IT Implementation Research: A Technological Diffusion Approach – Cooper & Zmud (1990)

Randolph B. Cooper and Robert W. Zmud’s “Information Technology Implementation Research: A Technological Diffusion Approach,” published in Management Science (36(2), 1990), provided a landmark synthesis of fragmented IT implementation research streams into a unified framework organized around six stages of implementation and the critical role of task-technology compatibility. The model established that IT implementation success depends not on the technical quality of the system alone, but on the fit between technology capabilities and the task characteristics of the adopting organization.

Cooper and Zmud addressed a critical gap: organizations faced widespread IT implementation difficulties despite considerable investment, yet there was little integration of implementation research and no unified framework to guide future research or organizational practice. Their work synthesized factors research, process research, and political research into a coherent model grounded in technological diffusion literature.

Why Was the Model Created?

Cooper and Zmud developed this model to address a critical gap in IT implementation research. Organizations faced significant pressure to make their operational, tactical, and strategic processes more efficient and effective through information technology adoption. However, despite considerable investment in IT systems, organizations experienced widespread implementation difficulties. The researchers identified that while substantial research existed on IT implementation problems, there was little integration of implementation research streams and no unified directing and organizing framework to guide future research.

The model emerged from recognition that IT implementation success depends on managing multiple complex factors operating across different stages of the implementation process. Prior IT implementation research had been fragmented across several perspectives—factors research (examining individual, organizational, and technological factors), process research (examining social change activities), and political research (examining stakeholder interests and power dynamics). The Cooper-Zmud model synthesized these perspectives into a unified framework organized around implementation stages.

A key motivation was to examine Material Requirements Planning (MRP) implementation specifically, as this was a critical application in manufacturing firms. MRP adoption provided a real-world context for testing the model’s propositions about task-technology compatibility and the interaction between organizational and technological characteristics across implementation stages.

The researchers drew on technological diffusion literature and innovation adoption research to propose that IT implementation should be conceptualized as an organizational effort directed toward diffusing appropriate information technology within a user community. This perspective recognized that implementation success was not simply a matter of rational decision-making, but involved complex organizational dynamics, learning, political maneuvering, and the actual fit between technology capabilities and work requirements.

Several preceding models and theories shaped the Cooper-Zmud framework:

• Lewin’s (1952) Change Model: The unfreezing-change-refreezing framework provided foundational thinking about organizational change processes that Cooper and Zmud applied to IT implementation.
• Laudon’s (1985) Environmental and Institutional Models: Contributed understanding of how external environment and institutional factors shape system development and adoption.
• Technological Diffusion Literature (Rogers):Rogers’ work on how innovations spread provided the diffusion metaphor central to the model’s framing of IT implementation as an organizational diffusion process.
• Kwon & Zmud (1987): Earlier work on innovation adoption theory directly informed the six-stage framework.
• Tait and Vessey (1988): Work on IT implementation effectiveness provided empirical grounding for the task-technology compatibility emphasis.

Core Concepts and Definitions

The Cooper-Zmud model rests on two foundational concepts. First, IT implementation is a staged process—organizations do not move from no-adoption to full adoption in a single step but progress through distinct phases with different challenges and success factors. Second, task-technology compatibility is central to implementation success—the fit between the capabilities of the technology and the characteristics of the work to be done determines whether adoption is likely and whether it succeeds.

The model identifies six distinct stages of IT implementation, each characterized by different processes, products, and outcomes:

1. Initiation: Active or passive scanning of organizational problems/opportunities and IT solutions; matching IT solutions to organizational needs.
2. Adoption: Rational and political negotiations to secure organizational backing; deciding to invest resources for implementation.
3. Adaptation: Developing and installing the IT application; revising organizational procedures; training organizational members in new procedures and IT applications.
4. Acceptance: Inducing organizational members to commit to IT application usage; achieving organizational embedding of the IT application.
5. Routinization: Encouraging use of IT application as a normal activity; adjusting organizational governance systems.
6. Infusion: Obtaining increased organizational effectiveness through IT application use; using IT within the organization to its fullest potential.

Task characteristics relevant to the specific IT being implemented include manufacturing method, demand orientation, complexity, and process interdependence. Technology characteristics include system design quality, user-designer interaction quality, and IT design fit with work requirements. Organizational and user characteristics include user job tenure and education, organizational structure (specialization, centralization, formalization), and resistance to change.

Internal Validity

The Cooper-Zmud model’s internal validity was tested through an empirical cross-sectional field survey of Material Requirements Planning (MRP) implementations. The researchers conducted a survey of American Production and Inventory Control Society (APICS) members across manufacturing firms in the United States, using telephone interviews to reduce terminology confusion and encourage high response rates. The initial random sample contained 100 APICS members, with 52 manufacturing facilities ultimately included in the analysis (a 97% response rate among contacted applicable members).

The study employed logistic regression analysis with dichotomous dependent variables (adoption versus non-adoption, and level of infusion classification). Logistic regression was selected because it provided more efficient and flexible analysis than standard linear regression. The study tested four major research hypotheses relating task-technology compatibility and complexity to likelihood of MRP adoption.

The results showed strong support for the task-technology compatibility premise. The overall logistic regression model for MRP adoption was significant (p < 0.02), with three of four hypotheses supported (p ≤ 0.05). Manufacturing method emerged as the strongest predictor: continuous, repetitive manufacturing methods were significantly more likely to adopt MRP than intermittent job shop methods. Bill-of-material levels and parts complexity also significantly predicted adoption. The research demonstrated that MRP adoption correlates with manufacturing environments characterized by continuous production methods, higher product complexity, and deterministic demand characteristics.

The model measures implementation success and adoption through multiple dimensions. Primary dependent variables include adoption versus non-adoption (a dichotomous measure indicating whether an organization has decided to implement MRP) and level of MRP infusion (classified into Class A through D implementation levels). Class A represents a closed-loop system with priority and capacity planning used by top management; Class B has capability for priority and capacity planning with somewhat inflated master production schedule; Class C is an order launching system with priority only; Class D is MRP existing mainly in data processing while the informal system runs the company.

External Validity

The Cooper-Zmud model’s external validity was established through several mechanisms:

• Sample Composition and Representativeness: The research employed a random sample survey designed to reduce confusion and encourage high response rates. The MRP adoption classification of the survey sample was similar to industry adoption rates found in previous research.
• Industry Profile Validation:The study documented the survey’s industry and MRP classification profile, comparing it directly to the Anderson et al. (1981) survey sample. The researchers found “a strong similarity between the Anderson sample and this study’s sample of firms” in terms of industry distribution and MRP adoption patterns.
• Geographic Representativeness: The study included respondents across 21 different U.S. states, with significant representation from states with manufacturing concentrations (California 17.3%, New York 15.5%, Michigan 7.7%, Massachusetts 5.9%, and others).
• Theoretical Generalization: The model built upon established theoretical frameworks from technology diffusion literature and organizational innovation research, suggesting that findings about task-technology interaction should apply beyond MRP systems to other IT implementations where task characteristics can be similarly analyzed.

The researchers explicitly acknowledged external validity limitations: all firms were from APICS (potentially biasing toward MRP-aware organizations); responses were collected at a single point in time, limiting ability to examine dynamic processes; limited control variables were included; and response bias concerns existed. Despite these limitations, the findings showed reasonable consistency with prior research.

Key Contributions

Comprehensive Integration of Research Streams:The model uniquely integrates three previously fragmented research approaches—factors research, process research, and political research—into a single coherent framework. Rather than treating these as separate domains, the model shows how individual factors, organizational processes, and political dynamics interact across implementation stages to influence outcomes.

Stage-Based Framework: The six-stage model provides clear structure and progression, allowing researchers and practitioners to understand implementation as an evolving process rather than a discrete event. This staging framework acknowledges that different factors may be relevant at different implementation phases, and that successful transition through early stages does not guarantee success in later stages.

Theoretical Grounding in Diffusion Literature: Building on established technological diffusion theories (Rogers) and organizational change models (Lewin), the model benefits from rich theoretical foundations. This grounding enhances credibility and suggests broader applicability beyond IT systems.

Focus on Task-Technology Compatibility: The model emphasizes that technology adoption success depends not merely on rational decision-making or technical superiority but on the fit between technology capabilities and task requirements. This challenged purely technical or change-management perspectives and recognized that organizational context profoundly influences implementation outcomes.

Empirical Testing and Validation: The researchers conducted systematic empirical testing through a cross-sectional survey with logistic regression analysis. The study tested specific hypotheses about task-technology relationships, providing empirical evidence rather than purely theoretical propositions.

Practical Multi-Factor Perspective: By incorporating user characteristics, organizational characteristics, task characteristics, and technology characteristics, the model acknowledges the complexity of IT implementation and provides a multidimensional lens that is more realistic than single-factor explanations.

Clear Identification of Research Gaps: The paper explicitly identifies that prior research concentrated heavily on factors and process research, with relatively less attention to political research and the interaction of multiple factors across implementation stages. The model addresses these gaps directly.

Relevance to Technology Adoption

The Cooper-Zmud model provides a framework for practitioners to understand and manage IT implementation across multiple organizational dimensions and implementation stages. Organizations can use the six stages to structure their implementation planning and to recognize where implementation initiatives currently stand. Each stage has specific activities and management focuses that require different resource allocation and leadership attention.

For technology adoption broadly, the model’s task-technology compatibility insight is highly transferable. Organizations should assess not just whether a technology is technically superior but whether it fits their specific work characteristics, process complexity, and organizational context. MRP adoption is most successful in environments with continuous manufacturing methods, make-to-stock strategies, high bill-of-material levels, and product complexity—environments where MRP’s underlying assumptions are met. Organizations with different task characteristics should recognize that higher complexity and more sophisticated solutions may be required, necessitating greater management attention to user education, system customization, and organizational change management.

The infusion stage of the model has particular contemporary relevance. Many organizations successfully adopt technology (reach Acceptance and Routinization) but fail to achieve Infusion—using IT to its fullest organizational potential. This gap between routine use and strategic value creation is a critical adoption challenge that the model specifically names and addresses.

Organizations can use the model to anticipate implementation challenges based on task-technology fit. When task characteristics violate a technology’s underlying assumptions, organizations face higher implementation difficulty and should be prepared with enhanced support for implementation activities, more extensive user involvement, and modified systems better suited to their context.

Several limitations should be acknowledged. The empirical study employed a cross-sectional survey design, preventing examination of implementation dynamics and temporal progression through stages. The study included only 52 manufacturing facilities from APICS members, limiting generalization to diverse industries and non-manufacturing sectors. Technological obsolescence is a concern: the model was developed in the context of MRP systems in 1990, and specific task-technology characteristics examined may not be equally relevant to contemporary IT systems (cloud computing, enterprise resource planning, artificial intelligence) that operate under different assumptions. Political factors receive limited empirical investigation despite being identified as important. The dependent variable classification into broad categories (Class A through D) may mask important variations.

Despite these limitations, the Cooper-Zmud model’s synthesis of implementation research streams, its six-stage framework, and its emphasis on task-technology compatibility have made it a foundational reference in IT implementation research. Its core insights about implementation as a staged, context-dependent diffusion process continue to inform both academic research and organizational practice in technology adoption.

Note: This article provides an overview based on the comprehensive literature review. Readers are encouraged to consult the original publication for complete details.