Configurational Pathways to Digital Transformation in Human Resource Service Firms: A Grounded Theory and fsQCA Study in China

Abstract

Digital transformation is reshaping the value creation and competitive logic of human resource service (HRS) firms, yet its antecedents are commonly examined as independent net effects. This study adopts an exploratory mixed-method design to identify and test configurational pathways to digital transformation. First, grounded theory analysis of interviews with 20 HRS firms in Beijing identified five antecedent conditions: digital technology application, digital strategic planning, firm social capital, the digital policy environment, and competitive intensity. Second, fuzzy-set qualitative comparative analysis (fsQCA) was applied to survey data from 97 firms. The analysis yielded five solution terms for high digital transformation and two configurations for non-high digital transformation. The five high-outcome solution terms were organized into three broader configurational patterns: technology–strategy synergy, technology-driven competition transformation, and strategy–ecosystem synergy. The non-high-outcome configurations reflected either weak digital foundations combined with ecosystem disengagement or relational-resource maintenance accompanied by transformation inertia. No single condition was necessary for either outcome, and the configurations associated with high and non-high digital transformation exhibited causal asymmetry. The findings show that digital transformation in HRS firms depends on the alignment of technological, organizational, and environmental conditions rather than on technology investment alone.

1. Introduction

The rapid development of big data, cloud computing, artificial intelligence, and other digital technologies is reshaping how firms allocate resources, create value, and define competitive boundaries. Digital technologies have enabled new business models and organizational forms while compelling firms to reassess established operating models and sources of competitive advantage [1]. Digital transformation has therefore become more than a defensive response to technological change; it is a strategic process through which firms redesign business processes, innovate value creation, and support long-term development [2,3]. Rather than representing the mere deployment of information technology, digital transformation entails the systematic reconfiguration of strategy, resources, organizational processes, and business models around digital technologies [4].

Human resource service (HRS) firms provide a distinctive setting in which to examine this process. Their activities center on information search, talent assessment, labor-market matching, and relationship coordination. They are knowledge-intensive, information-intensive, interaction-intensive, and highly dependent on professional judgment. Data analytics, intelligent assessment, algorithmic recommendation, and digital platforms can reduce search and matching costs and improve recruitment, assessment, training, flexible staffing, and career services. At the same time, HRSs cannot be reduced to standardized information processing. Their value continues to depend on sector-specific expertise, professional judgment, client trust, relational networks, and institutional legitimacy. Digital transformation in this sector therefore reflects an interaction between the efficiency logic of digital technologies and the professional, relational, and regulatory logics of HRS provision.

Beyond improving firm efficiency and competitiveness, digital transformation in HRS firms also has broader implications for sustainable development. By enhancing information processing, talent matching, and digital service delivery, HRS firms can reduce labor-market information frictions, expand access to employment services, and improve the efficiency and inclusiveness of human resource allocation. These changes can support service innovation and broaden access to labor-market opportunities, thereby contributing to more sustainable human resource service provision. From this perspective, digital transformation represents not only organizational upgrading but also a potential pathway toward sustainable value creation in the HRS sector.

This interaction creates a transformation paradox. On the one hand, HRS activities have substantial potential for digitalization and platformization, allowing firms to overcome the scale constraints of labor-intensive service delivery. On the other hand, the conversion of digital technologies into process redesign and value creation depends on a clear digital strategy, mobilizable relational resources, a supportive policy context, and the capacity to respond to competition. Firms operating under similar technological and institutional conditions may consequently display markedly different transformation outcomes. Established providers may possess extensive client networks, accumulated industry knowledge, and institutional resources but remain constrained by organizational inertia and rigid processes. Technology-oriented entrants may possess advanced digital capabilities but lack client trust, industry legitimacy, or relational resources. These differences suggest that digital transformation is not a linear sequence from technology investment to performance; it is a complex process of alignment among technological, strategic, relational, and environmental conditions.

China’s rapidly developing digital economy offers a useful institutional and industrial context for studying these configurations. The digital economy, industrial digitalization, and the upgrading of modern services have received sustained policy support. As an intermediary sector linking human capital with the real economy, the HRS industry is increasingly expected to improve labor-market allocation and modernize service systems through digital transformation. Beijing combines a dense concentration of science and technology resources, HRS providers, policy resources, and market competition. Its HRS ecosystem includes integrated service providers, executive-search firms, flexible staffing platforms, online recruitment companies, assessment firms, and digitally enabled training providers. The coexistence of diverse ownership structures, firm sizes, and business models provides a suitable setting for examining how different resource endowments generate alternative pathways to digital transformation.

Existing research has identified technological infrastructure, organizational resources, strategic capabilities, government support, and competitive pressure as important antecedents of digital transformation. However, the complex mechanisms through which these factors combine in HRS firms remain underexplored. First, much of the literature follows a variable-centered linear logic that estimates the independent net effect of each factor. Such models are useful for assessing average associations, but they tend to treat other conditions as separable background variables and are less able to capture complementarity, substitution, or suppression among technological, organizational, and environmental conditions [5,6]. Firms do not possess identical resources, and high digital transformation may be achieved through different combinations rather than a single best practice.

Second, the digital transformation literature has concentrated heavily on manufacturing and broad service-sector samples. In manufacturing, transformation frequently concerns equipment, production, supply chains, and operational systems. HRS firms instead depend on talent data, specialized knowledge, client relationships, reputation, and institutional credentials. Whether technology produces substantive transformation therefore depends not only on technical tools but also on digital strategy, social capital, policy support, and competitive conditions. Applying generic transformation models without adaptation may overlook the relational and institutional foundations of this industry.

Third, theory-driven variable selection may fail to capture locally salient conditions in a policy-intensive and network-embedded industry. An empirical strategy that first derives antecedent conditions from firms’ transformation practices and then integrates them within an established framework can reduce this risk. Moreover, prior studies have paid more attention to the presence of high transformation than to its absence. The conditions that generate a high outcome cannot be assumed to have mirror-image effects when absent.

Accordingly, this study addresses three questions: (1) Which antecedent conditions are central to the digital transformation of HRS firms? (2) How do technological, organizational, and environmental conditions combine into configurations associated with high and non-high digital transformation? (3) Do the configurations associated with high and non-high digital transformation exhibit causal asymmetry?

We adopt an exploratory mixed-method design [7]. In the qualitative phase, interview data from 20 Beijing-based HRS firms were analyzed using grounded theory procedures to identify contextually grounded antecedent conditions. The five conditions were subsequently organized within the TOE framework: digital technology application represents the technological context; digital strategic planning represents the organizational context; and firm social capital, the digital policy environment, and competitive intensity capture the relational, institutional, and market aspects of the environmental context, respectively. In the quantitative phase, fuzzy-set qualitative comparative analysis (fsQCA) was applied to survey data from 97 HRS firms to identify sufficient configurations for high and non-high digital transformation.

This study makes three contributions. First, it extends digital transformation research by revealing the context-specific role of relational resources in the HRS industry. Whereas much prior research, particularly in manufacturing settings, emphasizes technological readiness, production systems, and operational integration, HRS firms depend more heavily on client relationships, institutional connections, professional reputation, and access to external resources. Our findings show that firm social capital is therefore an important component of the HRS transformation context, but not a universally sufficient or uniformly beneficial condition. Its contribution depends on how it is combined with digital technology application, strategic planning, policy support, and competitive pressure. Second, the findings identify three broader configurational patterns—technology–strategy synergy, technology–competition alignment, and strategy–ecosystem coordination—showing that HRS firms can reach high digital transformation through different forms of complementarity rather than through a single dominant model. Third, the sequential use of grounded theory and fsQCA enables the study to connect contextually grounded condition identification with configurational explanation. Its value therefore lies not simply in combining two methods, but in showing how conditions derived from HRS firms’ actual transformation practices operate jointly across firms.

2. Literature Review and Theoretical Framework

2.1. Conceptualizing Digital Transformation

Scholarly understanding of digital transformation has evolved from a technology-tool perspective toward a view of strategic and organizational change. Early work commonly treated digitalization as the conversion of information into digital form, the online redesign of processes, or the use of information systems to improve efficiency. As big data, cloud computing, artificial intelligence, and mobile connectivity have become embedded in organizational activity, technology adoption alone has become insufficient to explain changes in strategy, structure, and value creation [8].

The literature typically distinguishes digitization, digitalization, and digital transformation. Digitization converts physical information into machine-readable digital data. Digitalization uses digital technologies to improve existing processes and activities. Digital transformation goes further by combining information, computing, communication, and connectivity technologies to produce systemic changes in strategic direction, organizational structures, business processes, customer relationships, and value creation [9]. It therefore involves the integration of technology with strategy, organizational capabilities, and business models rather than the accumulation of digital tools.

HRS firms represent a particularly relevant context. Their core activities involve information search, talent evaluation, matching, and relationship coordination. Digital technologies can integrate data, automate assessments, improve recommendations, and connect labor-market participants, thereby increasing the efficiency and responsiveness of recruitment, assessment, training, flexible staffing, and career services [10,11]. Yet the quality of these services remains dependent on professional expertise, context-sensitive judgment, client trust, relational networks, and institutional qualifications. Digital transformation in HRS firms therefore combines digital efficiency with professional service, relational trust, and regulatory compliance.

Three interrelated forms of transformation can be observed. The first is process digitalization, in which enterprise resource planning, human resource management systems, electronic HRM, and related tools move payroll, attendance, recruitment, and contract management online. The second is intelligent decision support, in which integrated talent, client, and operating data are used for talent profiling, algorithmic matching, turnover prediction, and personalized services [10,12,13]. The third is platform-based service reconfiguration, through which digital platforms connect client firms, workers, and other service actors and reposition HRS providers as orchestrators of labor-market resources and service ecosystems [14,15,16,17,18]. These forms need not occur as a fixed sequence; firms may develop them in parallel, skip stages, or remain concentrated on one layer. In this study, digital transformation is defined as a systemic organizational change through which HRS firms use digital technologies to reconfigure business processes, decision mechanisms, service relationships, and value creation and thereby move from traditional information intermediation toward data-driven, intelligent, and platform-enabled service provision [19,20].

2.2. Antecedents of Digital Transformation

Digital transformation does not arise automatically from a single technology. It reflects interactions among internal resources, organizational agency, and the external environment [12]. Technologically, digital infrastructure and the capacity to apply digital tools provide the material foundation for data collection, system connectivity, and service innovation. The embedding of artificial intelligence, data analytics, and cloud technologies in core service processes determines whether technical resources can improve workflows and decision making [10,21]. Digital platforms and external technology partnerships may also reduce access costs for resource-constrained firms [22]. Nevertheless, ownership of technology does not guarantee transformation when applications remain fragmented or disconnected from strategy and service contexts.

Organizationally, strategic cognition, resource allocation, and organizational adaptation shape the depth and continuity of transformation. Digital transformation involves substantial investment, extended implementation periods, cross-functional consequences, and uncertain returns. Firms therefore need to position digitalization within long-term strategy, prioritize projects, and adjust structures, processes, and talent allocation. Dynamic capability research similarly emphasizes that firms must sense digital opportunities, seize market changes, and reconfigure resources in response to disruption [23].

Environmentally, relational networks, policy institutions, and market competition create important resources, opportunities, and pressures for digital transformation. Firm social capital is particularly relevant in HRSs because value creation is strongly embedded in relationships. Client networks, business partners, industry associations, and government agencies can provide information, data, technical cooperation, legitimacy, and access to projects. The ability to build and mobilize these external relationships therefore shapes firms’ access to resources and opportunities beyond their organizational boundaries. Digital economy policies, industrial support, subsidies, public digital infrastructure, and human capital programs can further lower adoption costs and signal development priorities. Policy transparency, accessibility, implementation fairness, data protection, and regulatory uncertainty also influence the willingness and cost of transformation. Competitive pressure from digital platforms, changing client expectations, rapid imitation, and price competition can erode the advantages of labor-intensive service models and encourage intelligent matching, service innovation, and platform development [16]. Competition may nevertheless crowd out long-term investment when firms lack internal capabilities.

These conditions are interdependent. The same technology may yield different outcomes under different strategic and relational conditions; policy support requires absorptive and implementation capacity; and competition may either trigger transformation or intensify resource constraints. Digital transformation is therefore more plausibly understood as the outcome of complementary, substitutive, and context-dependent condition combinations [24]. Resource orchestration research further suggests that firms create digital value by structuring, bundling, and leveraging technological and relational resources rather than by simply possessing them [25].

2.3. Research Gaps

Prior research has identified technological capability, strategic orientation, relational resources, policy support, and competitive pressure as important antecedents of digital transformation. However, their effects are often contingent rather than uniform. The same technological resources may generate different outcomes under different strategic and relational conditions; policy support depends on firms’ capacity to absorb and implement external resources; and competitive pressure may stimulate transformation while also intensifying resource constraints when internal capabilities are weak. These findings suggest that the antecedents of digital transformation are interdependent and that apparently inconsistent findings across studies may reflect differences in how these conditions are combined.

This issue is particularly relevant in HRS firms, where transformation depends not only on technology but also on talent data, professional knowledge, client relationships, institutional connections, and policy resources. Variable-centered approaches are less able to capture such complementarity and substitution among conditions, or to explain why different combinations may be sufficient for similar outcomes. Moreover, configurations associated with high digital transformation need not be mirror images of those associated with non-high transformation. A configurational approach is therefore appropriate for examining equifinality, conjunctural relationships, and asymmetry in the digital transformation of HRS firms.

2.4. TOE Framework and Configurational Perspective

The TOE framework, introduced by Tornatzky and Fleischer, explains organizational innovation through three interrelated contexts: technology, organization, and environment [26]. The technological context concerns available infrastructure, technological maturity, application capability, and fit with business processes [27]. The organizational context concerns resources, structure, managerial support, strategic orientation, learning, and adaptability [28]. The environmental context includes policy, regulation, clients, competition, and industry ecosystems. The framework is sufficiently open to accommodate sector-specific variables and is particularly suitable for organization-level digital transformation [29].

A configurational perspective complements the TOE framework by treating organizational outcomes as associated with configurations of interdependent conditions rather than isolated predictors. Three principles are central. Conjunctural causation means that the role of a condition depends on the other conditions with which it is combined. Equifinality means that distinct configurations may be sufficient for the same outcome. Causal asymmetry means that configurations sufficient for the presence of an outcome need not be mirror images of those sufficient for its absence [6]. fsQCA operationalizes these principles using set theory and Boolean algebra and is therefore appropriate for identifying alternative sufficient pathways to high and non-high digital transformation.

3. Materials and Methods

3.1. Qualitative Phase: Identification of Antecedent Conditions

3.1.1. Research Setting and Sample Selection

The qualitative phase sought to identify antecedent conditions with explanatory relevance to the HRS context. Because established digital transformation variables may not fully capture the importance of talent data, professional knowledge, client relationships, industry credentials, and institutional resources, we employed Straussian grounded theory procedures to derive core categories inductively.

Beijing was selected because of its concentration of digital industries, science and technology resources, HRS organizations, and policy institutions. The local industry includes integrated providers, recruitment platforms, assessment firms, executive-search agencies, flexible staffing companies, and digital training providers. This diversity creates substantial variation in technological foundations, relational resources, access to policy support, and transformation stage.

Purposive sampling was combined with maximum variation sampling. Between March and June 2026, the research team contacted candidate firms with support from industry associations and professional networks. Firms were included when they had operated for at least three years, had initiated digital technology application or transformation activities, could provide a key informant familiar with strategy or digital operations, and contributed variation in size, ownership, service type, and transformation stage. Firms unable to provide substantive transformation information or knowledgeable informants were excluded. The final qualitative sample comprised 20 HRS firms (

Table 1. Profile of the interviewed firms.

Anonymous Firm	Business Positioning	Anonymous Firm	Business Positioning
C01	Integrated HRS provider	C11	Integrated HRS provider
C02	Regional HRS provider	C12	Executive-search provider
C03	Employment and talent-allocation provider	C13	Integrated HR outsourcing provider
C04	HR outsourcing provider	C14	Digital training provider
C05	Talent assessment and consulting provider	C15	Talent development provider
C06	Regional talent-ecosystem operator	C16	Vertical-market HRS provider
C07	Group-affiliated HRS provider	C17	Flexible staffing provider
C08	Digital recruitment platform	C18	Integrated HRS provider
C09	Online recruitment platform	C19	Integrated HRS provider
C10	Digital-technology-driven provider	C20	Integrated HRS provider

).

3.1.2. Interviews and Multi-Source Data Collection

Semi-structured in-depth interviews constituted the primary data source. Interviewees included chairpersons, general managers, heads of digitalization or information systems, and executives familiar with digital business operations. The interview guide covered the external environment, strategic positioning, technology application, resource acquisition, organizational adjustment, business-model change, and transformation outcomes (

Table 2. Illustrative semi-structured interview guide.

Interview Module	Core Interview Questions
Firm background and external context	1. Please describe the firm’s main services, client groups, and recent development.
	2. What major changes have occurred in the HRS industry, and how have they affected the firm?
	3. What opportunities or constraints arise from Beijing’s industrial, talent, and policy environment?
Transformation motives and strategic positioning	4. Under what circumstances did the firm begin its digital transformation?
	5. Which business or management problems was the transformation intended to address?
	6. What role does digitalization play in corporate strategy, and has this role changed?
	7. How are digital priorities, investment sequencing, and resource allocation determined?
Technology application and implementation	8. In which business and management scenarios are digital technologies currently applied?
	9. What roles do AI, algorithms, and data analytics play in core processes?
	10. How does the firm acquire technologies and develop systems, and why were these approaches chosen?
	11. What major difficulties have arisen in digital projects, and how were they addressed?
	12. How are client and talent data collected, integrated, used, and protected?
Organizational adjustment and resource coordination	13. How has transformation affected organizational structure, workflows, and job division?
	14. What new employee capabilities and talent structures are required?
	15. What roles have senior managers and business units played?
	16. From which external actors has the firm obtained resources or support?
	17. How do relationships with government, industry bodies, and partners affect transformation?
Transformation outcomes and future development	18. How has transformation changed operations, services, and the business model?
	19. How does the firm assess transformation outcomes?
	20. Have any projects failed to meet expectations or produced new risks?
	21. What are the main bottlenecks and future priorities?

).

Each interview lasted approximately 1.5–2 h. With participants’ consent, interviews were recorded and transcribed. Firm and participant identities were anonymized, and firms were coded C01–C20. The final corpus contained approximately 250,000 Chinese characters of interview transcripts.

To reduce retrospective and single-source bias, the research team also collected corporate websites, social-media archives, promotional documents, authoritative media reports, industry reports, and policy documents to triangulate the interview evidence. Publicly verifiable claims concerning digital systems, partnerships, qualifications, and policy projects were cross-checked against these materials. Discrepancies were resolved by returning to the interview context or through follow-up communication.

3.1.3. Coding and Analysis

The qualitative data were analyzed following the systematic grounded theory procedures proposed by Strauss and Corbin, including open coding, axial coding, and selective coding [30]. The collected interview and secondary data were systematically examined through these three sequential coding stages, through which core concepts and theoretical categories were progressively derived from the original material, ultimately forming a theoretical framework that explains the key antecedent conditions of firms’ digital transformation.

During open coding, the interview transcripts and Supplementary Materials were examined line by line to identify meaningful statements related to firms’ digital transformation practices. Repetitive or purely conversational expressions were removed, composite statements were divided into separate meaning units, and initial labels were assigned as closely as possible to the empirical content [31]. Labels with similar meanings were compared and consolidated, whereas ambiguous concepts were re-examined against their original context and other data sources. This process generated 131 initial concepts. To enhance coding transparency, representative interview excerpts and corresponding open codes are provided in Appendix A.

Axial coding was then used to identify relationships among the initial concepts and to organize them into higher-order categories on the basis of common attributes, behavioral patterns, and underlying mechanisms [32]. For example, data-supported decision making, identification of key talent attributes, and algorithm-based person–job matching were grouped into the category of algorithm-assisted decision making. Similarly, digital foresight, managerial awareness of digitalization, and long-term value orientation were integrated into strategic cognition. Through repeated comparison and refinement, the 131 initial concepts were consolidated into 24 axial categories.

Selective coding further integrated the axial categories around the focal phenomenon of digital transformation [33]. Categories were compared in terms of their conceptual scope, explanatory relevance, and relationships with other categories. Five core antecedent categories were ultimately identified: digital technology application, digital strategic planning, firm social capital, digital policy environment, and competitive intensity. These categories capture, respectively, the technological foundation, strategic orientation, relational resources, institutional environment, and market pressure associated with digital transformation in HRS firms. Their conceptual structure and representative empirical indicators are summarized in

Table 3. Grounded theory coding structure for the antecedent conditions.

Selective Category	Axial Category	Illustrative Open Codes
Digital technology application	Business-scenario embedding	Digital recruitment processes; online talent assessment; intelligent person–job matching; digital training services; online client delivery; scenario-based feature iteration
	Algorithm-assisted decision making	Data-supported business decisions; identification of key talent attributes; algorithmic person–job matching; talent-mobility prediction; intelligent recommendation
	Digital system integration	Integration of fragmented systems; linkage of recruitment and talent databases; cross-department system connection; external platform interfaces; cloud platforms; unified data management
	Technology development accumulation	Increasing R&D investment; digital technology teams; proprietary intellectual property; collaborative development; continuous product upgrading
Digital strategic planning	Strategic cognition	Senior-management attention to digitalization; recognition of digital trends and traditional business limitations; long-term digital value orientation; transformation consensus
	Digital strategic positioning	Integration of digitalization into corporate strategy; movement from business support to strategic leadership; explicit digital objectives and plans
	Digital resource allocation	Higher priority for digital projects; continuing investment; allocation of digital talent; dedicated units; cross-functional coordination
	Digital business portfolio	Selection of digital business domains; focus on vertical service scenarios; platform exploration; data-enabled services; upgrading of traditional services
Firm social capital	Government ties	Communication with government agencies; policy information; interdepartmental coordination; participation in public projects; government–business cooperation
	Industry embeddedness	Participation in associations and alliances; cooperation with authoritative organizations; standards development; access to industry information and resources
	Business-partner collaboration	Joint development with technology firms; platform data authorization; data cooperation; shared clients and channels; ecosystem partnerships
	Client relationship and trust	Long-term clients; service reputation; client retention; repeat cooperation; deep client relationships; brand trust
Digital policy environment	Policy resource provision	Digital project support; subsidies; tax incentives; public service platforms; digital talent support
	Policy information accessibility	Timely policy information; effective communication; clear interpretation; accessible consultation channels
	Implementation convenience	Simple application procedures; timely implementation; clear eligibility; implementation fairness; government service efficiency
	Digital regulatory environment	Clear data compliance requirements; changes in flexible staffing regulation; administrative approval changes; stronger digital-business regulation; uncertain policy boundaries
Competitive intensity	Rivalry pressure	Growing service homogeneity; rapid imitation; market-share competition; shrinking traditional markets; more competitors
	Platform substitution pressure	Entry of digital-native platforms and cross-industry technology firms; pressure from online recruitment platforms; substitution by platform services
	Price and bargaining pressure	Intense price competition; unclear pricing standards; stronger client bargaining power; reduced pricing power and margins
	Client demand upgrading	Requirements for system integration; higher delivery standards; more customized demand; faster response; demand for data services
	Competition for projects	More intense bidding; higher preparation costs; delayed project information; higher entry thresholds; lost opportunities

.

The five core categories can be further organized according to the TOE framework. Digital technology application represents the technological context because it captures the extent to which digital technologies, data resources, and digital systems are embedded in firms’ service and operational activities. Digital strategic planning represents the organizational context because it reflects firms’ internal strategic orientation, managerial priorities, and resource allocation for digital transformation. Firm social capital, the digital policy environment, and competitive intensity are situated within the environmental context. Specifically, firm social capital reflects firms’ relational embeddedness in the HRS ecosystem and the external resources accessible through relationships with business partners, industry organizations, government agencies, and other relevant actors; the digital policy environment captures institutional support and regulatory conditions; and competitive intensity reflects market pressure arising from competitors and changing client demands. This classification is consistent with the TOE perspective, in which the environmental context encompasses external actors and conditions surrounding the focal organization, including industry relationships, market forces, and governmental institutions. Together, the five categories provide a contextually grounded representation of the technological, organizational, and environmental conditions associated with digital transformation in HRS firms.

A further examination of the interview narratives provided additional process-oriented evidence on how the identified conditions unfolded in relation to firms’ digital transformation. Digital technology application often developed through the gradual accumulation of systems, data, and implementation experience before broader transformation took shape. For example, C12 described developing internal systems around business needs at an early stage and accumulating data through ongoing operations. Digital strategic planning could also emerge before substantial implementation. In C07, digital transformation had already been incorporated into the firm’s draft Five-Year Plan when its digital foundation remained relatively weak, suggesting that strategic orientation preceded large-scale digital deployment. External relational resources likewise accumulated over time. C14 described long-standing cooperation with universities, industry associations, local governments, and enterprises as an important basis for the subsequent expansion of service activities and access to broader resource networks. Policy and regulatory changes also shaped the sequence of transformation activities. As illustrated by C13, changes in policy and compliance requirements were followed by adjustments in business practices and, in some cases, the development of related digital platforms. Competitive pressure showed a similar process pattern. C16 described growing concern about falling behind as clients and competitors advanced their own digital transformation, which increased the perceived need to improve existing services and explore digital solutions. Taken together, these cases reveal a recurring pattern in which technological foundations, strategic orientation, relational resources, policy conditions, and competitive pressures were already present or evolving during the initiation and development of firms’ digital transformation.

3.1.4. Theoretical Saturation

Throughout the coding process, the research team adhered to the principles of coding consistency and constant comparison. Interview materials, emerging concepts, and categories were repeatedly compared to identify similarities and differences, refine conceptual boundaries, and clarify relationships among categories. To enhance coding reliability, two doctoral students in human resource management independently coded the materials from the 20 focal firms and subsequently compared their coding results. Any disagreements regarding coding labels, conceptual interpretations, or category assignments were resolved through discussion and re-examination of the original interview materials, with the coding scheme refined when necessary.

Following the three-stage coding process based on the 20 focal firms, theoretical saturation was further assessed using interview materials from eight additional HRS firms. These additional materials were examined against the established coding framework using the same comparative logic. Saturation was considered achieved when the additional interviews no longer generated new core categories, required substantial revision of existing category boundaries, or altered the relationships among the established categories [34]. The eight additional interviews provided further empirical examples and contextual variation within the existing categories but did not result in the emergence of new core categories or changes to the five-category framework. Therefore, the theoretical model was considered to have reached saturation.

3.2. Quantitative Phase: Survey and fsQCA

The study adopted a sequential mixed-methods design in which the qualitative and quantitative phases were analytically connected but drew on independent samples. The 20 firms included in the grounded-theory phase were not included in the subsequent survey sample. The grounded-theory analysis identified five contextually relevant antecedent conditions—digital technology application, digital strategic planning, firm social capital, the digital policy environment, and competitive intensity. These constructs were subsequently operationalized using established scales adapted to the HRS context and specified as the antecedent conditions in the fsQCA model, with digital transformation as the outcome condition.

3.2.1. Survey Sample and Data Collection

The survey targeted HRS firms in Beijing. Questionnaires were distributed through the Beijing human resource services industry network, alumni contacts, and firms employing MBA participants. Because the study concerned firm-level digital strategy, technology, external relationships, and operating conditions, respondents were selected using a key-informant approach. Eligible respondents included chairpersons, general managers, heads of digitalization or information systems, and senior managers familiar with the firm’s digital business.

Before formal administration, three researchers and managers reviewed the questionnaire for clarity, construct meaning, and industry relevance. Five eligible respondents then completed a pilot survey. Ambiguous, repetitive, or contextually inappropriate wording was revised, and pilot responses were excluded from the final sample.

A total of 125 questionnaires were collected. Responses were excluded when the respondent was not a sufficiently knowledgeable key informant; the firm’s main business was not HRS provision; the questionnaire contained extensive missing values, patterned responses, or implausibly short completion times; or multiple responses came from the same firm, in which case the most complete response from the most appropriate informant was retained. The final sample contained 97 firm-level observations, one per firm.

The sample included firms of different ages, sizes, ownership forms, and digital organizational arrangements (

Table 4. Characteristics of the survey sample (N = 97).

Characteristic	Category	Frequency	Percentage
Firm age	Less than 3 years	6	6.2%
	3–5 years	10	10.3%
	6–10 years	14	14.4%
	11–20 years	27	27.8%
	More than 20 years	40	41.2%
Firm size	Fewer than 10 employees	1	1.0%
	10–99 employees	24	24.7%
	100–299 employees	16	16.5%
	300 employees or more	56	57.7%
Ownership	Private firm	37	38.1%
	State-owned or state-controlled	42	43.3%
	Foreign/Hong Kong/Macao/Taiwan invested	5	5.2%
	Public institution, association, or other	13	13.4%
Digital organization	Dedicated IT or digitalization department	71	73.2%
	Dedicated IT staff without a separate department	17	17.5%
	IT services partially or fully outsourced	6	6.2%
	Other arrangement	3	3.1%

). Firms operating for at least 11 years represented 69.0% of the sample, and firms with at least 300 employees represented 57.7%. State-owned or state-controlled firms accounted for 43.3%, private firms for 38.1%, foreign- or Hong Kong/Macao/Taiwan-invested firms for 5.2%, and other organizational forms for 13.4%. Most firms had an internal digital foundation: 73.2% had a dedicated IT or digitalization department, and 17.5% employed dedicated IT staff without a separate department.

3.2.2. Measures

The questionnaire contained firm characteristics and six focal constructs: digital technology application, digital strategic planning, firm social capital, the digital policy environment, competitive intensity, and digital transformation level. The grounded theory phase identified the relevant construct domains; the quantitative phase selected established scales that matched these domains and adapted wording to the Chinese HRS context.

For English-language scales, translation and back-translation were performed. Two graduate researchers with management backgrounds independently translated the measures into Chinese, and a third researcher who had not seen the original items back-translated them. The research team compared the versions and revised discrepancies using feedback from the pilot study. All construct items used five-point Likert scales ranging from 1 (strongly disagree) to 5 (strongly agree).

Digital technology application was adapted from the digital embeddedness scale developed by Liu et al. [35]. It assessed the extent to which big data, AI, and digital systems were embedded in core services and operations, including person–job matching, transaction processing, client service, and managerial decision making.

Digital strategic planning was adapted from the digital strategy scale of Proksch et al. [36]. It assessed whether digitalization was integrated into long-term corporate strategy and whether digital projects received priority in strategic decisions and resource allocation. The items covered attention to digital trends, the strategic importance of digitalization, project prioritization, strategic adjustment, and the aspiration to lead digitally.

Firm social capital was adapted from the organizational networking scale of Park and Luo [37]. It assessed the relational resources available to the firm through its external ties with business partners, industry organizations, and government agencies, including access to information, resources, and collaborative support. Because these resources are embedded in relationships beyond the focal firm’s boundary, firm social capital is treated as an environmental-context condition in the TOE model.

The digital policy environment was measured by adapting the perceived government support scale used by Zhao and Liu [38] to the digital policy context. The items covered policy communication and interpretation, diversity and accessibility of support, ease of application, and implementation fairness.

Competitive intensity was adapted from Jaworski and Kohli [39]. It captured the intensity of price competition, shifts in competitors’ behavior, speed of competitive response, and overall market pressure.

Digital transformation level was adapted from related measures used by Chi et al. and Lu et al. [40,41]. It assessed substantive changes in processes, service models, and value creation enabled by digital technologies.

Although digital technology application and digital transformation are related, they capture distinct constructs. Digital technology application reflects the deployment and embedding of digital technologies in existing operations, whereas digital transformation captures broader changes in business processes and value creation enabled by those technologies. Accordingly, the former is treated as a technological antecedent and the latter as the outcome condition.

3.2.3. Measurement Quality and Descriptive Statistics

SPSS 25.0 and Mplus 9.0 were used to assess reliability and validity. Cronbach’s alpha ranged from 0.859 to 0.916, and composite reliability ranged from 0.861 to 0.916, exceeding conventional standards for internal consistency [42,43,44]. KMO values ranged from 0.741 to 0.883, and all Bartlett tests of sphericity were significant (p < 0.001), supporting factor analysis [45]. Standardized loadings ranged from 0.613 to 0.929, and average variance extracted ranged from 0.608 to 0.773, indicating convergent validity, as reported in

Table 5. Reliability and validity statistics.

Construct	Cronbach’s α	KMO	Factor Loading	AVE	CR
Digital technology application	0.916	0.839	0.828–0.912	0.732	0.916
Digital strategic planning	0.912	0.851	0.792–0.894	0.723	0.912
Firm social capital	0.860	0.810	0.613–0.884	0.617	0.863
Digital policy environment	0.900	0.883	0.785–0.833	0.644	0.901
Competitive intensity	0.859	0.788	0.720–0.837	0.608	0.861
Digital transformation level	0.909	0.741	0.844–0.929	0.773	0.911

. The six-factor model showed good fit (χ² = 280.420, df = 237, CFI = 0.975, TLI = 0.970, RMSEA = 0.043, SRMR = 0.059) and outperformed a one-factor model. All constructs met the Fornell–Larcker criterion, and HTMT values ranged from 0.353 to 0.840, supporting discriminant validity [44,46]. Harman’s single-factor test showed that the first unrotated factor accounted for 37.09% of the variance, providing no indication that a single common method factor dominated the covariance structure. To further assess common-method bias, an unmeasured latent method construct (ULMC) analysis was conducted by introducing a common method factor into the measurement model. The baseline model without the common method factor showed good fit (χ²/df = 1.191, RMSEA = 0.044, CFI = 0.973, IFI = 0.974, TLI = 0.968). Adding the common method factor did not significantly improve model fit (Δχ² = 31.847, Δdf = 24, p = 0.131), indicating that common-method bias was not a significant concern in this study.

Descriptive statistics are reported in

Table 6. Descriptive statistics.

Construct	Minimum	Maximum	Mean	Standard Deviation
Digital technology application	1.250	5.000	3.294	1.098
Digital strategic planning	1.250	5.000	3.384	1.070
Firm social capital	1.250	5.000	3.528	0.823
Digital policy environment	1.800	5.000	3.577	0.809
Competitive intensity	1.250	5.000	3.322	0.968
Digital transformation level	1.333	5.000	3.354	1.100

. Means ranged from 3.294 to 3.577. The digital policy environment had the highest mean (M = 3.577), followed by firm social capital (M = 3.528), whereas digital technology application had the lowest mean (M = 3.294). Standard deviations ranged from 0.809 to 1.100, with the greatest dispersion in digital transformation level and digital technology application.

3.2.4. Fuzzy-Set Calibration

Calibration converted the original scores into fuzzy-set membership scores between 0 and 1 [47]. Because the conditions and outcome were measured as multi-item perceptual constructs, and no established HRS-industry thresholds correspond directly to full membership, the crossover point, or full non-membership in these sets, the direct calibration method was applied using empirical distributional anchors. The 95th, 50th, and 5th percentiles were specified as the anchors for full membership, the crossover point, and full non-membership, respectively [48]. Cases located exactly at 0.50 were adjusted slightly to 0.501 to avoid maximum ambiguity.

Table 7. Calibration anchors for the fuzzy sets.

Type	Set	Full Membership	Crossover Point	Full Non-Membership
Antecedent	Digital technology application	4.750	3.500	1.450
	Digital strategic planning	5.000	3.500	1.500
	Firm social capital	4.750	3.500	2.000
	Digital policy environment	5.000	3.600	2.160
	Competitive intensity	4.800	3.250	1.500
Outcome	Digital transformation level	5.000	3.333	1.333

presents the calibration anchors. The calibrated fuzzy-set membership scores for all cases are provided in Table S1 in Supplementary Materials.

3.2.5. Necessity, Sufficiency, and Robustness Procedures

The fsQCA analysis proceeded from necessity to sufficiency [49]. Each antecedent and its negation were tested as necessary conditions for high digital transformation and for the absence of high digital transformation. A consistency score of 0.90 was used as the criterion for necessity.

For the sufficiency analysis, truth tables were constructed using fsQCA 4.1 and minimized using Boolean algebra. The case-frequency threshold was set at 2. A raw consistency of 0.80 was used as the minimum benchmark and PRI consistency was set at 0.75 [50,51]. Inspection of the natural breaks in the empirical truth-table distributions produced final raw consistency cutoffs of 0.939 for high digital transformation and 0.925 for non-high digital transformation. Truth-table rows meeting the frequency, consistency, and PRI criteria were coded as producing the outcome.

fsQCA generates complex, intermediate, and parsimonious solutions. The substantive interpretation in this study is based on the intermediate solution. The parsimonious solution is used as a comparative reference for distinguishing core from peripheral conditions rather than as the primary basis for interpreting configurational pathways [47]. Conditions appearing in both the intermediate and parsimonious solutions are classified as core conditions, whereas those appearing only in the intermediate solution are classified as peripheral conditions. The complex solution, which does not incorporate logical remainders, is retained as a conservative reference [48]. Complete truth tables and the three solution forms are reported in Tables S2–S5 in Supplementary Materials to make the role of logical remainders and counterfactual assumptions transparent. A blank cell means that the condition is unrestricted, not absent.

To assess the robustness of the configurational findings, three sensitivity analyses were conducted. First, alternative calibration anchors were applied by replacing the original 95th/50th/5th percentile anchors with 90th/50th/10th percentile anchors. Second, the case-frequency threshold was increased from 2 to 3 to examine whether the configurations remained stable under a stricter empirical frequency requirement. Third, the PRI consistency threshold was increased from 0.75 to 0.80 to assess whether the identified configurations were sensitive to a stricter consistency criterion. The detailed results of these sensitivity analyses are reported in Tables S6–S11 in Supplementary Materials [52].

4. Results

4.1. Necessary Condition Analysis

Table 8. Necessary condition analysis.

Condition	High Digital Transformation: Consistency	High Digital Transformation: Coverage	Non-High Digital Transformation: Consistency	Non-High Digital Transformation: Coverage
Digital technology application	0.815	0.869	0.492	0.458
~Digital technology application	0.491	0.525	0.859	0.803
Digital strategic planning	0.846	0.883	0.488	0.445
~Digital strategic planning	0.468	0.511	0.871	0.832
Firm social capital	0.767	0.761	0.644	0.558
~Firm social capital	0.554	0.641	0.723	0.731
Digital policy environment	0.733	0.786	0.591	0.554
~Digital policy environment	0.584	0.620	0.772	0.717
Competitive intensity	0.793	0.792	0.609	0.531
~Competitive intensity	0.531	0.608	0.761	0.763

Note: The symbol “~” denotes set negation.

reports necessity consistency and coverage. No antecedent condition or its negation reached the 0.90 consistency threshold for either high digital transformation or non-high digital transformation. To assess the sensitivity of the necessity findings to calibration, we repeated the analysis using a stricter 97.5th/50th/2.5th percentile calibration. No condition or its negation exceeded the 0.90 consistency threshold, indicating that the conclusion regarding the absence of individually necessary conditions remained unchanged. Thus, no single condition constituted a necessary condition for either outcome. The result supports a conjunctural explanation in which transformation depends on combinations of technology, strategy, relational resources, and environmental conditions rather than on any isolated antecedent.

4.2. Sufficiency Analysis

The sufficiency analysis yielded five solution terms for high digital transformation and two for non-high digital transformation (

Table 9. Configurations associated with high and non-high digital transformation.

Antecedent Condition	High DT S1	High DT S2	High DT S3	High DT S4	High DT S5	Non-High DT N1	Non-High DT N2
Digital technology application	●	●	●	●		⊗	⊗
Digital strategic planning	●	●			●	⊗	⊗
Firm social capital	•		⊗	•	•	⊗	•
Digital policy environment			⊗	•	•	⊗
Competitive intensity		●	●	●	●		⊗
Consistency	0.917	0.960	0.967	0.962	0.957	0.923	0.938
Raw coverage	0.604	0.646	0.390	0.498	0.542	0.610	0.439
Unique coverage	0.069	0.037	0.012	<0.001	0.056	0.276	0.106
Solution consistency	0.913					0.922
Solution coverage	0.795					0.716

Note: Table reports the intermediate solutions. Large circles indicate core conditions appearing in both the intermediate and parsimonious solutions; small circles indicate peripheral conditions appearing only in the intermediate solution. Filled circles indicate presence, crossed circles indicate absence, and blank cells indicate that the condition is unrestricted. DT = digital transformation. S4 is retained to report the complete intermediate solution. Given its unique coverage below 0.001, it is interpreted as an overlapping variant within the technology–competition pattern rather than as an empirically independent pathway.

). For the high outcome, configuration consistency ranged from 0.917 to 0.967. Solution consistency was 0.913 and solution coverage was 0.795, indicating a strong subset relationship between the overall solution and high digital transformation.

The five high-outcome solution terms can be organized into three broader configurational patterns according to their shared core structures. S1 and S2 constitute the technology–strategy pattern, as both contain core digital technology application and digital strategic planning. S3 and S4 constitute the technology–competition pattern, as both contain core digital technology application and competitive intensity. S5 represents the strategy–ecosystem pattern, combining core digital strategic planning and competitive intensity with peripheral firm social capital and the digital policy environment.

In terms of raw coverage, S2 had the highest value (0.646), followed by S1 (0.604), S5 (0.542), S4 (0.498), and S3 (0.390). Although S4 showed relatively high raw coverage, its unique coverage was below 0.001, indicating that almost all of its empirical coverage overlapped with other configurations. S4 is therefore retained in Table 9 as part of the complete intermediate solution but is interpreted as an overlapping variant within the broader technology–competition pattern rather than as an empirically independent pathway.

For non-high digital transformation, solution consistency was 0.922 and solution coverage was 0.716. N1 combined the core absence of digital technology application, digital strategic planning, and firm social capital with the peripheral absence of a favorable digital policy environment; competitive intensity was unrestricted. N2 combined the core absence of digital technology application, digital strategic planning, and competitive intensity with peripheral firm social capital; the digital policy environment was unrestricted. The two non-high pathways shared the core absence of digital technology application and digital strategic planning but differed in their external resource and competitive conditions.

The high and non-high solutions were not mirror images. Conditions that were central in one high-outcome pathway could be unrestricted in another, while the absence of high transformation consistently involved a joint deficit in technology application and strategic planning. This pattern is consistent with the principles of equifinality and causal asymmetry.

4.3. Robustness Check

Three sensitivity analyses were conducted by varying the calibration anchors, case-frequency threshold, and PRI consistency cutoff. Recalibrating all conditions and the outcome using the 90th, 50th, and 10th percentiles yielded configurational results substantively consistent with the baseline analysis. Increasing the case-frequency threshold from 2 to 3 produced only minor changes in coverage and consistency, while the principal configurational patterns remained unchanged. Raising the PRI cutoff from 0.75 to 0.80 also preserved the principal solution structure, although S4 was no longer retained. Because S4 had negligible unique coverage in the baseline analysis (<0.001) and was already interpreted as an overlapping variant within the technology–competition pattern, its disappearance does not alter the substantive interpretation of the findings. Overall, the results are robust across these alternative analytical specifications.

5. Discussion

To further interpret the mechanisms underlying the identified configurations, we draw on illustrative evidence from the qualitative phase. These interview cases were independent from the 97 firms included in the fsQCA analysis and are used to provide contextual insights into how the identified conditions may operate in practice. Accordingly, the cases presented below should be understood as qualitative illustrations that complement the configurational findings rather than as empirical cases representing specific fsQCA pathways.

5.1. Configurational Pathways to High Digital Transformation

The five high-outcome solution terms are interpreted through three broader substantive patterns: technology–strategy synergy, technology–competition-driven transformation, and strategy–ecosystem synergy. S1 and S2 constitute the first pattern, S3 and S4 jointly constitute the second, and S5 constitutes the third. Because S4 contributes virtually no unique empirical coverage, it is not interpreted as a separate substantive pathway.

5.1.1. Technology–Strategy Synergy

S1 and S2 share the core presence of digital technology application and digital strategic planning. Their common structure indicates that technology is most consequential when it is directed by a strategic logic that specifies investment priorities, implementation sequencing, and integration with service scenarios [53]. In HRS firms, digital transformation is not achieved by adding information systems in isolation. It requires the embedding of technology in recruitment, talent assessment, training, payroll and tax services, and flexible staffing, thereby reconfiguring service delivery and value creation.

S1 adds firm social capital as a peripheral condition. It represents an internally accumulated transformation process supplemented by relational resources. Professional knowledge, client relationships, and business data provide application scenarios, while social capital supplies client insight, collaborative opportunities, and industry knowledge. The interview case of C05 provides a qualitative illustration of this mechanism: its accumulated assessment expertise and talent data were integrated with AI-enabled assessment, talent databases, and online consulting platforms, converting professional capability into digital products and platform services.

S2 adds competitive intensity as a core condition. Strong competition increases the urgency of transformation, while a clear digital strategy converts that pressure into sustained R&D, scenario innovation, and service-model adjustment. Firm C08, operating in online recruitment, incorporated digitalization into long-term strategy and continuously applied AI, large language models, intelligent recommendation, and talent profiling to job matching and platform services. The case illustrates how strategy can translate competitive pressure into sustained technology development rather than fragmented short-term responses.

5.1.2. Technology–Competition-Driven Transformation

S3 and S4 share the core presence of digital technology application and competitive intensity and are therefore interpreted jointly as a technology–competition-driven pattern. In information-intensive services, competitive pressure arising from service homogenization, declining margins, platform entrants, and changing client requirements increases the demand for faster, more efficient, and differentiated service delivery. Digital technology provides the operational basis for responding to these pressures through workflow automation, data-enabled decision making, and service innovation [54].

The two solution terms differ in their peripheral relational and policy conditions. S3 combines the peripheral absence of firm social capital and a favorable digital policy environment, whereas S4 combines their peripheral presence. However, because S4 has a unique coverage below 0.001, this peripheral distinction is not interpreted as constituting a separate substantive pathway. The theoretical interpretation therefore focuses on the shared technology–competition core of S3 and S4.

Firm C10 provides a qualitative illustration of this broader mechanism. The firm responded to intensifying competition in outsourcing services by developing business management systems, a software-as-a-service platform, and automation tools that redesigned client management, employee services, and transaction processing. This example illustrates how technological capability can be mobilized in response to competitive pressure even when relational and policy support is limited.

5.1.3. Strategy–Ecosystem Synergy

S5 combines core digital strategic planning and competitive intensity with peripheral firm social capital and the digital policy environment. Digital technology application is unrestricted. This configuration highlights the role of strategy in integrating market demand with external ecosystem resources. Under sustained competition, strategy clarifies priority service domains and coordinates policy resources, client networks, and partners within a broader process of business reconfiguration. Social capital supports interorganizational cooperation and resource exchange, while the policy environment supplies institutional support, projects, and public resources. Their integration can support movement from isolated digital projects toward platform-based and ecosystem-oriented services [55,56].

The unrestricted status of digital technology does not imply that technology is absent or unimportant. It means that membership in the set of high digital technology application is not a defining requirement for this pathway. For diversified HRS firms with extensive client networks and strong resource-integration capabilities, transformation may be expressed primarily through the redesign of business portfolios, organizational boundaries, and service ecosystems, with technologies deployed as instruments of strategic integration.

Firm C01 reflects this mechanism. The firm integrated recruitment, talent management, payroll and tax services, and employee services into a digital platform covering the HRS lifecycle. The key mechanism was not any isolated technology but the strategic coordination of business units, market requirements, government relationships, clients, and partners. This interpretation is consistent with research showing that platform innovation depends on the orchestration of heterogeneous resources rather than on technology possession alone [25].

5.2. Configurational Pathways to Non-High Digital Transformation

Both non-high configurations share the absence of digital technology application and digital strategic planning, but differ in the constraints they represent. N1 reflects weak internal digital foundations combined with limited external resource access, whereas N2 reflects transformation inertia in which relational resources sustain existing operations without supporting digital renewal.

5.2.1. Capability Deficit and Limited Ecosystem Embeddedness

N1 combines the absence of digital technology application, digital strategic planning, firm social capital, and a favorable digital policy environment, while competitive intensity is unrestricted. This configuration reflects more than a shortage of individual resources. It indicates a cumulative capability constraint in which weak internal foundations and limited external support coexist.

Internally, insufficient technological application limits firms’ experience in integrating digital tools into core services, while the absence of strategic direction weakens the coordination of digital investment, capability development, and business redesign. These two conditions reinforce a fragmented pattern of digitalization in which isolated technological initiatives are difficult to translate into broader organizational change. Externally, weak social capital and an unfavorable policy environment further restrict access to knowledge, partners, policy resources, and other forms of support that could compensate for internal capability gaps. The problem is therefore not simply that several favorable conditions are absent, but that firms lack both the internal basis for capability accumulation and the external channels through which those capabilities might be supplemented or accelerated [57].

This combination can keep digital development at a relatively low level. Without sustained technological learning, strategic coordination, or external resource mobilisation, firms have limited scope to move from incremental digital applications toward more systematic transformation. C16 illustrates this pattern. The firm remained dependent on traditional processes and established clients, lacked a systematic digital strategy and core digital platform, and had limited engagement with policy projects, technology providers, and industry networks. Its digital development consequently remained fragmented and incremental.

5.2.2. Relational Buffering and Transformation Inertia

N2 combines the absence of digital technology application, digital strategic planning, and competitive intensity with the peripheral presence of firm social capital; the digital policy environment is unrestricted. Unlike N1, this configuration is not characterized by a general shortage of resources. Instead, it reflects a situation in which existing relational resources support organizational continuity but are not redirected toward digital renewal.

Firm social capital can provide stable clients, business opportunities, institutional connections, and access to external resources. When competitive pressure is weak, however, these relationships can also buffer firms from the immediate need to reconsider established service models. The absence of digital strategic planning further limits the organizational direction required to redeploy relational resources toward technology development, process redesign, or new digital services. As a result, resources that are valuable for sustaining current operations may remain tied to existing business routines rather than being recombined for digital capability building.

This configuration therefore reflects a deeper problem of path dependence and resource reconfiguration. Established relationships make the existing business model viable, while weak competitive pressure reduces the urgency of departing from familiar routines. In the absence of a strategic mechanism for redirecting resources, firms may continue to exploit established relational advantages rather than use them to support digital renewal. Over time, this pattern can reinforce transformation inertia: the firm is not constrained by resource scarcity, but by the persistence of a resource-allocation logic that favors continuity over reconfiguration [58,59].

C06 illustrates this pattern. Its regional clients and cooperation networks provided a stable basis for ongoing services, while its digital systems remained focused on information aggregation and basic service functions rather than being integrated into a broader digital transformation strategy.

5.3. Theoretical Implications

This study offers three theoretical implications. First, it extends the TOE perspective by showing that technological, organizational, and environmental conditions are better understood configurationally than as independent and additive factors associated with digital transformation. Prior research has established the relevance of technological capability, strategic orientation, and environmental support, whereas our findings further show that their importance depends on the configurations in which they are embedded. Digital transformation in HRS firms is therefore better understood as the alignment of complementary conditions rather than the accumulation of isolated advantages.

Second, the findings qualify technology-centered explanations of digital transformation. Consistent with prior research, digital technology application appears in most configurations associated with high transformation, confirming its important enabling role. However, it does not constitute a necessary condition in the set-theoretic sense, and high transformation can also occur when high digital technology application is not a defining requirement. Thus, the findings do not diminish the importance of technology; rather, they show that its role in high-transformation configurations varies with its combination with strategic direction, competitive pressure, and complementary resources.

Third, the study extends relational-resource research by revealing the configuration dependent role of firm social capital in HRS firms. External relationships can provide information, legitimacy, cooperation opportunities, and access to resources, but social capital does not uniformly support digital transformation. It complements high transformation in some configurations, yet can coexist with non-high transformation when technological and strategic foundations are weak. This finding highlights the importance of resource orchestration and further demonstrates that the configurations associated with high and non-high digital transformation are asymmetric.

5.4. Managerial and Policy Implications

HRS firms should select transformation pathways that fit their resource endowments rather than imitate a single industry benchmark. Firms with strong technology and professional knowledge should strengthen digital strategic planning so that applications are integrated into core service processes. Firms facing intense competition can use market pressure as a trigger for scenario-specific innovation, but short-term technology purchases should be avoided when they are disconnected from a coherent strategy.

For firms with extensive relational and institutional resources, the priority is to convert those resources into platforms, data partnerships, and service innovation rather than use them only to preserve existing business. Conversely, firms with weak technology, strategy, and ecosystem links require coordinated capability building; isolated subsidies or relationship-based support are unlikely to be sufficient.

Policy makers and industry associations should recognize that policy support has a conditional rather than deterministic role. Public programs can be more effective when they combine financial or project support with capability assessment, technical partnerships, shared infrastructure, and opportunities for firms to test digital services in real operational settings.

6. Conclusions

6.1. Main Conclusions

Using grounded theory and fsQCA, this study examined the configurational patterns associated with digital transformation in Beijing HRS firms. The qualitative phase identified five relevant conditions: digital technology application, digital strategic planning, firm social capital, the digital policy environment, and competitive intensity. The quantitative phase yielded five high-outcome solution terms, which were interpreted through three broader configurational patterns, together with two configurations associated with non-high digital transformation.

Four conclusions emerge. First, no single necessary condition or universally optimal pathway exists. Firms with different capabilities and environments can reach the same outcome through differentiated configurations. Second, digital technology application, digital strategic planning, and competitive intensity form the principal axes of the high-outcome configurations, but the role of each condition depends on the configuration in which it is embedded. Third, firm social capital and the digital policy environment mainly play contextual enabling roles; their value depends on technological absorption, strategic integration, and market responsiveness. Fourth, the configurations associated with high and non-high digital transformation exhibit causal asymmetry. Non-high transformation is associated with either a systemic absence of internal and external conditions or the coexistence of relational resources with weak technological and strategic foundations.

Overall, digital transformation in HRS firms is a context-dependent and configurational process. Its central challenge is not simply to invest in technology but to align technology, strategy, relational resources, policy opportunities, and market response within a coherent service innovation system. Such alignment also provides an organizational foundation for more sustainable HRS provision through improved service efficiency, labor-market matching, and access to employment-related services.

6.2. Limitations and Future Research

First, the findings are bounded by the Beijing context. Beijing has relatively strong technological, policy, and professional resources, while the present sample also includes a relatively high proportion of larger and state-owned or state-controlled firms. These characteristics may shape firms’ access to digital capabilities, institutional support, and external networks. The identified configurations should therefore be interpreted primarily as patterns observed among the sampled Beijing HRS firms rather than as representative of all HRS firms in China. Future research could examine firms across different regions and organizational sizes, particularly SMEs outside major metropolitan areas, to assess the transferability and boundary conditions of these configurations.

Second, although the interview narratives provide process-oriented evidence showing that several identified conditions were already present or evolving during the initiation and development of firms’ digital transformation, the quantitative analysis remains cross-sectional and therefore cannot establish temporal ordering conclusively or rule out reciprocal relationships. Higher levels of digital transformation may, in turn, reinforce strategic planning, facilitate resource accumulation, or expand firms’ external relational networks. The configurational results should therefore be interpreted primarily as contemporaneous set-theoretic sufficiency relationships rather than as evidence of unidirectional temporal causation. Future research using longitudinal case studies, panel data, or dynamic QCA could examine temporal sequencing, reciprocal reinforcement, and the evolution of configurations over time.

Third, the antecedents are concentrated at the firm and environmental levels. Managerial digital cognition, leadership, absorptive capacity, organizational inertia, employee digital skills, and change acceptance may shape how firms convert technology, policy, and relational resources into transformation outcomes. Future research could directly measure these mechanisms and develop multilevel configurational models.

Fourth, the outcome is digital transformation level rather than its downstream consequences. Different high-outcome pathways may generate different levels of operational efficiency, innovation, service quality, resilience, or business-model renewal. Future work could examine an antecedent configuration–transformation pathway–performance consequence framework and compare the durability and risks of alternative pathways.