regression analysis and confirmatory factorial analysis are widely used in this area of research (Crosno et al., 2020; Dong et al., 2017; Russo et al., 2017). Although some studies also conducted interviews (Friend and Johnson, 2017; Zondag and Brink, 2015), experiments (Cardozo and Cagley, 1971; Mir et al., 2017), conjoint analysis (Chakraborty et al., 2007; Wuyts et al., 2009) or case studies (Bolton and Myers, 2003; Krause and Ellram, 2014), the challenge is to go further and develop mixed-approach methods. We recommend the fuzzy-set qualitative comparative analysis to explore complex relationships and moderating effects.

Regarding samples, the size depends on the methodological tool used. Thus, case studies tend to be three on average per article. In structural equation model, the sample (using a cross-sectional approach) ranges between 100 and 600. Interviews have between 10 and 70 participants. Conjoint analysis and cluster analysis have between 100 and 200 participants. Experiments have between 50 and 100 participants. Therefore, previous studies lack longitudinal data collection, field experiments or multiple-case studies, which needs to be considered in future studies. One limitation of previous studies also pointed out as a possible research avenue is a need for collecting more data and the replication of models in other contexts to allow generalization (Davis-Sramek et al., 2009; Salo and Wendelin, 2013).

Research about B2B has adopted different perspectives about customer behavior in this domain, resulting in a rich body of literature trying to understand this phenomenon. Based on an overview of the 219 papers analyzed through a systematic literature review, this paper enhances the existing knowledge about B2B customer behavior, identifies and summarizes the main topics in this field of research, contributes to an explanatory framework, identifies future research avenues and offers a research agenda.

There is still no similar systematic review on the topic to the best of the author’s knowledge. Research has been conducted to map B2B advertising (Swani et al., 2020), B2B market segmentation (Mora Cortez et al., 2021) or B2B branding (Leek and Christodoulides, 2011), but a systematic review of customer behavior in the B2B domain is still not available, which puts in evidence the timely and relevant contribution of this paper. This novelty can benefit scholars and practitioners, who can take advantage of our integrative viewpoint on this topic. This paper offers a systematic review and applies a text-mining procedure using R software to adequately capture the relevant topic discussed and open new research avenues, which is still uncommon, particularly in the marketing and B2B domains. Moreover, our initial descriptive analysis offers a helicopter view of the type and number of studies already conducted, the journals publishing in this field of research and the geographic coverage of the empirical studies. This information allowed future researchers to understand the prior studies’ characteristics and the potential journals that can be open to receiving more studies on the topic. Indeed, through the support of the TCCM framework, this literature highlights several gaps for future research, such as a firmer theoretical foundation and development, better contextual positioning or more exploratory methodologies. Our arguments may also assist practitioners in understanding the various direct and indirect connections between antecedents and outcomes of customer behavior in B2B settings, helping to formulate appropriate marketing strategies in a structured and systematic way, such as the insights gained from the analysis of buyer–supplier relationships, the role of supplier selection, cooperation and interactions or the effect of commitment and cultural differences.

This paper proposes an integrated overview of customer behavior in B2B that can be useful for practitioners and academics in future endeavors. We offered the detailed implications for future research agenda in previous Section 5, which we summarized here in Table 4.

As with any systematic review, this paper’s findings should be taken with caution within the context of this method’s limitations. The review has resorted to the WOS and SCOPUS databases to assess the quality of publications. Although it comprises diverse publishers (e.g. Emerald, Sage, Elsevier, Wiley or Taylor & Francis), one may assume that interesting research may not be incorporated in the final pool of papers analyzed as conference proceedings papers or other non-top tier publications. Second, the keywords used may limit the process even if inspired by top reference articles. Additionally, the screening process may have other biases, such as the researchers’ data handling. However, the authors believe that the rigorous procedure of this systematic review has reduced the probability that the omitted research would have contained information that would critically alter our conclusions. The gaps and avenues for future research have been identified through the TCCM framework, and the authors prepared a summative table (Table 4) compiling the suggestions of several prior studies that have not been achieved so far.

Funding: This research did not receive any specific grant from public, commercial or not-for-profit funding agencies.

Table A1