Matrix organization

People report to two bosses at once, usually a function head and a product or project head, so two dimensions are managed together.

What it is

A matrix organization puts the same people under two lines of authority at once. Typically an engineer belongs to a functional department, which holds the technical standard, and to a project, whose manager sets deadlines; in multinationals the lines are often product and area. A shared superior sits above both. Davis and Lawrence (1977) define it as a multiple command system, with dual budgeting, dual evaluation, and an associated culture.

The matrix puts the same specialist under two lines of authority

Select a unit or the letter to read it; the key above the drawing names the lines and the letter used.

Reports toAdjust to each other directlyMMutual adjustment2Two superiors

Select a unit or a letter to read it. M, mutual adjustment: People coordinate their own work by informal communication with each other.Shared superior. Layer: Strategic apex. Reports to no one. Sits above both lines, so a conflict the lines do not settle below goes to one person.Functional head. Layer: The two lines. Reports to Shared superior. By function. Holds the technical standard for the specialists in the department.Project manager. Layer: The two lines. Reports to Shared superior. By project. Sets the project's deadlines and has real say over the same people's work.Specialist in both lines. Layer: Two bosses. Reports to Functional head and Project manager. Belongs to a functional department and to a project at the same time, with dual budgeting and dual evaluation (Davis and Lawrence 1977).M, mutual adjustment. People coordinate their own work by informal communication with each other.

The structure in words

Strategic apex
- Shared superior, reports to no one. Sits above both lines, so a conflict the lines do not settle below goes to one person.
The two lines
- Functional head, reports to Shared superior, grouped by function. Holds the technical standard for the specialists in the department.
- Project manager, reports to Shared superior, grouped by project. Sets the project's deadlines and has real say over the same people's work.
Two bosses
- Specialist in both lines, reports to Functional head and Project manager. Belongs to a functional department and to a project at the same time, with dual budgeting and dual evaluation (Davis and Lawrence 1977).

Functional head and Project manager: adjust to each other directly (Adjust between the lines), coordinated by mutual adjustment.

The specialist belongs to a functional department and to a project at once, and a shared superior sits above both lines. The two lines coordinate by mutual adjustment, and a conflict between them is settled by the two-boss specialist or goes up to the shared superior.

Source: Davis and Lawrence 1977; Galbraith 1971.

Theory places it on three dimensions: grouping on two bases at once (Galbraith 1971), mutual adjustment as the prime coordinating mechanism, and the matrix as the top of Galbraith's (1974) lateral ladder. The rest are our reading: specialization spans medium to high, formalization is medium, centralization mixed, and layers moderate. Decision rights span limited vertical to selective vertical and horizontal decentralization (Mintzberg 1980, a synthesis offering propositions without a new test). Differentiation is high and integration heavy, though the most differentiated firms in Lawrence and Lorsch (1967), a multi-firm comparative field study, used an integrating department instead. Exchange is bureaucratic (Ouchi 1980), the boundary selective (Ashkenas and colleagues 1995).

Where the matrix organization sits on the eleven dimensions

The filled segments are the form's levels; select a row for its reason and sources.

Select a dimension to read why the form sits there, and on what basis.Basis of grouping: Two or more at once. Basis: Theory. Two bases are managed at once, usually function and product or project, so the same people report to two lines. Sources: Galbraith 1971; Davis and Lawrence 1977.Specialization: Medium to high. Basis: Judgment. The functional line pools specialists whom the product or project line draws on; judgment for the band. Sources: Galbraith 1971.Standardization and formalization: Medium. Basis: Judgment. Dual reporting needs defined roles and processes at the junctions, while much of the work is settled by negotiation; judgment. Sources: Davis and Lawrence 1977; Burton and colleagues 2015a.Centralization: Mixed. Basis: Judgment. Conflicts between the two lines are settled by the two-boss managers or by the shared superior above them; judgment. Sources: Davis and Lawrence 1977; Kuprenas 2003.Layers and span of control: Moderate. Basis: Judgment. A matrix adds a second line rather than more levels; judgment. Sources: Galbraith 1971.Decision rights and the locus of authority: Limited vertical decentralization to selective vertical and horizontal decentralization. Basis: Judgment. Product or region heads hold delegated power in parallel, as in Type C, and in project matrices decisions go to the teams where the lines meet, as in Type D; judgment. Sources: Mintzberg 1980; Galbraith 1971.Prime coordinating mechanism: Mutual adjustment. Basis: Theory. Mutual adjustment between the two lines, under a dual authority structure. Sources: Davis and Lawrence 1977; Galbraith 1971.Lateral mechanisms: Matrix. Basis: Theory. The matrix is the last step of Galbraith's ladder of lateral processes, added when integrating and linking roles are overloaded; in hospitals, adoption also followed sociometric location and the cumulative force of adoption in interorganizational networks, not task diversity alone. Sources: Galbraith 1974; Burns and Wholey 1993.Differentiation against integration: High differentiation, heavy integration. Basis: Judgment. The heaviest integrating device for units that differ sharply; judgment, since Lawrence and Lorsch's most differentiated firms used an integrating department rather than a matrix. Sources: Galbraith 1974; Lawrence and Lorsch 1967a.Mode of exchange and control: Bureaucracy. Basis: Judgment. Authority, budgets, and evaluation run through two hierarchies inside one firm; judgment. Sources: Ouchi 1980.Boundary permeability: Selective. Basis: Judgment. The horizontal boundary between function and product is opened by design, while the firm's edge stays firm; judgment. Sources: Ashkenas and colleagues 1995.

Three levels rest on theory: two bases of grouping at once, mutual adjustment as the prime mechanism, and the matrix at the top of Galbraith's lateral ladder. The other eight are judgment, and no row is contested.

Source: Galbraith 1971; Davis and Lawrence 1977; Burton and colleagues 2015a, a conceptual essay with no new data; Kuprenas 2003, one case study; Mintzberg 1980, a synthesis of earlier research, offering propositions without a new test; Galbraith 1974; Burns and Wholey 1993, a quantitative study of matrix adoption and abandonment in a group of hospitals; Lawrence and Lorsch 1967a, a multi-firm comparative field study; Ouchi 1980, a conceptual article; Ashkenas and colleagues 1995.

Where it came from

The matrix was practice first: in U.S. aerospace and defense in the 1950s and 1960s, development projects were laid over functional engineering departments, and neither technical depth nor delivery could be subordinated to the other. Galbraith (1971) traced it to aerospace and set out the designs between project and functional forms. Its multinational version, a product-by-area grid, was what Stopford and Wells (1972), a study of 187 U.S. multinational enterprises, expected to emerge, its nature unclear; Knight (1976) reviewed its spread; and Davis and Lawrence (1977) gave it a book.

The schools of organization design and the forms' origins, 1900 to now

Each row is a school's era; each square under them is a form at its first year. Select a row or a square to read it; on a form's page, its own square and span are marked, with the schools that treat it.

The forms by first year

Schools that treat this form: Structural contingency theory; Information processing; Institutional theory; and Computational organization design and search.

Select a school's row or a form's square to read it.Select a form's square to read it; open a school below.Bureaucracy and its critics1900s to 1950s, revived from the 1990sWhat it claimsWeber held that rational-legal authority exercised through an office hierarchy, written rules, trained officials, and promotion by qualification is technically superior in precision, speed, and reliability. Critics studied its unintended consequences, Michels's iron law sharpest.What it has shownMerton (1940) argued, with no test, that rules can turn from means into ends; case studies followed. The Aston group concluded that the type's dimensions vary independently, and Wagner's meta-analysis (2021) found that many of their interrelationships have weakened.Who disputes itHuman relations and System 4 sets the costs of impersonal rules to motivation and cooperation against the technical superiority Weber stresses.Bureaucracy and its critics in The Schools of Organization DesignClassical administration1910s to 1940s, and in German to the 1970sWhat it claimsThere are general principles of administration, among them division of work, unity of command, a scalar chain of authority, and a limited span of control, and a structure built to them will be efficient in most settings. Taylor added that planners should set work methods.What it has shownThe school offered reasoning and experience, not tests. Woodward (1965) found that the spans and levels of the more successful firms in her sample differed by production system, and Simon (1946) showed that the principles come in pairs that point in opposite directions.Who disputes itStructural contingency theory found that the structure of successful firms varies with technology and environment, where classical writers treat their principles as general.Classical administration in The Schools of Organization DesignHuman relations and System 41920s to 1960sWhat it claimsOutput depends on the social relations of work, the norms of informal groups, and the style of supervision as much as on formal design. Likert's System 4, overlapping work groups joined by linking-pin managers under participative management, outperforms authoritative systems.What it has shownThe founding evidence is the Hawthorne studies, which later reanalyses dispute (Carey 1967; Levitt and List 2011). Likert's support came largely from his own institute's surveys. Wagner (1994) found participation's effects on performance and satisfaction significant but small.Who disputes itLabor process and critical theory reads human relations as a technique of managerial control rather than a change in the distribution of power.Human relations and System 4 in The Schools of Organization DesignTransaction cost economics and the M-form hypothesis1937, then the 1970s to the presentWhat it claimsFirms exist, and are organized as they are, to economize on transaction costs. Hierarchy replaces the market when assets are specific, uncertainty is high, and exchange is frequent, and the multidivisional form should outperform the functional form in large, diversified firms.What it has shownArmour and Teece (1978) found, in United States petroleum firms from 1955 to 1973, a positive relationship between the multidivisional form and profitability while the form was spreading, but not once an equilibrium in organizational form was reached.Who disputes itThe network perspective on organizations: Granovetter argued that the theory undersocializes economic action, which is embedded in networks of personal relations that also produce trust and order.Transaction cost economics and the M-form hypothesis in The Schools of Organization DesignThe Carnegie school and the attention-based view1940s to the presentWhat it claimsPeople in organizations are boundedly rational, so structure matters because it sets the premises of their decisions and channels their attention: who sees which problems, which goals they pursue, and which routines they follow. Ocasio makes structure distribute attention.What it has shownMarch and Simon (1958) drew propositions from the existing literature, and Cyert and March (1963) added process models and simulations. Two studies show hierarchy shaping what is sent up as well as what is chosen (Reitzig and Maciejovsky 2015; Keum and See 2017).Who disputes itOrganizational economics: decision rights, authority, and complementarities models a principal and agents with defined objectives; the Carnegie school, firms as coalitions with unresolved goal conflict.The Carnegie school and the attention-based view in The Schools of Organization DesignSocio-technical systems1950s to 1980sWhat it claimsAn organization is a social system and a technical system that must be designed together. The same technology admits more than one way of organizing work, and groups that own a whole task and control its variances at the source perform better and are more satisfied.What it has shownTrist and Bamforth (1951) found that longwall mechanization broke up self-regulating work groups. A review of 134 experiments found them extremely successful overall, and Stewart (2006) linked team autonomy to performance; studies of socialist brigades and Volkswagen qualify it.Who disputes itLabor process and critical theory: Kelly argued that labor intensification, not joint optimization, accounts for what socio-technical redesign achieved, and that the part of pay incentives was seriously underestimated.Socio-technical systems in The Schools of Organization DesignThe economics of the labor-managed firm and of enterprise ownership1958 to the presentWhat it claimsA firm run by its workers behaves differently from one run for investors: Ward modeled the Yugoslav-style firm as maximizing income per worker. Hansmann argued that an ownership form's success depends on the balance between the costs of market contracting and of ownership.What it has shownBurdín (2014) found a 29 percent lower hazard of dissolution for Uruguayan worker-managed firms than for conventional ones, and Fakhfakh, Pérotin, and Gago (2012) found French labor-managed firms as productive or more so in every industry studied.Who disputes itOrganizational economics (Jensen and Meckling) contrasts the labor-managed firm, in which tradable residual claims are legally prohibited, with cooperatives and partnerships that arise out of voluntary contracting.The economics of the labor-managed firm and of enterprise ownership in The Schools of Organization DesignModularity and the mirroring hypothesis1960s to the presentWhat it claimsComplex systems are built of nearly decomposable parts, and an organization's structure tends to mirror the architecture of what it makes: modular products allow loosely coupled organizations, and tightly coupled products need tightly coupled teams. Conway's law runs in reverse.What it has shownColfer and Baldwin (2016) review the empirical studies and report both support and principled exceptions. Direct tests point both ways: MacCormack, Baldwin, and Rusnak (2012) found open source products more modular than their commercial pairs; dynamic components break the mirror.Who disputes itNo school in the record disputes it head on; it answers information processing, which meets the coordination load inside the organization; mirroring writers locate it in the product's architecture.Modularity and the mirroring hypothesis in The Schools of Organization DesignStrategy and structure1960s to 1980sWhat it claimsStructure follows strategy: firms that diversify by product or geography outgrow the centralized functional structure and, after a lag and usually a crisis, adopt divisions. Chandler drew this from the histories of DuPont, General Motors, Standard Oil of New Jersey, and Sears.What it has shownChandler (1962) is comparative business history of four firms set against a wider survey. Rumelt (1974) related diversification and structure to performance, and Egelhoff (1988), with 34 multinationals, found some of Stopford and Wells's relationships supported and others not.Who disputes itInstitutional theory: in Fligstein's model divisions spread with a product-related or unrelated strategy, presidents from sales or finance, and other firms in the industry changing; strategy was one of three.Strategy and structure in The Schools of Organization DesignStructural contingency theory1960s to the presentWhat it claimsThere is no one best structure. Performance depends on fit between structure and contingencies such as environment, technology, size, and strategy, on an open-system view of the organization. Drazin and Van de Ven separate fit as selection, interaction, and systems.What it has shownBurns and Stalker, Woodward, and the Aston group related structure to context; Drazin and Van de Ven (1985) supported selection and systems fit only. Pennings (1975) found that fit did not explain effectiveness; Schoonhoven (1981) found traditional hypotheses unsupported.Who disputes itStrategic choice: Child argued that contingency theory explains structure by the constraints of context and leaves out the political process by which power-holders choose strategic action, structural forms included.Structural contingency theory in The Schools of Organization DesignThe French bureaucratic school (strategic analysis)1960s to 1980sWhat it claimsA bureaucracy is a system of power relations organized around zones of uncertainty. Impersonal rules and centralization protect members from depending on one another, those who control uncertainties hold power, and, unable to correct its errors, it changes only through crisis.What it has shownLe phénomène bureaucratique (1963; English 1964) rests on two case studies, a clerical agency and a state industrial monopoly. L'acteur et le système (1977; English 1980) is theoretical.Who disputes itSocietal effect and comparative business systems: D'Iribarne argues that national logics of conduct, such as a French logic of honor, explain differences that Crozier attributed to power games.The French bureaucratic school (strategic analysis) in The Schools of Organization DesignThe Japanese-firm literature1960s to the presentWhat it claimsThe large Japanese firm coordinates operating decisions through horizontal information sharing among semi-autonomous shops and uses its ranking hierarchy mainly as an incentive scheme (Aoki). It creates knowledge through middle-up-down management (Nonaka and Takeuchi).What it has shownAoki (1986) is a formal model. Lincoln, Hanada, and McBride (1986) compared 51 Japanese and 55 United States manufacturing plants, and Nonaka and Takeuchi (1995) rest on case studies, chiefly of Japanese firms. Fruin (1994) found the enterprise system defined at three levels.Who disputes itLabor process and critical theory reads shop-floor delegation as control: Sewell and Wilkinson argue that just-in-time and total quality regimes delegate tactical responsibility while centralizing strategic control.The Japanese-firm literature in The Schools of Organization DesignThe Scandinavian and Dutch socio-technical schools1960s to 2000sWhat it claimsThe Scandinavian program sought democracy at work through field experiments with semi-autonomous groups, then through democratic dialogue. The Dutch school's design theory turns complex organizations with simple jobs into simple organizations with complex jobs.What it has shownEmery and Thorsrud reported field experiments in Norwegian industry; Gustavsen (1992) records that they spread less than hoped, his reason for turning to dialogue. Berggren studied work organization in the Swedish auto industry. The Dutch work rests on design cases and argument.Who disputes itThe Japanese-firm literature: defenders of the Toyota system hold that standardized work with employee participation learns faster; Berggren reports Swedish long-cycle plants matching a line plant on productivity.The Scandinavian and Dutch socio-technical schools in The Schools of Organization DesignConfiguration theory1970s to the presentWhat it claimsStructures cluster into a small number of internally coherent configurations whose elements reinforce one another, so organizations move between a few types in large jumps, not small steps. Mintzberg's five configurations each rest on one coordinating mechanism.What it has shownDoty, Glick, and Huber (1993) found that fit to Miles and Snow's ideal types predicted effectiveness while fit to Mintzberg's did not. A meta-analysis of 40 tests (Ketchen and colleagues 1997) found that configuration partly explains performance, more so with broad definitions.Who disputes itStructural contingency theory: Donaldson counters that most organizations sit at intermediate points on a continuum of fits, not in a few configurations, and that studies over time find mostly incremental change.Configuration theory in The Schools of Organization DesignInformation processing1970s to the presentWhat it claimsThe more uncertain the task, the more information must be processed during its execution. An organization either reduces that need, through slack resources or self-contained tasks, or raises its capacity, through vertical information systems or lateral relations up to the matrix.What it has shownGalbraith (1973, 1974) built the framework from cases and from Simon's theory of bounded rationality, and Tushman and Nadler (1978) set it out as an integrating concept for design. Daft and Lengel (1986) argued that structure must reduce equivocality as well as supply data.Who disputes itThe Carnegie school and the attention-based view: Ocasio holds that firm behavior results from how the firm channels and distributes attention, so structure governs where attention goes, not only how much is processed.Information processing in The Schools of Organization DesignLabor process and critical theory1970s to the presentWhat it claimsStructure is above all a system of control over labor. Braverman argued that management deskills work by separating conception from execution; Edwards traced a sequence from simple control by the boss to technical control through machines to bureaucratic control through rules.What it has shownHistorical and theoretical (Braverman 1974; Edwards 1979). Delbridge (1995), observing a Japanese-owned plant in England, found that just-in-time and total quality regimes combined the control of labor with management's economic goals more completely than piece rates had.Who disputes itBureaucracy and its critics: Adler argues that labor process theory ignores the progressive socialization of the labor process, which drives long-term skill upgrading, and disputes deskilling as the trend.Labor process and critical theory in The Schools of Organization DesignStrategic choice1970s to the presentWhat it claimsStructure is the outcome of choices by those with power in the organization, who decide on strategic action, including the structural form, the manipulation of environmental features, and the performance standards to be met. Context constrains but does not determine structure.What it has shownLargely conceptual: Child (1972) built the argument from gaps in contingency findings, and the empirical case rests mostly on the modest share of structural variance contingencies explain. A 2026 review of 42 studies finds upper echelons theory their main theoretical anchor.Who disputes itStructural contingency theory: Donaldson replies that organizations mostly adopt the structures their contingencies dictate, to avoid the loss of performance misfit brings, so choice over structure is at most limited.Strategic choice in The Schools of Organization DesignThe German codetermination literature1970s to the presentWhat it claimsWorks councils and employee board seats form a structural layer through which coordination can become negotiation between management and employees. Defenders argue that it supports long-term employment, training, and cooperation; critics, that it weakens shareholder control.What it has shownJäger, Schoefer, and Heining (2021) found board-level codetermination left wages unchanged and, if anything, raised capital formation. Jirjahn (2011) reports studies pointing to positive works council effects; a meta-analysis found codetermination laws tied to lower productivity.Who disputes itOrganizational economics (Gorton and Schmid) found that firms with equal employee and shareholder representation did worse for their shareholders than firms with one-third employee representation.The German codetermination literature in The Schools of Organization DesignOrganizational economics: decision rights, authority, and complementarities1976 to the presentWhat it claimsDesign is the joint choice of who holds which decision rights, how performance is measured, and how it is rewarded. Rights should be placed where the specific knowledge lies, formal authority differs from real authority, and practices are complements that pay off in clusters.What it has shownIchniowski, Shaw, and Prennushi (1997) found, on 36 steel finishing lines, that lines using a set of innovative work practices achieved substantially higher productivity, consistent with complementarities. Bloom and colleagues found more decentralization where trust was higher.Who disputes itGhoshal argued that management research has shaped managers' worldview with ideologically inspired amoral theories, built on a scientific model that excludes human intentionality and so denies moral considerations.Organizational economics: decision rights, authority, and complementarities in The Schools of Organization DesignInstitutional theory1977 to the presentWhat it claimsFormal structures are often adopted because they are legitimate in the organization's field, as rationalized myths, and then decoupled from the work done. Organizations in a field grow alike through coercive, mimetic, and normative isomorphism, whether or not it is efficient.What it has shownTolbert and Zucker (1983) found that city characteristics predicted early but not later adoption of civil service reform, which they read as institutionalization. Palmer, Jennings, and Zhou (1993) found support for institutional and economic accounts of late M-form adoption.Who disputes itStructural contingency theory: Kraatz and Zajac found little support for neoinstitutional predictions in 631 liberal arts colleges, which grew less alike; the change they studied was curricular, not structural.Institutional theory in The Schools of Organization DesignPopulation ecology1977 to the presentWhat it claimsOrganizations face strong structural inertia, because selection favors the reliability and accountability of stable structures. Change in a population's forms comes mainly from founding and failure, not adaptation, and changing core features resets the liability of newness.What it has shownAmburgey, Kelly, and Barnett (1993) found that the effects of change on the hazard of failure vary by type of change and over time, and that rates of change depend on how often and how recently an organization has changed. Carroll and Hannan (2000) set out the program's methods.Who disputes itFor structural contingency theory, Donaldson argues that ecology understates managers' capacity to adapt structure.Population ecology in The Schools of Organization DesignResource dependence1978 to the presentWhat it claimsOrganizations depend on outside parties for critical resources, and structure follows the management of that dependence: units that cope with critical dependencies gain power, and boundaries are redrawn through cooptation, joint ventures, and mergers to absorb constraint.What it has shownPfeffer and Salancik (1978) assembled studies of boards, mergers, and joint ventures. A meta-analysis of 157 tests (Drees and Heugens 2013) corroborated the main prediction: organizations meet resource dependencies with interorganizational arrangements that raise their autonomy.Who disputes itTransaction cost economics and the M-form hypothesis explains boundaries by efficiency; resource dependence, by power and constraint.Resource dependence in The Schools of Organization DesignThe business-groups literature1978 to the presentWhat it claimsGroups of legally independent firms bound by ownership and social ties are a durable form of their own, neither one firm nor a market. Where institutions function poorly, groups can benefit member firms and can destroy value; they are sometimes paragons, sometimes parasites.What it has shownKhanna and Palepu (2000) found that affiliates of the most diversified Indian groups outperformed unaffiliated firms. A meta-analysis of 141 studies (Carney and colleagues 2011) found that affiliation lowers performance in general, less so where institutions are underdeveloped.Who disputes itOrganizational economics: agency writers read pyramids and cross-holdings as devices that separate control from ownership; Almeida and Wolfenzon offer a financing rationale for pyramids instead.The business-groups literature in The Schools of Organization DesignScandinavian institutionalism1980s to the presentWhat it claimsIdeas of organizing do not diffuse unchanged; they travel and are translated as they are adopted, so a form taken up in one place is not the same form elsewhere. Organizations can also be built of organizations (meta-organizations), which behave differently.What it has shownA discussion paper by Czarniawska and Joerges (1995) contrasts the diffusion of ideas with their translation, drawing on examples such as Swedish municipal reform; the work on meta-organizations (Ahrne and Brunsson 2008) is conceptual.Who disputes itInstitutional theory can read the same processes, seen from a translocal view, as institutional isomorphism, though translation holds that each adopter modifies an idea, so an institution re-created elsewhere is new.Scandinavian institutionalism in The Schools of Organization DesignSocietal effect and comparative business systems1980s to the presentWhat it claimsFirms matched on size, technology, and industry differ systematically by country, because how a society educates, trains, recruits, and promotes its workforce, and its industrial relations, shape how work is divided and supervised, so structure is partly a national product.What it has shownMaurice, Sorge, and Warner (1980) compared closely matched manufacturing units in France, West Germany, and Great Britain and found more pervasive differences in their organization than usually recognized. Whitley (1999) set out a comparative account of business systems.Who disputes itStructural contingency theory stresses culture-free elements of structure; matched comparisons find nationally different shapes of organization.Societal effect and comparative business systems in The Schools of Organization DesignThe network perspective on organizations1985 to the presentWhat it claimsOrganizations and their members are nodes joined by ties, and the pattern of ties shapes what they can do. Economic action is embedded in networks of personal relations (Granovetter 1985), and network forms are a distinct form, not a hybrid of market and hierarchy (Powell 1990).What it has shownPodolny and Page (1998) review the research and find that network forms foster learning, confer status and legitimacy, bring economic benefits, help manage resource dependencies, and give employees considerable autonomy, and they ask for attention to networks' dysfunctions.Who disputes itTransaction cost economics: Williamson treats the hybrid as a third generic form between market and hierarchy; for Powell, network forms are a distinct form, not a mix of the two.The network perspective on organizations in The Schools of Organization DesignComputational organization design and search1990s to the presentWhat it claimsDesign choices interact, so design space is a rugged landscape with many local peaks, and structure affects how well an organization searches it. Decentralizing interdependent decisions can trap units on poor peaks; temporary decentralization, then reintegration, can help.What it has shownMostly simulation, with one field test: Csaszar (2012), with more than 150,000 stock-picking decisions by 609 mutual funds, found that decision-making structure affects the number of initiatives pursued and the omission and commission errors made.Who disputes itNo school in the record disputes it head on; it disputes structural contingency theory, whose bivariate fit, its simulations show, cannot capture how the best design depends on interactions among choices.Computational organization design and search in The Schools of Organization DesignThe new-forms debate1990s to the presentWhat it claimsOrganizations must solve four universal problems (dividing tasks, allocating them, providing rewards, and providing information), and a form is new when it solves them in new ways. Self-managing organizations radically decentralize authority; skeptics hold hierarchy reappears.What it has shownFoss (2003) traced the rise and partial reversal of Oticon's spaghetti organization to the problem of selective intervention. A meta-analysis of 54 studies of 13,914 teams (Greer and colleagues 2018) found that hierarchy within teams on net slightly lowers team performance.Who disputes itBureaucracy and its critics: removing formal structure leaves informal elites whose power is unaccountable (Freeman), and democratic and network organizations are much less hierarchy-free than their labels suggest.The new-forms debate in The Schools of Organization DesignThe algorithmic-management literature2015 to the presentWhat it claimsSoftware does much of the directing, evaluating, and disciplining that managers and rules once did. Kellogg, Valentine, and Christin (2020) call algorithmic control a new contested terrain; Puranam (2021) treats dividing decisions between people and machines as a design problem.What it has shownMostly qualitative field studies of gig work. Wood and colleagues (2019), with 107 interviews in six countries and 679 survey respondents, found that algorithmic management tends to give remote gig workers flexibility and autonomy, but also low pay, isolation, and overwork.Who disputes itLabor process and critical theory reads control at work as contested terrain; Kellogg, Valentine, and Christin take up that view but argue that algorithmic control is distinct from earlier forms.The algorithmic-management literature in The Schools of Organization DesignFunctional structureOrigin: 1911 to 1937.Family: Classical hierarchies.Line and staffOrigin: 1916 to 1943.Family: Classical hierarchies.Multidivisional (M-form) by productOrigin: 1920s practice; named and analyzed 1962 to 1975.Family: Divisionalized forms.Machine bureaucracyOrigin: 1922 (Weber); 1979 (Mintzberg).Family: Classical hierarchies.Divisional by customer or market segmentOrigin: 1937; 2005.Family: Divisionalized forms.Matrix organizationOrigin: 1950s to 1960s practice; analyzed 1971 to 1977.Family: Lateral and multidimensional forms.Co-determined organizationOrigin: 1951 to 1976 (the laws); 1984 to 2022 (the main studies).Family: Forms from other traditions.Socio-technical autonomous work groupsOrigin: 1951 to 1965.Family: Team-based and self-managing forms.Stratified hierarchy (requisite organization)Origin: 1951 to 1989.Family: Classical hierarchies.Cooperative organizationOrigin: 1958 to 1999.Family: Forms from other traditions.Linking-pin organization (System 4)Origin: 1961 to 1967.Family: Team-based and self-managing forms.Divisional by regionOrigin: 1962 to 1972.Family: Divisionalized forms.Adhocracy (innovative organization)Origin: 1968 to 1979.Family: Professional, project, and knowledge forms.J-form (the Japanese firm)Origin: 1968 to 1990.Family: Forms from other traditions.Professional bureaucracyOrigin: 1968 to 1979.Family: Professional, project, and knowledge forms.Holding company (H-form)Origin: 1972 to 1975; 1987.Family: Divisionalized forms.Ambidextrous organizationOrigin: 1976 to 2004.Family: Professional, project, and knowledge forms.Simple structureOrigin: 1979.Family: Classical hierarchies.Hybrid structureOrigin: 1980 to the present.Family: Lateral and multidimensional forms.Missionary (ideological) organizationOrigin: 1980 to 1989.Family: Ideological forms.Network, modular, and virtual organizationOrigin: 1986 to 1990.Family: Network and market-like forms.Transnational and multidimensional organizationOrigin: 1987 to 1989.Family: Lateral and multidimensional forms.Business group (keiretsu, chaebol, family house)Origin: 1988 to 2007.Family: Forms from other traditions.Shamrock organizationOrigin: 1988 to 1989.Family: Network and market-like forms.Federal organizationOrigin: 1989 to 1992.Family: Network and market-like forms.Self-managing organization (less-hierarchical forms)Origin: 1989 to 2017.Family: Team-based and self-managing forms.Boundaryless organizationOrigin: 1990 to 1995.Family: Network and market-like forms.Front-back organizationOrigin: 1990s to 2007.Family: Lateral and multidimensional forms.Professional partnership (P2 form)Origin: 1990.Family: Professional, project, and knowledge forms.The corporate center and shared servicesOrigin: 1991 to 2000.Family: Divisionalized forms.Internal market organizationOrigin: 1993 to 1994.Family: Network and market-like forms.Process (horizontal) organizationOrigin: 1993 to 1999.Family: Lateral and multidimensional forms.Cellular organizationOrigin: 1994 to 1997.Family: Network and market-like forms.Hypertext organizationOrigin: 1994 to 1995.Family: Forms from other traditions.Team-based organizationOrigin: 1995.Family: Team-based and self-managing forms.Project-based organizationOrigin: 2000 to 2006.Family: Professional, project, and knowledge forms.Agile at scaleOrigin: 2001 to 2016.Family: Contemporary and emerging forms.Distributed (all-remote) organizationOrigin: 2002 to 2021.Family: Contemporary and emerging forms.Open-source and community-based organizationOrigin: 2002 to 2007.Family: Contemporary and emerging forms.Algorithmically managed organizationOrigin: 2015 to 2020.Family: Contemporary and emerging forms.Team of teamsOrigin: 2015.Family: Team-based and self-managing forms.

The schools by first year

Bureaucracy and its critics1900s to 1950s, revived from the 1990sWhat it claimsWeber held that rational-legal authority exercised through an office hierarchy, written rules, trained officials, and promotion by qualification is technically superior in precision, speed, and reliability. Critics studied its unintended consequences, Michels's iron law sharpest.What it has shownMerton (1940) argued, with no test, that rules can turn from means into ends; case studies followed. The Aston group concluded that the type's dimensions vary independently, and Wagner's meta-analysis (2021) found that many of their interrelationships have weakened.Who disputes itHuman relations and System 4 sets the costs of impersonal rules to motivation and cooperation against the technical superiority Weber stresses.
Classical administration1910s to 1940s, and in German to the 1970sWhat it claimsThere are general principles of administration, among them division of work, unity of command, a scalar chain of authority, and a limited span of control, and a structure built to them will be efficient in most settings. Taylor added that planners should set work methods.What it has shownThe school offered reasoning and experience, not tests. Woodward (1965) found that the spans and levels of the more successful firms in her sample differed by production system, and Simon (1946) showed that the principles come in pairs that point in opposite directions.Who disputes itStructural contingency theory found that the structure of successful firms varies with technology and environment, where classical writers treat their principles as general.
Human relations and System 41920s to 1960sWhat it claimsOutput depends on the social relations of work, the norms of informal groups, and the style of supervision as much as on formal design. Likert's System 4, overlapping work groups joined by linking-pin managers under participative management, outperforms authoritative systems.What it has shownThe founding evidence is the Hawthorne studies, which later reanalyses dispute (Carey 1967; Levitt and List 2011). Likert's support came largely from his own institute's surveys. Wagner (1994) found participation's effects on performance and satisfaction significant but small.Who disputes itLabor process and critical theory reads human relations as a technique of managerial control rather than a change in the distribution of power.
Transaction cost economics and the M-form hypothesis1937, then the 1970s to the presentWhat it claimsFirms exist, and are organized as they are, to economize on transaction costs. Hierarchy replaces the market when assets are specific, uncertainty is high, and exchange is frequent, and the multidivisional form should outperform the functional form in large, diversified firms.What it has shownArmour and Teece (1978) found, in United States petroleum firms from 1955 to 1973, a positive relationship between the multidivisional form and profitability while the form was spreading, but not once an equilibrium in organizational form was reached.Who disputes itThe network perspective on organizations: Granovetter argued that the theory undersocializes economic action, which is embedded in networks of personal relations that also produce trust and order.
The Carnegie school and the attention-based view1940s to the presentWhat it claimsPeople in organizations are boundedly rational, so structure matters because it sets the premises of their decisions and channels their attention: who sees which problems, which goals they pursue, and which routines they follow. Ocasio makes structure distribute attention.What it has shownMarch and Simon (1958) drew propositions from the existing literature, and Cyert and March (1963) added process models and simulations. Two studies show hierarchy shaping what is sent up as well as what is chosen (Reitzig and Maciejovsky 2015; Keum and See 2017).Who disputes itOrganizational economics: decision rights, authority, and complementarities models a principal and agents with defined objectives; the Carnegie school, firms as coalitions with unresolved goal conflict.
Socio-technical systems1950s to 1980sWhat it claimsAn organization is a social system and a technical system that must be designed together. The same technology admits more than one way of organizing work, and groups that own a whole task and control its variances at the source perform better and are more satisfied.What it has shownTrist and Bamforth (1951) found that longwall mechanization broke up self-regulating work groups. A review of 134 experiments found them extremely successful overall, and Stewart (2006) linked team autonomy to performance; studies of socialist brigades and Volkswagen qualify it.Who disputes itLabor process and critical theory: Kelly argued that labor intensification, not joint optimization, accounts for what socio-technical redesign achieved, and that the part of pay incentives was seriously underestimated.
The economics of the labor-managed firm and of enterprise ownership1958 to the presentWhat it claimsA firm run by its workers behaves differently from one run for investors: Ward modeled the Yugoslav-style firm as maximizing income per worker. Hansmann argued that an ownership form's success depends on the balance between the costs of market contracting and of ownership.What it has shownBurdín (2014) found a 29 percent lower hazard of dissolution for Uruguayan worker-managed firms than for conventional ones, and Fakhfakh, Pérotin, and Gago (2012) found French labor-managed firms as productive or more so in every industry studied.Who disputes itOrganizational economics (Jensen and Meckling) contrasts the labor-managed firm, in which tradable residual claims are legally prohibited, with cooperatives and partnerships that arise out of voluntary contracting.
Modularity and the mirroring hypothesis1960s to the presentWhat it claimsComplex systems are built of nearly decomposable parts, and an organization's structure tends to mirror the architecture of what it makes: modular products allow loosely coupled organizations, and tightly coupled products need tightly coupled teams. Conway's law runs in reverse.What it has shownColfer and Baldwin (2016) review the empirical studies and report both support and principled exceptions. Direct tests point both ways: MacCormack, Baldwin, and Rusnak (2012) found open source products more modular than their commercial pairs; dynamic components break the mirror.Who disputes itNo school in the record disputes it head on; it answers information processing, which meets the coordination load inside the organization; mirroring writers locate it in the product's architecture.
Strategy and structure1960s to 1980sWhat it claimsStructure follows strategy: firms that diversify by product or geography outgrow the centralized functional structure and, after a lag and usually a crisis, adopt divisions. Chandler drew this from the histories of DuPont, General Motors, Standard Oil of New Jersey, and Sears.What it has shownChandler (1962) is comparative business history of four firms set against a wider survey. Rumelt (1974) related diversification and structure to performance, and Egelhoff (1988), with 34 multinationals, found some of Stopford and Wells's relationships supported and others not.Who disputes itInstitutional theory: in Fligstein's model divisions spread with a product-related or unrelated strategy, presidents from sales or finance, and other firms in the industry changing; strategy was one of three.
Structural contingency theory1960s to the presentWhat it claimsThere is no one best structure. Performance depends on fit between structure and contingencies such as environment, technology, size, and strategy, on an open-system view of the organization. Drazin and Van de Ven separate fit as selection, interaction, and systems.What it has shownBurns and Stalker, Woodward, and the Aston group related structure to context; Drazin and Van de Ven (1985) supported selection and systems fit only. Pennings (1975) found that fit did not explain effectiveness; Schoonhoven (1981) found traditional hypotheses unsupported.Who disputes itStrategic choice: Child argued that contingency theory explains structure by the constraints of context and leaves out the political process by which power-holders choose strategic action, structural forms included.
The French bureaucratic school (strategic analysis)1960s to 1980sWhat it claimsA bureaucracy is a system of power relations organized around zones of uncertainty. Impersonal rules and centralization protect members from depending on one another, those who control uncertainties hold power, and, unable to correct its errors, it changes only through crisis.What it has shownLe phénomène bureaucratique (1963; English 1964) rests on two case studies, a clerical agency and a state industrial monopoly. L'acteur et le système (1977; English 1980) is theoretical.Who disputes itSocietal effect and comparative business systems: D'Iribarne argues that national logics of conduct, such as a French logic of honor, explain differences that Crozier attributed to power games.
The Japanese-firm literature1960s to the presentWhat it claimsThe large Japanese firm coordinates operating decisions through horizontal information sharing among semi-autonomous shops and uses its ranking hierarchy mainly as an incentive scheme (Aoki). It creates knowledge through middle-up-down management (Nonaka and Takeuchi).What it has shownAoki (1986) is a formal model. Lincoln, Hanada, and McBride (1986) compared 51 Japanese and 55 United States manufacturing plants, and Nonaka and Takeuchi (1995) rest on case studies, chiefly of Japanese firms. Fruin (1994) found the enterprise system defined at three levels.Who disputes itLabor process and critical theory reads shop-floor delegation as control: Sewell and Wilkinson argue that just-in-time and total quality regimes delegate tactical responsibility while centralizing strategic control.
The Scandinavian and Dutch socio-technical schools1960s to 2000sWhat it claimsThe Scandinavian program sought democracy at work through field experiments with semi-autonomous groups, then through democratic dialogue. The Dutch school's design theory turns complex organizations with simple jobs into simple organizations with complex jobs.What it has shownEmery and Thorsrud reported field experiments in Norwegian industry; Gustavsen (1992) records that they spread less than hoped, his reason for turning to dialogue. Berggren studied work organization in the Swedish auto industry. The Dutch work rests on design cases and argument.Who disputes itThe Japanese-firm literature: defenders of the Toyota system hold that standardized work with employee participation learns faster; Berggren reports Swedish long-cycle plants matching a line plant on productivity.
Configuration theory1970s to the presentWhat it claimsStructures cluster into a small number of internally coherent configurations whose elements reinforce one another, so organizations move between a few types in large jumps, not small steps. Mintzberg's five configurations each rest on one coordinating mechanism.What it has shownDoty, Glick, and Huber (1993) found that fit to Miles and Snow's ideal types predicted effectiveness while fit to Mintzberg's did not. A meta-analysis of 40 tests (Ketchen and colleagues 1997) found that configuration partly explains performance, more so with broad definitions.Who disputes itStructural contingency theory: Donaldson counters that most organizations sit at intermediate points on a continuum of fits, not in a few configurations, and that studies over time find mostly incremental change.
Information processing1970s to the presentWhat it claimsThe more uncertain the task, the more information must be processed during its execution. An organization either reduces that need, through slack resources or self-contained tasks, or raises its capacity, through vertical information systems or lateral relations up to the matrix.What it has shownGalbraith (1973, 1974) built the framework from cases and from Simon's theory of bounded rationality, and Tushman and Nadler (1978) set it out as an integrating concept for design. Daft and Lengel (1986) argued that structure must reduce equivocality as well as supply data.Who disputes itThe Carnegie school and the attention-based view: Ocasio holds that firm behavior results from how the firm channels and distributes attention, so structure governs where attention goes, not only how much is processed.
Labor process and critical theory1970s to the presentWhat it claimsStructure is above all a system of control over labor. Braverman argued that management deskills work by separating conception from execution; Edwards traced a sequence from simple control by the boss to technical control through machines to bureaucratic control through rules.What it has shownHistorical and theoretical (Braverman 1974; Edwards 1979). Delbridge (1995), observing a Japanese-owned plant in England, found that just-in-time and total quality regimes combined the control of labor with management's economic goals more completely than piece rates had.Who disputes itBureaucracy and its critics: Adler argues that labor process theory ignores the progressive socialization of the labor process, which drives long-term skill upgrading, and disputes deskilling as the trend.
Strategic choice1970s to the presentWhat it claimsStructure is the outcome of choices by those with power in the organization, who decide on strategic action, including the structural form, the manipulation of environmental features, and the performance standards to be met. Context constrains but does not determine structure.What it has shownLargely conceptual: Child (1972) built the argument from gaps in contingency findings, and the empirical case rests mostly on the modest share of structural variance contingencies explain. A 2026 review of 42 studies finds upper echelons theory their main theoretical anchor.Who disputes itStructural contingency theory: Donaldson replies that organizations mostly adopt the structures their contingencies dictate, to avoid the loss of performance misfit brings, so choice over structure is at most limited.
The German codetermination literature1970s to the presentWhat it claimsWorks councils and employee board seats form a structural layer through which coordination can become negotiation between management and employees. Defenders argue that it supports long-term employment, training, and cooperation; critics, that it weakens shareholder control.What it has shownJäger, Schoefer, and Heining (2021) found board-level codetermination left wages unchanged and, if anything, raised capital formation. Jirjahn (2011) reports studies pointing to positive works council effects; a meta-analysis found codetermination laws tied to lower productivity.Who disputes itOrganizational economics (Gorton and Schmid) found that firms with equal employee and shareholder representation did worse for their shareholders than firms with one-third employee representation.
Organizational economics: decision rights, authority, and complementarities1976 to the presentWhat it claimsDesign is the joint choice of who holds which decision rights, how performance is measured, and how it is rewarded. Rights should be placed where the specific knowledge lies, formal authority differs from real authority, and practices are complements that pay off in clusters.What it has shownIchniowski, Shaw, and Prennushi (1997) found, on 36 steel finishing lines, that lines using a set of innovative work practices achieved substantially higher productivity, consistent with complementarities. Bloom and colleagues found more decentralization where trust was higher.Who disputes itGhoshal argued that management research has shaped managers' worldview with ideologically inspired amoral theories, built on a scientific model that excludes human intentionality and so denies moral considerations.
Institutional theory1977 to the presentWhat it claimsFormal structures are often adopted because they are legitimate in the organization's field, as rationalized myths, and then decoupled from the work done. Organizations in a field grow alike through coercive, mimetic, and normative isomorphism, whether or not it is efficient.What it has shownTolbert and Zucker (1983) found that city characteristics predicted early but not later adoption of civil service reform, which they read as institutionalization. Palmer, Jennings, and Zhou (1993) found support for institutional and economic accounts of late M-form adoption.Who disputes itStructural contingency theory: Kraatz and Zajac found little support for neoinstitutional predictions in 631 liberal arts colleges, which grew less alike; the change they studied was curricular, not structural.
Population ecology1977 to the presentWhat it claimsOrganizations face strong structural inertia, because selection favors the reliability and accountability of stable structures. Change in a population's forms comes mainly from founding and failure, not adaptation, and changing core features resets the liability of newness.What it has shownAmburgey, Kelly, and Barnett (1993) found that the effects of change on the hazard of failure vary by type of change and over time, and that rates of change depend on how often and how recently an organization has changed. Carroll and Hannan (2000) set out the program's methods.Who disputes itFor structural contingency theory, Donaldson argues that ecology understates managers' capacity to adapt structure.
Resource dependence1978 to the presentWhat it claimsOrganizations depend on outside parties for critical resources, and structure follows the management of that dependence: units that cope with critical dependencies gain power, and boundaries are redrawn through cooptation, joint ventures, and mergers to absorb constraint.What it has shownPfeffer and Salancik (1978) assembled studies of boards, mergers, and joint ventures. A meta-analysis of 157 tests (Drees and Heugens 2013) corroborated the main prediction: organizations meet resource dependencies with interorganizational arrangements that raise their autonomy.Who disputes itTransaction cost economics and the M-form hypothesis explains boundaries by efficiency; resource dependence, by power and constraint.
The business-groups literature1978 to the presentWhat it claimsGroups of legally independent firms bound by ownership and social ties are a durable form of their own, neither one firm nor a market. Where institutions function poorly, groups can benefit member firms and can destroy value; they are sometimes paragons, sometimes parasites.What it has shownKhanna and Palepu (2000) found that affiliates of the most diversified Indian groups outperformed unaffiliated firms. A meta-analysis of 141 studies (Carney and colleagues 2011) found that affiliation lowers performance in general, less so where institutions are underdeveloped.Who disputes itOrganizational economics: agency writers read pyramids and cross-holdings as devices that separate control from ownership; Almeida and Wolfenzon offer a financing rationale for pyramids instead.
Scandinavian institutionalism1980s to the presentWhat it claimsIdeas of organizing do not diffuse unchanged; they travel and are translated as they are adopted, so a form taken up in one place is not the same form elsewhere. Organizations can also be built of organizations (meta-organizations), which behave differently.What it has shownA discussion paper by Czarniawska and Joerges (1995) contrasts the diffusion of ideas with their translation, drawing on examples such as Swedish municipal reform; the work on meta-organizations (Ahrne and Brunsson 2008) is conceptual.Who disputes itInstitutional theory can read the same processes, seen from a translocal view, as institutional isomorphism, though translation holds that each adopter modifies an idea, so an institution re-created elsewhere is new.
Societal effect and comparative business systems1980s to the presentWhat it claimsFirms matched on size, technology, and industry differ systematically by country, because how a society educates, trains, recruits, and promotes its workforce, and its industrial relations, shape how work is divided and supervised, so structure is partly a national product.What it has shownMaurice, Sorge, and Warner (1980) compared closely matched manufacturing units in France, West Germany, and Great Britain and found more pervasive differences in their organization than usually recognized. Whitley (1999) set out a comparative account of business systems.Who disputes itStructural contingency theory stresses culture-free elements of structure; matched comparisons find nationally different shapes of organization.
The network perspective on organizations1985 to the presentWhat it claimsOrganizations and their members are nodes joined by ties, and the pattern of ties shapes what they can do. Economic action is embedded in networks of personal relations (Granovetter 1985), and network forms are a distinct form, not a hybrid of market and hierarchy (Powell 1990).What it has shownPodolny and Page (1998) review the research and find that network forms foster learning, confer status and legitimacy, bring economic benefits, help manage resource dependencies, and give employees considerable autonomy, and they ask for attention to networks' dysfunctions.Who disputes itTransaction cost economics: Williamson treats the hybrid as a third generic form between market and hierarchy; for Powell, network forms are a distinct form, not a mix of the two.
Computational organization design and search1990s to the presentWhat it claimsDesign choices interact, so design space is a rugged landscape with many local peaks, and structure affects how well an organization searches it. Decentralizing interdependent decisions can trap units on poor peaks; temporary decentralization, then reintegration, can help.What it has shownMostly simulation, with one field test: Csaszar (2012), with more than 150,000 stock-picking decisions by 609 mutual funds, found that decision-making structure affects the number of initiatives pursued and the omission and commission errors made.Who disputes itNo school in the record disputes it head on; it disputes structural contingency theory, whose bivariate fit, its simulations show, cannot capture how the best design depends on interactions among choices.
The new-forms debate1990s to the presentWhat it claimsOrganizations must solve four universal problems (dividing tasks, allocating them, providing rewards, and providing information), and a form is new when it solves them in new ways. Self-managing organizations radically decentralize authority; skeptics hold hierarchy reappears.What it has shownFoss (2003) traced the rise and partial reversal of Oticon's spaghetti organization to the problem of selective intervention. A meta-analysis of 54 studies of 13,914 teams (Greer and colleagues 2018) found that hierarchy within teams on net slightly lowers team performance.Who disputes itBureaucracy and its critics: removing formal structure leaves informal elites whose power is unaccountable (Freeman), and democratic and network organizations are much less hierarchy-free than their labels suggest.
The algorithmic-management literature2015 to the presentWhat it claimsSoftware does much of the directing, evaluating, and disciplining that managers and rules once did. Kellogg, Valentine, and Christin (2020) call algorithmic control a new contested terrain; Puranam (2021) treats dividing decisions between people and machines as a design problem.What it has shownMostly qualitative field studies of gig work. Wood and colleagues (2019), with 107 interviews in six countries and 679 survey respondents, found that algorithmic management tends to give remote gig workers flexibility and autonomy, but also low pay, isolation, and overwork.Who disputes itLabor process and critical theory reads control at work as contested terrain; Kellogg, Valentine, and Christin take up that view but argue that algorithmic control is distinct from earlier forms.

The forms by first year

Functional structure: 1911 to 1937
Line and staff: 1916 to 1943
Multidivisional (M-form) by product: 1920s practice; named and analyzed 1962 to 1975
Machine bureaucracy: 1922 (Weber); 1979 (Mintzberg)
Divisional by customer or market segment: 1937; 2005
Matrix organization: 1950s to 1960s practice; analyzed 1971 to 1977
Co-determined organization: 1951 to 1976 (the laws); 1984 to 2022 (the main studies)
Socio-technical autonomous work groups: 1951 to 1965
Stratified hierarchy (requisite organization): 1951 to 1989
Cooperative organization: 1958 to 1999
Linking-pin organization (System 4): 1961 to 1967
Divisional by region: 1962 to 1972
Adhocracy (innovative organization): 1968 to 1979
J-form (the Japanese firm): 1968 to 1990
Professional bureaucracy: 1968 to 1979
Holding company (H-form): 1972 to 1975; 1987
Ambidextrous organization: 1976 to 2004
Simple structure: 1979
Hybrid structure: 1980 to the present
Missionary (ideological) organization: 1980 to 1989
Network, modular, and virtual organization: 1986 to 1990
Transnational and multidimensional organization: 1987 to 1989
Business group (keiretsu, chaebol, family house): 1988 to 2007
Shamrock organization: 1988 to 1989
Federal organization: 1989 to 1992
Self-managing organization (less-hierarchical forms): 1989 to 2017
Boundaryless organization: 1990 to 1995
Front-back organization: 1990s to 2007
Professional partnership (P2 form): 1990
The corporate center and shared services: 1991 to 2000
Internal market organization: 1993 to 1994
Process (horizontal) organization: 1993 to 1999
Cellular organization: 1994 to 1997
Hypertext organization: 1994 to 1995
Team-based organization: 1995
Project-based organization: 2000 to 2006
Agile at scale: 2001 to 2016
Distributed (all-remote) organization: 2002 to 2021
Open-source and community-based organization: 2002 to 2007
Algorithmically managed organization: 2015 to 2020
Team of teams: 2015

Twenty-three of the 29 schools run to the present, and 16 of those began in the 1970s or later. The forms arrive in waves: 14 of the 41 first appear between 1985 and 1994.

Source: Fayol 1949; Weber 1978; Mayo 1933; Trist and Bamforth 1951, an interview-based field study in one industry; Burns and Stalker 1961, a field study of twenty firms; Mintzberg 1979; Child 1972a; Coase 1937; Meyer and Rowan 1977; Hannan and Freeman 1977; Galbraith 1973; Jensen and Meckling 1976; Crozier 1963; Emery and Thorsrud 1969; Yoshino 1968; Streeck 1984a; Kellogg and colleagues 2020, a review of earlier research through one theoretical lens, with no new data; Puranam and colleagues 2014, a conceptual argument; Chandler 1962, several independent case histories by a historian; Pfeffer and Salancik 1978; Simon 1946; Maurice and colleagues 1980, a comparison of matched factories in three countries; Levinthal 1997; Simon 1962; Braverman 1974; Hamilton and Biggart 1988, a comparison of three economies by scholars; Ward 1958; Czarniawska and Joerges 1995; Granovetter 1985.

Where it fits

Size. Fits medium and large; partial small: two lines need enough work to repay them, and Burton, Obel, and Håkonsson advise a matrix only for strong reasons (2015, a conceptual essay with no new data). No study tests size.
Age. Contested. Greiner (1972), a conceptual model with no systematic test, puts a matrix-type structure in the latest growth phase, though his later commentary confines it largely to senior management: partial young, fits established and old. DeSantola, Gulati, and Zhelyazkov (2023), a longitudinal quantitative study of thousands of young ventures, found divergent paths, not uniform phases. Hannan and Freeman (1984): inertia rises with age and attempts at reorganization raise death rates, so reorganizing an old organization is a hazard, partial at old.
Environmental uncertainty. Contested. In theory (Galbraith 1973; Burton and Obel 2004): strain stable, fits shifting and unpredictable. Patanakul, Chen, and Lynn (2012), a survey of several hundred projects: partial at unpredictable, autonomous teams doing best on technically novel projects; it measured project novelty, not the environment, and reads as the matrix only by judgment.
Environmental complexity. Strain simple, partial moderate, fits complex. In theory, Mintzberg (1980) places the adhocracy, built with matrix structures, in complex, dynamic environments; extending that to every matrix is the cell's reading, and strain at simple ours, from the strong-reasons advice.
Munificence and hostility. Strain hostile, partial elsewhere, our judgment, applying Burton and Obel's (2004) call for less organizational complexity under hostility to a dual hierarchy. Pennings (1975), a quantitative field test of the structural-contingency model, found environment and structure fit did not explain effectiveness, a limit on all three environment fits.
Diversity. Contested, evidence. The Stopford and Wells model, its diversity tables known only through later accounts: strain single business, partial related, fits unrelated, the unrelated reading ours, since it counts product lines abroad, not relatedness. Egelhoff (1988), a multi-firm study of 34 multinationals, found only some relationships held and revised it; Wolf and Egelhoff (2002), a statistical study of 95 German firms, found general support for the traditional fits and extended the theory to matrix types. Qiu and Donaldson (2012), a preliminary test on a sample of German multinationals, tie matrix use to corporate integration and area diversification, which the cell's judgment reads as partial single, fits related, strain unrelated.
Strategy. Contested; theory, untested against strategy type. Burton and Obel (2004): partial defender, fits analyzer and prospector. Miles and Snow (1978) and Burton, Obel, and Håkonsson's later book (2015): partial defender and prospector, fits analyzer, since they pair the prospector with product or divisional structures.
Technology. Contested, evidence. Strain routine, fits mixed and nonroutine: Larson and Gobeli's favorable matrix results in 540 development projects (1988) and in a quantitative study of 546 (1989), the first's ratings possibly sharing a source with the structure report, read as the upper levels by judgment; the routine level rests on theory (Galbraith 1973; Burton and Obel 2004). Patanakul and colleagues, reading heavyweight teams as the matrix (a judgment): partial at nonroutine. This cell is limited by three tests naming no matrix: Hickson, Pugh, and Pheysey (1969), a multi-organization comparative survey in the English Midlands, found technology shaping only structure near the workflow; Mohr (1971), a quantitative field study of technology and structure, found no support for effectiveness turning on technology's match with structure, a null Donaldson (2001) lays, with Pennings's, to method; and Schoonhoven (1981), a quantitative study of hospital operating room suites, found traditional Galbraith hypotheses unsupported and more precise ones better supported.
Interdependence. In theory: strain pooled, partial sequential, fits reciprocal (Galbraith 1973; Thompson 1967; Burton and Obel 2004: divisible work does not need a matrix). Mutual adjustment, the matrix's mechanism, is Thompson's costliest; strain at pooled is the table's reading.
Knowledge intensity. Partial low, fits moderate and high, in theory: Mintzberg's (1980) adhocracy houses experts in functions and deploys them in teams through matrix structure, extended to every matrix by the cell's reading; low is ours.
Global integration pressure. Contested, evidence. Qiu and Donaldson: fits high, partial below, but only where integration pressure comes with area diversification, not for a global firm with little of it, an association with use, not performance. Devinney, Midgley, and Venaik (2000), a formal model with no new data, and Wolf and Egelhoff (2013), a survey of 82 German multinationals: partial throughout, since firms configure many ways and product-by-region matrices carried more conflict.
Local responsiveness. Partial throughout, our judgment: responsiveness alone is served by area divisions (after Qiu and Donaldson).

The matrix organization against the twelve contingency factors

Each factor at its three levels, Fits, Partial, or Strain, or Contested where a contested cell's views disagree; select a row for its reason, sources, and grade.

Select a factor to read its reason, sources, basis, and grade.Size. Two lines of authority need enough managers and work on both dimensions to be worth their cost; Burton and Obel still allow a matrix configuration in a small firm when the situation calls for it. Basis: Theory. Grade: conceptual only, no empirical test. Sources: Burton and Obel 2004; Galbraith 1971; Burton and colleagues 2015a.Age. A matrix crosses existing functional and product or project lines, so it usually comes after both have been built; Greiner places a matrix-type structure in the latest phase of growth. Basis: Theory. Grade: conceptual only, no empirical test. Sources: Greiner 1972; Worren 2017; Hannan and Freeman 1984; Amburgey and colleagues 1993. View 1: Young partial, Established fits, Old fits. A matrix crosses existing functional and product or project lines, so it usually comes after both have been built; Greiner places a matrix-type structure in the latest phase of growth. Sources: Greiner 1972; Worren 2017. View 2: Young partial, Established fits, Old partial. Inertia rises with age, and attempts at reorganization raise death rates, so reorganizing an old organization into this form is itself a hazard at the old level. Sources: Hannan and Freeman 1984.Environmental uncertainty (unpredictability, dynamism). Dual lines add the information-processing capacity that uncertainty demands; in a stable environment that capacity costs more than it returns. Basis: Theory. Grade: conceptual only, no empirical test. Sources: Galbraith 1973; Burton and Obel 2004; Mintzberg 1980; Burton and colleagues 2015a; Patanakul and colleagues 2012; Larson and Gobeli 1988; Burton and colleagues 2020a; Burns and Wholey 1993; Pennings 1975. View 1: Stable strain, Shifting fits, Unpredictable fits. The matrix is the top of Galbraith's lateral relations, the answer to the highest information-processing load, and Burton and Obel's model lets an uncertain situation fit a matrix configuration. Sources: Galbraith 1973; Burton and Obel 2004; Burton and colleagues 2015a. View 2: Stable strain, Shifting fits, Unpredictable partial. In 555 development projects, autonomous teams were more effective than the other structures for high technological novelty or radical innovation, so at the highest uncertainty the matrix gives way to a dedicated team. Sources: Patanakul and colleagues 2012.Environmental complexity. Two lines of authority are worth their cost when the environment presents many dissimilar demands at once; in a simple environment the matrix is overhead without a purpose. Basis: Theory. Grade: conceptual only, no empirical test. Sources: Mintzberg 1980; Burton and colleagues 2015a; Pennings 1975.Munificence and hostility. Two bosses and negotiated decisions are slow and costly, the opposite of the simpler, faster response with less organizational complexity that a hostile environment calls for. Basis: Judgment. Sources: Burton and Obel 2004; Burton and colleagues 2015a; Pennings 1975.Diversity of products and markets. Works disagree on what drives a multinational matrix: high product diversity joined to high foreign sales, or the need to integrate units across areas. Basis: Evidence. Grade: a multi-firm empirical study or several independent case studies by scholars. Sources: Stopford and Wells 1972; Qiu and Donaldson 2012; Wolf and Egelhoff 2002. View 1: Single or dominant business strain, Related diversification partial, Unrelated diversification fits. In the Stopford and Wells model, which a test of 95 German firms broadly supported and extended to matrix types, the matrix fits multinationals high on both foreign product diversity and foreign sales. Sources: Stopford and Wells 1972; Wolf and Egelhoff 2002. View 2: Single or dominant business partial, Related diversification fits, Unrelated diversification strain. In the revised model corporate integration and area diversification, not product diversification, go with matrix use; reading that as the matrix fitting related businesses best, since unrelated ones have least to integrate, is this cell's judgment. Sources: Qiu and Donaldson 2012.Strategy type. Miles and Snow, Burton and Obel (2004), and Burton, Obel, and Håkonsson (2015) all pair the analyzer with the matrix; the 2004 rules also list it for a prospector, which Miles and Snow and the later book (2015) do not, and a defender is given a functional design instead. Basis: Theory. Grade: conceptual only, no empirical test. Sources: Miles and Snow 1978; Burton and Obel 2004; Burton and colleagues 2015b; Burton and colleagues 2020a. View 1: Defender (exploitation) partial, Analyzer (exploitation and exploration) fits, Prospector (exploration) fits. Burton and Obel's 2004 rules give the analyzer with innovation a matrix or divisional configuration and list the matrix for a prospector; a defender is given a functional design. Sources: Burton and Obel 2004. View 2: Defender (exploitation) partial, Analyzer (exploitation and exploration) fits, Prospector (exploration) partial. Miles and Snow pair the prospector with a product structure, not a matrix, and the later book (2015) gives the prospector the divisional configuration and the analyzer the matrix. Sources: Miles and Snow 1978; Burton and colleagues 2015b.Technology and routineness. Matrix project structures compared favorably with dedicated project teams in two studies of development projects, though autonomous teams did best on technically novel projects in another; with routine work, rules and planning make the dual line an unneeded cost. Basis: Evidence. Grade: a multi-firm empirical study or several independent case studies by scholars. Sources: Patanakul and colleagues 2012; Galbraith 1973; Galbraith 1971; Burton and Obel 2004; Larson and Gobeli 1988; Larson and Gobeli 1989; Hickson and colleagues 1969; Mohr 1971. View 1: Routine strain, Mixed fits, Nonroutine fits. In 540 and then 546 development projects the balanced and project matrix compared favorably with dedicated project teams and functional designs did worst; with routine work, rules and planning make the dual line an unneeded cost. Sources: Larson and Gobeli 1988; Larson and Gobeli 1989; Burton and Obel 2004; Galbraith 1973. View 2: Routine strain, Mixed fits, Nonroutine partial. Among functional, lightweight, heavyweight, and autonomous project teams, heavyweight teams did best on incremental innovation and autonomous teams on technically novel projects, which, reading heavyweight teams as the matrix, puts a separate team ahead of the matrix for nonroutine work. Sources: Patanakul and colleagues 2012; Burton and Obel 2004.Interdependence. Two lines of authority are the costliest lateral device, worth it when units must adjust to each other constantly; when work divides cleanly, Burton and Obel say the matrix is not needed. Basis: Theory. Grade: conceptual only, no empirical test. Sources: Galbraith 1973; Thompson 1967; Burton and Obel 2004.Knowledge intensity and professionalization. The matrix lets specialists keep a functional home while they work on projects or products, the arrangement Mintzberg's adhocracy uses for highly trained experts; with work learned on the job there is less depth of expertise for the second line to protect. Basis: Theory. Grade: conceptual only, no empirical test. Sources: Mintzberg 1980; Galbraith 1971; Larson and Gobeli 1988.Global integration pressure. Multinationals pursuing corporate integration across many areas tend to use a product-by-area matrix, which holds the global and the local line together; product-by-region matrices also carried more internal conflict than other structures. Basis: Evidence. Grade: a multi-firm empirical study or several independent case studies by scholars. Sources: Qiu and Donaldson 2012; Wolf and Egelhoff 2002; Stopford and Wells 1972; Devinney and colleagues 2000; Wolf and Egelhoff 2013. View 1: Low partial, Moderate partial, High fits. Multinationals pursuing corporate integration together with diversification across many areas tended to use a product-by-area matrix, which holds the global and the local line together. Sources: Qiu and Donaldson 2012. View 2: Low partial, Moderate partial, High partial. Global firms can meet competing integration and responsiveness pressures with a variety of configurations, and matrices of product divisions crossed with regions carried more conflict than single-dimension structures. Sources: Devinney and colleagues 2000; Wolf and Egelhoff 2013.Local responsiveness pressure. In a product-by-area matrix the area line carries local adaptation, but the second line pays off only when integration pressure is also strong, so responsiveness alone gives no reason for or against the matrix. Basis: Judgment. Sources: Qiu and Donaldson 2012; Stopford and Wells 1972.

Counting each contested cell by its first view, the 36 settings read Fits at 16, Partial at 14, and Strain at 6, the strains at the low end of six factors: a stable, simple, or hostile environment, a single business, routine work, and pooled interdependence. Six of the twelve cells are contested, and two rest on judgment.

Source: Burton and Obel 2004; Galbraith 1971; Burton and colleagues 2015a, a conceptual essay with no new data; Greiner 1972, a conceptual model with no systematic test; Worren 2017, one case study; Hannan and Freeman 1984; Amburgey and colleagues 1993, a quantitative study of organizational change and failure rates; Galbraith 1973; Mintzberg 1980, a synthesis of earlier research, offering propositions without a new test; Patanakul and colleagues 2012, a survey of several hundred projects; Larson and Gobeli 1988, a comparison of 540 development projects; Burton and colleagues 2020a; Burns and Wholey 1993, a quantitative study of matrix adoption and abandonment in a group of hospitals; Pennings 1975, a quantitative field test of the structural-contingency model; Stopford and Wells 1972, a study of 187 U.S. multinational enterprises; Qiu and Donaldson 2012, a preliminary test on a sample of German multinationals; Wolf and Egelhoff 2002, a statistical study of 95 German firms; Miles and Snow 1978; Burton and colleagues 2015b; Larson and Gobeli 1989, a quantitative study of 546 development projects; Hickson and colleagues 1969, a multi-organization comparative survey; Mohr 1971, a quantitative field study of technology and structure; Thompson 1967; Devinney and colleagues 2000, a formal model with no new data; Wolf and Egelhoff 2013, a survey of 82 German multinationals.

Your own situation against the matrix organization

Set any factor to outline its cell in the strip above; the lines under the switches give the counts, the form's place among the 41, and the misfit rules that fire.

Size

Age

Environmental uncertainty (unpredictability, dynamism)

Environmental complexity

Munificence and hostility

Diversity of products and markets

Strategy type

Technology and routineness

Interdependence

Knowledge intensity and professionalization

Global integration pressure

Local responsiveness pressure

Set: 0 of 12 factors. Fits on 0, Partial on 0, Strain on 0, Contested on 0.

Its place in the explorer's rankingEvery factor is on Any: all 41 forms tie, and the explorer lists them alphabetically.1st of 411st of 41, tied with one other1st of 41, tied with 0 othersSunk by a misfit rule: 1st of 41Sunk by a misfit rule: 1st of 41, tied with one otherSunk by a misfit rule: 1st of 41, tied with 0 others

Sinks the form: A defender run through a matrix. Source: Burton and Obel 2004; Burton and colleagues 2015b; Burton and colleagues 2020a.
Flagged (its sources disagree): A prospector run through a matrix. Disputed: Burton and colleagues 2015b; against it, Burton and Obel 2004; Larson and Gobeli 1988.
Sinks the form: Routine work run through a matrix. Source: Burton and Obel 2004.
Flagged (its sources disagree): Separable work run through a matrix. Disputed: Burton and Obel 2004; against it, Siggelkow and Levinthal 2003.
Flagged (its sources disagree): A matrix in a stable environment. Disputed: Burton and Obel 2004; Burton and colleagues 2015b; against it, Burton and Obel 2004; Qiu and Donaldson 2012; Wolf and Egelhoff 2002.

Ranking rule: by count of Fits on the factors the reader has set, then fewest Strains, then alphabetical; a fired misfit rule that no source disputes sinks the form to the bottom when every level of the form's profile meets it, a disputed one, or one that meets the form only through one end of a spanned level, one view of a contested profile cell, or one part of a two-part form, is flagged on the row (with both views where it is disputed), and a misfit between two of the reader's own settings is shown once above the list; a contested cell counts only where its views agree at the reader's level, and otherwise shows as Contested; factors on "Any" count for nothing; with every factor on "Any" the list is alphabetical and says so.

Open these settings in the explorer

The misfit rules that can fire for this form

A defender run through a matrix fires when Strategy type at Defender (exploitation): sinks the form. Source: Burton and Obel 2004; Burton and colleagues 2015b; Burton and colleagues 2020a.
A prospector run through a matrix fires when Strategy type at Prospector (exploration): flagged (its sources disagree). Disputed: Burton and colleagues 2015b; against it, Burton and Obel 2004; Larson and Gobeli 1988.
Routine work run through a matrix fires when Technology and routineness at Routine: sinks the form. Source: Burton and Obel 2004.
Separable work run through a matrix fires when Interdependence at Pooled: flagged (its sources disagree). Disputed: Burton and Obel 2004; against it, Siggelkow and Levinthal 2003.
A matrix in a stable environment fires when Environmental uncertainty (unpredictability, dynamism) at Stable: flagged (its sources disagree). Disputed: Burton and Obel 2004; Burton and colleagues 2015b; against it, Burton and Obel 2004; Qiu and Donaldson 2012; Wolf and Egelhoff 2002.

Five of the 39 contingency misfit rules name the matrix. Two are undisputed and would sink it, a defender and routine work; three are disputed and would only flag it, a prospector, separable work, and a stable environment.

Source: Burton and Obel 2004; Burton and colleagues 2020a.

What it does well

It keeps what a single line trades away: specialists retain a functional home while the project line holds them to delivery. In 540 projects the balanced and project matrix compared favorably with dedicated teams on cost, schedule, and technical performance (Larson and Gobeli 1988), and a survey of 500 managers experienced in product development judged the project matrix most effective (1987). Katz and Allen (1985), a survey of 86 matrixed project teams in nine R&D organizations (read in the working paper; the 1997 reprint agrees), found performance highest when project managers held influence in the wider organization and functional managers over technical details.

On Galbraith's (1973, 1974) theory it adds information capacity where hierarchy overloads; Ocasio's (1997) attention-based view answers that structure governs where attention goes, not only information flow. Sytch and Wohlgezogen's (2026) narrative review reads cross-sectional studies as suggesting gains in technical success, new-product implementation, and collaboration.

What it costs

A second line of management is the plain cost, which Burton, Obel, and Håkonsson's 2015 book says only a dual focus on efficiency and effectiveness justifies. In Joyce (1986), a quasi-experiment in one firm's engineering division, with a control group, formal communication rose in quantity and fell in quality, though the effects landed almost wholly on a relocated drafting group; engineers with prior liaison roles showed few or none, and neither group showed more role conflict. Sytch and Wohlgezogen's review reads the same studies as suggesting matrices may also reduce role clarity; Sytch, Wohlgezogen, and Zajac (2018), a quantitative study of many firms' alliances with interviews, found the market discounting matrix firms in multifunctional alliances.

How it fails

How the matrix organization fails, and the coordination failures behind it

The entry's failure modes at the left, each tied to the Coordination Failures patterns it names at the right; a pattern links its entry there.

Jurisdictional gap or overlap Consensus split Veto deadlock Regulator overload Self-competitive procurement Coordination overhead overrun

How many ties the failure has. Highest: No one is sure who owns whatTwo bosses give conflicting ordersDecisions stall between the lines and climb to the shared bossProjects fight over shared specialistsMore communication, less understandingJurisdictional gap or overlapConsensus splitVeto deadlockRegulator overloadSelf-competitive procurementCoordination overhead overrun

Select a failure to read its place in the network.No one is sure who owns what. Degree 1Two bosses give conflicting orders. Degree 1Decisions stall between the lines and climb to the shared boss. Degree 2Projects fight over shared specialists. Degree 1More communication, less understanding. Degree 1Jurisdictional gap or overlap. Degree 1Consensus split. Degree 1Veto deadlock. Degree 1Regulator overload. Degree 1Self-competitive procurement. Degree 1Coordination overhead overrun. Degree 1

Failure	Degree
No one is sure who owns what	1
Two bosses give conflicting orders	1
Decisions stall between the lines and climb to the shared boss	2
Projects fight over shared specialists	1
More communication, less understanding	1
Jurisdictional gap or overlap	1
Consensus split	1
Veto deadlock	1
Regulator overload	1
Self-competitive procurement	1
Coordination overhead overrun	1

Five failure modes name six patterns, and only the stall between the lines names two: veto deadlock, and regulator overload at the shared boss.

Source: Kuprenas 2003, one case study; Joyce 1986, a quasi-experiment in one firm's engineering division, with a control group; Katz and Allen 1985, a survey of 86 matrixed project teams in nine R&D organizations; Wolf and Egelhoff 2013, a survey of 82 German multinationals; Sy and Côté 2004; Davis and Lawrence 1977; Galbraith 1974.

No one is sure who owns what. A jurisdictional gap or overlap. Kuprenas (2003) found conflict over the roles of functional and project managers, Joyce (1986) worse role perceptions, and Katz and Allen (1985) most projects lacking the best split of influence.

Two bosses give conflicting orders. Each line claims the last word, a consensus split. The evidence is narrower: Wolf and Egelhoff (2013) found more conflict than in single-dimension structures only where products cross regions.

Decisions stall and climb. The lines block each other (veto deadlock) and the shared superior overloads (regulator overload), on our reading of where conflicts go and Sy and Côté's (2004) list of slow decisions, offered without new data. Galbraith (1974) reads it the other way: lateral relations up to the matrix relieve overload at the top.

Projects fight over shared specialists. Self-competitive procurement: Kuprenas found projects and resources politicized.

More communication, less understanding. Coordination overhead overrun, Joyce's finding.

Greenwood and colleagues (2010) call the two-dimensional matrix a failed form (their claim, not a test) and set their multiplex form, under Professional partnership, against it.

A defender sinks the form, paying for a second line without the payoff, as does routine work run through a structure built for exceptions (Burton and Obel 2004): two undisputed misfit rules, both propositions. Three are disputed. A prospector: the later book (2015) pairs it with divisions, while the 2004 rules recommend a matrix and matrices did well in development projects (Larson and Gobeli 1988), a prospector's core work by our reading, though at the project level. Separable work: a misfit for Burton and Obel, though a simulation with no field data found even decomposable problems gained from temporary interdependence (Siggelkow and Levinthal 2003). A stable environment: a misfit in the propositions, while the 2004 rules recommend a matrix under high equivocality, high complexity, and low uncertainty, multinationals' matrix use followed integration and area strategies, and Wolf and Egelhoff's (2002) German firms supported the traditional fits, extended to matrix types, so strategy may set its fit as much as the environment.

Burton, Lauridsen, and Obel (2002), a quantitative study of 224 Danish small and medium-sized firms read in its abstract, tested an earlier edition's rules only as a set: misfits went with lower returns, and added misfits brought no further loss. A single factor-to-structure pairing is the interaction form of fit, which Drazin and Van de Ven (1985), a quantitative study of 629 work units in two states' employment security agencies, found unsupported. Volberda and colleagues (2012), a survey of 3,259 respondents in 1,904 companies, found contingency and institutional fit complementary, though firms in perfect institutional fit lose slightly striving for contingency fit.

Where a decision stalls in a matrix

Select a stage to read whose desk the decision is on, or play the sequence; the dashed stages are where it stalls, on our reading of where conflicts go.

Back to the two lines (Shared superior to Function head)

Back to the two lines (Shared superior to Function head)

Select a stage to read whose desk the decision is on, or play the sequence.Two-boss manager. A matrix puts the same people under two lines of authority at once: typically an engineer belongs to a functional department, which holds the technical standard, and to a project, whose manager sets deadlines. A decision that touches both needs both.Function head. The functional line weighs the decision against the technical standard its department holds.Project manager. The project line weighs the same decision against the deadline it sets.Both line heads. Stalls here When the two lines disagree, each claims the last word, and the lines block each other. On our reading of where conflicts go, this is a veto deadlock.Shared superior. Stalls here A conflict the lines cannot settle climbs to the superior who sits above both, and that desk overloads: regulator overload, on our reading. Sy and Côté (2004) list slow decisions among the matrix's problems, without new data.Back at the two-boss manager. The shared superior settles the question, or sends it back to the two lines; once they agree, the decision returns to the person who carries it out.

Two-boss manager. A matrix puts the same people under two lines of authority at once: typically an engineer belongs to a functional department, which holds the technical standard, and to a project, whose manager sets deadlines. A decision that touches both needs both.
Function head. The functional line weighs the decision against the technical standard its department holds.
Project manager. The project line weighs the same decision against the deadline it sets.
Both line heads. Stalls here. When the two lines disagree, each claims the last word, and the lines block each other. On our reading of where conflicts go, this is a veto deadlock.
Shared superior. Stalls here. A conflict the lines cannot settle climbs to the superior who sits above both, and that desk overloads: regulator overload, on our reading. Sy and Côté (2004) list slow decisions among the matrix's problems, without new data.
Back at the two-boss manager. The shared superior settles the question, or sends it back to the two lines; once they agree, the decision returns to the person who carries it out.

A decision that needs both lines can stall twice: between the two line heads, who block each other, and at the shared superior, who overloads as conflicts climb. In Galbraith's (1974) reading, lateral relations up to the matrix relieve overload at the top instead.

Source: Sy and Côté 2004; Davis and Lawrence 1977; Galbraith 1974.

What the research found

Egelhoff, Wolf, and Adzic (2013), a survey of 57 German multinationals that use a matrix, supported information-processing hypotheses, comparing matrix types, not matrix against other forms.

Burns and Wholey (1993), a quantitative study of matrix adoption in a group of hospitals, found it followed network position and peers' adoption as well as task diversity, the imitation and legitimacy Meyer and Rowan (1977) and DiMaggio and Powell (1983) describe. The contingency side answers with Donaldson's critique (1995), which casts institutional theory as picturing facade structures conforming to myths and argues from the evidence for adaptation to regain fit, and with Kraatz and Zajac (1996), a longitudinal study of 631 private liberal arts colleges, which found little support for neoinstitutional predictions, though for curricular, not structural, change.

Fit to Miles and Snow's types predicted 24 percent of the variance in effectiveness in Doty, Glick, and Huber (1993), a quantitative test across a sample of organizations, which found no support for resembling Mintzberg's ideal types making organizations more effective, limiting the complexity and knowledge-intensity fits. Ketchen and colleagues' (1997) meta-analysis of 40 tests found configurations partly explain performance, though it pools definitions, many of them strategy types, and tests of pure structural types stay mixed; and Krabberød (2015) argues the negative test was not shown to be exhaustive.

Sytch and Wohlgezogen (2026), the most recent item, conclude that robust evidence on matrix performance is scant: most claims rest on anecdotes or single cases. The review states no search method, and both its authors, with Zajac, wrote one study it reviews.

The evidence ledger holds 21 studies filed under this form.

In practice: one published case

The City of Los Angeles Bureau of Engineering, in one case study (Kuprenas 2003). A new mayor elected in 1993 would no longer tolerate past cost overruns and delays, and studies from 1994 to 1996 recommended moving the Bureau from its functional organization toward project management. In February 1997 the Bureau's executives and the Board of Public Works mandated a matrix with a project manager responsible for each project.

The Bureau chose what the source calls a hybrid functional matrix, built around projects within four programs, each program manager holding complete authority over its projects.

Implementation brought problems research predicts, among them conflict over the roles of functional and project managers and the politicization of projects and resources. The Bureau answered with ten fundamental tasks for each kind of manager, written Handshake Agreements on functional commitments, a project control system, and training.

Design costs did not change significantly under the matrix. Time from work order to construction award fell 40 percent in the Street Program, and 39 of 41 Street projects and 25 of 25 Stormwater projects were delivered.

Despite implementation problems matching the research literature, the source concludes, performance under the matrix improved with respect to project delivery; it calls for continued monitoring and more performance measures.

Source: Kuprenas 2003. Implementation and performance of a matrix organization structure, International Journal of Project Management 21(1), 51 to 62; one case study.

Variants and near relatives

Functional, balanced, and project matrix. Where authority sits, from function heads, through shared authority, to project managers; Larson and Gobeli (1987) name the three types. Midler (1995), one longitudinal case study, follows Renault to strong Project Directors in 1988, balanced on purpose against strong departmental identities. Past the project end lies the Project-led organization (Hobday), where the needs of projects outweigh functional influence on decisions (Dosi, Hobday, and Marengo 2003), taught under the Project-based organization; Hobday (2000), one case study with a conceptual model, finds the pure form weak where the matrix is strong and liable to work against corporate strategy.

Five ways to run a development project, and what three studies found

Larson and Gobeli's five structures in order from the functional form to the project team, the two extremes Galbraith (1971) sets the intermediate matrix designs between. Select a structure to read where authority sits and what each study found; a study that left a structure out says so. All three compare structures of projects, not of whole organizations.

Functional organization

Studied in: 1988, 1989
Where authority sits: The studies name it; the records say nothing of where its authority sits.
500 managers, 1987: Not covered: the survey asked only about the three matrix types.
540 projects, 1988: Little support for it.
546 projects, 1989: Less successful than the balanced matrix, project matrix, and project team, with other determinants of success accounted for.

Functional matrix

Studied in: 1987 to 1989
Where authority sits: Most authority with the function heads.
500 managers, 1987: Used at a rate comparable to the balanced and project matrix.
540 projects, 1988: Little support for it.
546 projects, 1989: Less successful than the balanced matrix, project matrix, and project team, with other determinants of success accounted for.

Balanced matrix

Studied in: 1987 to 1989
Where authority sits: Shared between the function heads and the project managers.
500 managers, 1987: Used at a rate comparable to the functional and project matrix.
540 projects, 1988: Compared favorably with dedicated project teams on cost, schedule, and technical performance.
546 projects, 1989: More successful than the functional organization and functional matrix; behind the project matrix in meeting schedule.

Project matrix

Studied in: 1987 to 1989
Where authority sits: Most authority with the project managers.
500 managers, 1987: Used at a comparable rate, and judged the most effective of the three matrix types.
540 projects, 1988: Compared favorably with dedicated project teams on cost, schedule, and technical performance, and managers recommended it more strongly than the project team.
546 projects, 1989: More successful than the functional organization and functional matrix; ahead of the balanced matrix in meeting schedule and of the project team in controlling cost.

Project team

Studied in: 1988, 1989
Where authority sits: The studies name it; the records say nothing of where its authority sits.
500 managers, 1987: Not covered: the survey asked only about the three matrix types.
540 projects, 1988: The benchmark: the balanced and project matrix compared favorably with dedicated teams, against Peters and Waterman's popular conclusion that teams are superior for new products. Managers recommended it less strongly than the project matrix.
546 projects, 1989: More successful than the functional organization and functional matrix; behind the project matrix in controlling cost.

The functional organization and the functional matrix draw little support in 1988 and fare worst in 1989, and the project matrix comes out ahead on the measures the studies name: judged the most effective matrix type by managers in 1987, recommended over the project team in 1988, and ahead of the balanced matrix on schedule and of the team on cost in 1989. The 1987 ratings are the managers' own judgment, the 1988 abstract does not say who rated performance, and only the 1989 abstract was read, so how that study measured success, and who reported it, was not checked.

Source: Larson and Gobeli 1987, a survey of 500 managers experienced in product development; Larson and Gobeli 1988, a comparison of 540 development projects; Larson and Gobeli 1989, a quantitative study of 546 development projects; Galbraith 1971.

Development-center organization. Concurrent projects grouped into centers by shared platform, plus a center for shared components: Toyota in 1992 and 1993 (Nobeoka 1996, one case study built on interviews). It groups projects; the range above apportions authority within one.

Matrix as a frame of mind. Bartlett and Ghoshal's (1990) title argument: a frame of mind, not a structure.

Global matrix. Worldwide product and area divisions established together, the grid Stopford and Wells (1972) expected. The Transnational and multidimensional organization gives subsidiaries different roles in a network, not every manager the same two bosses.

Dual authority in government. Most offices in China's local governments answer to the territorial government (kuai) and their functional line (tiao) (Zhou and Sui 2025). In Mertha (2005), one scholar's study of one national reform, read in its abstract, fragmentation appears to have shifted from geographic to functional lines once regulators were moved up to provincial management. The prefectoral system (Smith 1967) is taught under Divisional by region.

Academic matrix. Discipline crossed with enterprise, large, permanent, and unplanned (Clark 1983), a matrix of membership more than of two bosses, near the Professional bureaucracy.

The Hybrid structure is the nearest neighbor, which gives different units different bases, each with one boss; the matrix lays two bases over the same people. Front-back organization builds coordination into linking processes between two halves, not a second boss, and a Hypertext organization member sits in one layer at a time. The Adhocracy also uses matrix structure (Mintzberg 1980), but its teams form and dissolve under no stable line, where the matrix keeps two standing lines.

The matrix beside its nearest neighbor, the hybrid

Select a number to fill what it marks in both drawings; the list under them gives the dimensions on which the two profiles differ.

Where a single line gives every unit a single boss, it puts the same people under two lines of authority at once, a function and a project, product, or region, each with real say over their work.

Matrix organization

Reports toAdjust to each other directlyMMutual adjustment2Two superiors

Hybrid structure

Reports toAdjust to each other directlyWStandardization of work processesOStandardization of outputs

The matrix's specialist answers to two bosses at once; every unit in the hybrid answers to one.
The matrix's function, holding the technical standard, and project, setting deadlines, cover the same specialist; the hybrid's bases cover different parts, central functions beside divisions running their own business.
The matrix leans on its two lines adjusting to each other directly, where the hybrid's departments coordinate by standardized work and its divisions by standardized outputs.

Differs on 7 of 11 dimensions

Basis of grouping Two or more at once; Hybrid structure: Function or process and product or purpose
Specialization Medium to high; Hybrid structure: Medium
Decision rights and the locus of authority Limited vertical decentralization to selective vertical and horizontal decentralization; Hybrid structure: Limited horizontal decentralization to limited vertical decentralization
Prime coordinating mechanism Mutual adjustment; Hybrid structure: Standardization of work processes and standardization of outputs
Lateral mechanisms Matrix; Hybrid structure: Direct contact and liaison roles
Differentiation against integration High differentiation, heavy integration; Hybrid structure: Moderate differentiation, some integrating roles
Boundary permeability Selective; Hybrid structure: Closed

The matrix lays its two bases of grouping over the same people; the hybrid, which also has two, gives each part of the organization one basis and one boss.

Source: Davis and Lawrence 1977; Galbraith 1971; Daft 1995; Mintzberg 1980, a synthesis of earlier research, offering propositions without a new test.

Moving into or out of it

Galbraith's (1973, 1974) theory, built from cases, has a firm move in from functions through lateral devices, reaching the matrix when those fall short. Burns and Wholey (1993) found the network factors behind adoption did little to explain abandonment.

Galbraith's lateral processes, added one at a time

Start from rules, hierarchy, and goals; each time the means in place is overloaded, add the next process, or step back.

Rules, hierarchy, and goals
Direct contact
Liaison roles
Task forces
Teams
Integrating roles
Managerial linking roles
Matrix

Rules, hierarchy, and goals. Coordination runs through rules and programs, referral up the hierarchy, and goal setting, with no designed lateral links.Lateral mechanisms, level 1 of 5: Hierarchy, rules, and goals only.Direct contact. Managers who share a problem settle it between themselves.Lateral mechanisms, level 2 of 5: Direct contact and liaison roles.Liaison roles. Where contact between two departments is heavy, a liaison role is set up to handle it.Lateral mechanisms, level 2 of 5: Direct contact and liaison roles.Task forces. Temporary groups of representatives from several departments solve shared problems.Lateral mechanisms, level 3 of 5: Task forces and teams.Teams. Recurring problems turn the temporary groups into permanent teams.Lateral mechanisms, level 3 of 5: Task forces and teams.Integrating roles. A product, program, or project manager integrates across departments, at first with expert power only.Lateral mechanisms, level 4 of 5: Integrating roles and managerial linking roles.Managerial linking roles. The integrating manager gains formal power over budgets.Lateral mechanisms, level 4 of 5: Integrating roles and managerial linking roles.Matrix. Two lines of authority are held at once, so some positions report both to a function and to a product or project.Lateral mechanisms, level 5 of 5: Matrix.

His three other strategies, beside the ladder: slack resources, self-contained tasks, and vertical information systems.

Galbraith's order: the processes 'evolve as follows with increases in uncertainty', each taken up when the one before no longer suffices. The figure takes no uncertainty input and shows no cost per step: he gives none, saying only that more lateral relations mean more managerial intensity.

Rules, hierarchy, and goals. Coordination runs through rules and programs, referral up the hierarchy, and goal setting, with no designed lateral links. Level 1 of 5: Hierarchy, rules, and goals only.
Direct contact. Managers who share a problem settle it between themselves. Level 2 of 5: Direct contact and liaison roles.
Liaison roles. Where contact between two departments is heavy, a liaison role is set up to handle it. Level 2 of 5: Direct contact and liaison roles.
Task forces. Temporary groups of representatives from several departments solve shared problems. Level 3 of 5: Task forces and teams.
Teams. Recurring problems turn the temporary groups into permanent teams. Level 3 of 5: Task forces and teams.
Integrating roles. A product, program, or project manager integrates across departments, at first with expert power only. Level 4 of 5: Integrating roles and managerial linking roles.
Managerial linking roles. The integrating manager gains formal power over budgets. Level 4 of 5: Integrating roles and managerial linking roles.
Matrix. Two lines of authority are held at once, so some positions report both to a function and to a product or project. Level 5 of 5: Matrix.

Galbraith gives an order, not thresholds: as uncertainty rises, each lateral process is taken up when the one before no longer suffices, up to the matrix, and the more of them in use, the greater the managerial intensity. Beside the ladder he names three other strategies: slack resources, self-contained tasks, and vertical information systems.

Source: Galbraith 1974; Galbraith 1973.

Worren (2017), one case study, follows a division that entered from geography, not functions: product units added in 2011 gave some regional managers two bosses until 2015, a transitory matrix giving way to a simpler multidimensional structure. Turato, Whiteoak, and Oprescu (2024), one case study, describe a hospital moving to single points of accountability: staff surveyed afterward preferred the profession-based structure, with no performance data reported in the abstract. Hannan and Freeman (1984) hold that inertia rises with age and attempts at reorganization raise death rates; Amburgey, Kelly, and Barnett (1993), a quantitative study of organizational change and failure rates whose abstract names no sample, find that the effects of change on failure vary by type of change and over time. Change & Transformation is where running a reorganization belongs.

The thinking behind it

Jay Galbraith (1971, 1974): the designs between project and functional forms, topped by the matrix.

Stanley Davis and Paul Lawrence (1977, 1978): the multiple command system, and its pathologies.

Erik Larson and David Gobeli (1987, 1988): the three matrix types against project teams.

Lawton Burns and Douglas Wholey (1993): adoption by imitation as well as task.

Richard Burton and Børge Obel (2004): the multi-contingency fit and misfit rules; with Dorthe Døjbak Håkonsson, three untested conditions for a working matrix (2015 article) and a fit table pairing it with turbulence and analyzers (2015 book).

William Egelhoff and Joachim Wolf (Egelhoff 1988; with Wolf 2002, 2013): revising Stopford and Wells, and conflict by matrix type.

Christopher Bartlett and Sumantra Ghoshal (1990): the matrix as a frame of mind.

Maxim Sytch and Franz Wohlgezogen (2026): how thin the evidence is.

Also called

Two-boss structure; multiple command system (Davis and Lawrence 1977); grid structure; matrix management (Larson and Gobeli; Bartlett and Ghoshal); cross-functional structure (Ford and Randolph 1992); and matrix project management (Kieser and Kubicek 1992).

Works cited

Ahrne and Brunsson 2008. Meta-organizations, Edward Elgar, Cheltenham.
Amburgey and colleagues 1993. Resetting the clock: the dynamics of organizational change and failure, Administrative Science Quarterly 38(1), from p. 51; a quantitative study of organizational change and failure rates.
Aoki 1986. Horizontal vs. vertical information structure of the firm, American Economic Review 76(5), 971 to 983; a formal economic model.
Armour and Teece 1978. Organizational structure and economic performance: a test of the multidivisional hypothesis, Bell Journal of Economics 9(1), 106 to 122; a multi-firm statistical study in one industry.
Ashkenas and colleagues 1995. The Boundaryless Organization: Breaking the Chains of Organizational Structure, Jossey-Bass, San Francisco.
Bartlett and Ghoshal 1990. Matrix management: not a structure, a frame of mind, Harvard Business Review (July to August 1990).
Brandstätter and Graf-Vlachy 2026. Strategic leadership and organizational design: a review and research agenda, Journal of Organization Design 15(3), 165 to 210.
Braverman 1974. Labor and monopoly capital: the degradation of work in the twentieth century, Monthly Review Press, New York; the linked publisher's page is the 1998 twenty-fifth anniversary edition.
Burdín 2014. Are worker-managed firms more likely to fail than conventional enterprises? Evidence from Uruguay, ILR Review 67(1), 202 to 238; a survival study of many firms in one country.
Burns and Stalker 1961. The Management of Innovation, Tavistock Publications, London (first edition 1961); the DOI is for the Oxford University Press revised edition of 1994; a field study of twenty firms.
Burns and Wholey 1993. Adoption and abandonment of matrix management programs: effects of organizational characteristics and interorganizational networks, Academy of Management Journal 36(1), 106 to 138; a quantitative study of matrix adoption and abandonment in a group of hospitals.
Burton and colleagues 2002. Return on assets loss from situational and contingency misfits, Management Science 48(11), 1461 to 1485; a quantitative study of 224 Danish small and medium-sized firms.
Burton and colleagues 2015a. How to get the matrix organization to work, Journal of Organization Design 4(3), 37 to 45; a conceptual essay with no new data.
Burton and colleagues 2015b. Organizational Design: A Step-by-Step Approach, Cambridge University Press (third edition; hardback ISBN 978-1-107-09765-0, paperback ISBN 978-1-107-48361-3).
Burton and colleagues 2020a. Organizational design: a step-by-step approach (fourth edition), Cambridge University Press.
Burton and Obel 2004. Strategic organizational diagnosis and design: the dynamics of fit, Kluwer Academic Publishers (third edition).
Carey 1967. The Hawthorne studies: a radical criticism, American Sociological Review 32(3), 403 to 416.
Carney and colleagues 2011. Business group affiliation, performance, context, and strategy: a meta-analysis, Academy of Management Journal 54(3), 437 to 460; a meta-analysis of 141 studies in 28 countries.
Carroll and Hannan 2000. The demography of corporations and industries, Princeton University Press, Princeton, NJ.
Chandler 1962. Strategy and Structure: Chapters in the History of the Industrial Enterprise, MIT Press, Cambridge, MA (xiv and 463 pages); several independent case histories by a historian.
Child 1972a. Organizational structure, environment and performance: the role of strategic choice, Sociology 6(1), 1 to 22.
Clark 1983. The Higher Education System: Academic Organization in Cross-National Perspective, University of California Press, Berkeley (xiii and 315 pages).
Coase 1937. The nature of the firm, Economica 4(16), 386 to 405 (November 1937).
Colfer and Baldwin 2016. The mirroring hypothesis: theory, evidence, and exceptions, Industrial and Corporate Change 25(5), 709 to 738; a review of 142 empirical studies.
Crozier 1963. Le phénomène bureaucratique: essai sur les tendances bureaucratiques des systèmes d'organisation modernes et sur leurs relations en France avec le système social et culturel, Éditions du Seuil, Paris, 414 pages; English translation The Bureaucratic Phenomenon, University of Chicago Press and Tavistock, 1964 (Crozier 1964).
Crozier 1964. The Bureaucratic Phenomenon, University of Chicago Press, Chicago; Tavistock, London; two case studies by a scholar.
Crozier and Friedberg 1977. L'acteur et le système: les contraintes de l'action collective, Éditions du Seuil, Paris (Sociologie politique series); English translation Actors and Systems: The Politics of Collective Action, University of Chicago Press, 1980 (Crozier and Friedberg 1980).
Crozier and Friedberg 1980. Actors and Systems: The Politics of Collective Action, University of Chicago Press, Chicago and London; translated by Arthur Goldhammer from L'acteur et le système (Seuil, 1977; Crozier and Friedberg 1977).
Csaszar 2012. Organizational structure as a determinant of performance: evidence from mutual funds, Strategic Management Journal 33(6), 611 to 632; a large-sample quantitative study of 609 mutual funds.
Cyert and March 1963. A behavioral theory of the firm, Prentice-Hall, Englewood Cliffs, NJ.
Czarniawska and Joerges 1995. Travels of ideas: organizational change as translation, Discussion paper FS II 95-501, Forschungsschwerpunkt Technik, Arbeit, Umwelt, Wissenschaftszentrum Berlin für Sozialforschung (WZB), Berlin (40 pages; a substantially revised version of the authors' essay 'Winds of organizational change' in Research in the Sociology of Organizations, 1994).
Daft 1995. Organization Theory and Design, West Publishing, Minneapolis/St. Paul, fifth edition.
Daft and Lengel 1986. Organizational information requirements, media richness and structural design, Management Science 32(5), 554 to 571.
Davis and Lawrence 1977. Matrix, Addison-Wesley, Reading, MA.
Davis and Lawrence 1978. Problems of matrix organizations, Harvard Business Review (May 1978).
Delbridge 1995. Surviving JIT: control and resistance in a Japanese transplant, Journal of Management Studies 32(6), 803 to 817; a participant-observation study of one plant.
DeSantola and colleagues 2023. External interfaces or internal processes? Market positioning and divergent professionalization paths in young ventures, Organization Science 34(1), 1 to 23; a longitudinal quantitative study of thousands of young ventures.
Devinney and colleagues 2000. The optimal performance of the global firm: formalizing and extending the integration-responsiveness framework, Organization Science 11(6), 674 to 695; a formal model with no new data.
DiMaggio and Powell 1983. The iron cage revisited: institutional isomorphism and collective rationality in organizational fields, American Sociological Review 48(2), from p. 147.
Donaldson 1995. American anti-management theories of organization: a critique of paradigm proliferation, Cambridge University Press.
Donaldson 2001. The contingency theory of organizations, SAGE Publications, Thousand Oaks, CA (Foundations for Organizational Science).
Dosi and colleagues 2003. Problem-solving behaviors, organizational forms, and the complexity of tasks, In Constance E. Helfat (ed.), The SMS Blackwell Handbook of Organizational Capabilities: Emergence, Development, and Change, Blackwell, Malden, MA, chapter 11.
Doty and colleagues 1993. Fit, equifinality, and organizational effectiveness: a test of two configurational theories, Academy of Management Journal 36(6), 1196 to 1250; a quantitative test across a sample of organizations.
Drazin and Van de Ven 1985. Alternative forms of fit in contingency theory, Administrative Science Quarterly 30(4), from p. 514; a quantitative study of 629 work units in two states' employment security agencies.
Drees and Heugens 2013. Synthesizing and extending resource dependence theory: a meta-analysis, Journal of Management 39(6), 1666 to 1698; a meta-analysis of 157 tests.
Edwards 1979. Contested terrain: the transformation of the workplace in the twentieth century, Basic Books.
Egelhoff 1988. Strategy and structure in multinational corporations: a revision of the Stopford and Wells model, Strategic Management Journal 9(1), 1 to 14; a multi-firm study of multinationals.
Egelhoff and colleagues 2013. Designing matrix structures to fit MNC strategy, Global Strategy Journal 3(3), 205 to 226; a survey of 57 German multinationals that use a matrix.
Emery and Thorsrud 1969. Form and Content in Industrial Democracy: Some Experiences from Norway and Other European Countries, Tavistock Publications, London; reissued by Routledge.
Fakhfakh and colleagues 2012. Productivity, capital, and labor in labor-managed and conventional firms: an investigation on French data, ILR Review 65(4), 847 to 879; a study of large representative samples of French firms.
Fayol 1949. General and Industrial Management, Translated by Constance Storrs, with a foreword by L. Urwick. Pitman, London.
Ford and Randolph 1992. Cross-functional structures: a review and integration of matrix organization and project management, Journal of Management 18(2), 267 to 294; a narrative review of published research, with no new data.
Foss 2003. Selective intervention and internal hybrids: interpreting and learning from the rise and decline of the Oticon spaghetti organization, Organization Science 14(3), 331 to 349; one case study.
Fruin 1994. The Japanese Enterprise System: Competitive Strategies and Cooperative Structures, Clarendon Press, Oxford; first published 1992 (ISBN 0198283180, LCCN 91033285); the DOI is the 1994 edition (ISBN 0198288980), in Oxford Scholarship Online from 2003; a historical study of Japan's 200 leading industrial firms by one scholar.
Galbraith 1971. Matrix organization designs: how to combine functional and project forms, Business Horizons 14(1), 29 to 40.
Galbraith 1973. Designing Complex Organizations, Addison-Wesley, Reading, MA.
Galbraith 1974. Organization design: an information processing view, Interfaces 4(3), 28 to 36.
Granovetter 1985. Economic action and social structure: the problem of embeddedness, American Journal of Sociology 91(3), 481 to 510.
Greenwood and colleagues 2010. The organizational design of transnational professional service firms, Organizational Dynamics 39(2), 173 to 183.
Greer and colleagues 2018. Why and when hierarchy impacts team effectiveness: a meta-analytic integration, Journal of Applied Psychology 103(6), 591 to 613; a meta-analysis of 54 studies.
Greiner 1972. Evolution and revolution as organizations grow, Harvard Business Review, July to August 1972; reprinted in Family Business Review 10(4), 397 to 409, 1997; a conceptual model with no systematic test.
Gustavsen 1992. Dialogue and Development: Theory of Communication, Action Research and the Restructuring of Working Life, Van Gorcum, Assen; Swedish Center for Working Life, Stockholm (viii and 127 pages).
Hamilton and Biggart 1988. Market, culture, and authority: a comparative analysis of management and organization in the Far East, American Journal of Sociology 94 (supplement), S52 to S94; a comparison of three economies by scholars.
Hannan and Freeman 1977. The population ecology of organizations, American Journal of Sociology 82(5), 929 to 964.
Hannan and Freeman 1984. Structural inertia and organizational change, American Sociological Review 49(2), 149 to 164.
Hickson and colleagues 1969. Operations technology and organization structure: an empirical reappraisal, Administrative Science Quarterly 14(3), from p. 378; a multi-organization comparative survey.
Hobday 2000. The project-based organisation: an ideal form for managing complex products and systems?, Research Policy 29(7-8), 871 to 893; one case study with a conceptual model.
Ichniowski and colleagues 1997. The effects of human resource management practices on productivity: a study of steel finishing lines, American Economic Review 87(3), 291 to 313; a quantitative study of 36 steel finishing lines in 17 companies.
Jäger and colleagues 2021. Labor in the boardroom, Quarterly Journal of Economics 136(2), 669 to 725; a natural experiment across many German firms.
Jensen and Meckling 1976. Theory of the firm: managerial behavior, agency costs and ownership structure, Journal of Financial Economics 3(4), 305 to 360.
Jirjahn 2011. Ökonomische Wirkungen der Mitbestimmung in Deutschland: ein Update, Schmollers Jahrbuch 131(1), 3 to 57 (in German); a review of many independent econometric studies.
Joyce 1986. Matrix organization: a social experiment, Academy of Management Journal 29(3), 536 to 561; a quasi-experiment in one firm's engineering division, with a control group.
Katz and Allen 1985. Project performance and the locus of influence in the R&D matrix, Academy of Management Journal 28(1), 67 to 87; a survey of 86 matrixed project teams in nine R&D organizations.
Kellogg and colleagues 2020. Algorithms at work: the new contested terrain of control, Academy of Management Annals 14(1), 366 to 410; a review of earlier research through one theoretical lens, with no new data.
Ketchen and colleagues 1997. Organizational configurations and performance: a meta-analysis, Academy of Management Journal 40(1), 223 to 240; a meta-analysis of 40 tests.
Keum and See 2017. The influence of hierarchy on idea generation and selection in the innovation process, Organization Science 28(4), 653 to 669; a field study combined with an experiment.
Khanna and Palepu 2000a. Is group affiliation profitable in emerging markets? An analysis of diversified Indian business groups, The Journal of Finance 55(2), 867 to 891; a statistical study of many firms in one country.
Kieser and Kubicek 1992. Organisation, Walter de Gruyter, Berlin, third edition, fully revised (de Gruyter Lehrbuch).
Knight 1976. Matrix organization: a review, Journal of Management Studies 13(2), 111 to 130.
Kraatz and Zajac 1996. Exploring the limits of the new institutionalism: the causes and consequences of illegitimate organizational change, American Sociological Review 61(5), 812 to 836; a longitudinal study of 631 private liberal arts colleges.
Krabberød 2015. Standing on the shoulders of giants? Exploring consensus on the validity status of Mintzberg's configuration theory after a negative test, SAGE Open 5(4), article 2158244015611185.
Kuprenas 2003. Implementation and performance of a matrix organization structure, International Journal of Project Management 21(1), 51 to 62; one case study.
Larson and Gobeli 1987. Matrix management: contradictions and insights, California Management Review 29(4), 126 to 138; a survey of 500 managers experienced in product development.
Larson and Gobeli 1988. Organizing for product development projects, Journal of Product Innovation Management 5(3), 180 to 190; a comparison of 540 development projects.
Larson and Gobeli 1989. Significance of project management structure on development success, IEEE Transactions on Engineering Management 36(2), 119 to 125; a quantitative study of 546 development projects.
Lawrence and Lorsch 1967a. Organization and Environment: Managing Differentiation and Integration, Division of Research, Graduate School of Business Administration, Harvard University, Boston (xv and 279 pages); a multi-firm comparative field study.
Levinthal 1997. Adaptation on rugged landscapes, Management Science 43(7), 934 to 950.
Levitt and List 2011. Was there really a Hawthorne effect at the Hawthorne plant? An analysis of the original illumination experiments, American Economic Journal: Applied Economics 3(1), 224 to 238; a reanalysis of one plant's original experimental data.
Lincoln and colleagues 1986. Organizational structures in Japanese and U.S. manufacturing, Administrative Science Quarterly 31(3), 338 to 364; a survey of about a hundred plants in two countries.
MacCormack and colleagues 2012. Exploring the duality between product and organizational architectures: a test of the "mirroring" hypothesis, Research Policy 41(8), 1309 to 1324; a matched comparison of five pairs of software products.
March and Simon 1958. Organizations, Wiley, New York (262 pages; with the collaboration of Harold Guetzkow); second edition, Blackwell, 1993.
Maurice and colleagues 1980. Societal differences in organizing manufacturing units: a comparison of France, West Germany, and Great Britain, Organization Studies 1(1), 59 to 86; a comparison of matched factories in three countries.
Mayo 1933. The human problems of an industrial civilization, Macmillan; reissued by Routledge, 2004.
Mertha 2005. China's 'soft' centralization: shifting tiao/kuai authority relations, The China Quarterly 184, 791 to 810; one scholar's study of one national reform, read in its abstract.
Merton 1940. Bureaucratic structure and personality, Social Forces 18(4), 560 to 568; a conceptual essay with no test.
Meyer and Rowan 1977. Institutionalized organizations: formal structure as myth and ceremony, American Journal of Sociology 83(2), 340 to 363.
Midler 1995. "Projectification" of the firm: the Renault case, Scandinavian Journal of Management 11(4), 363 to 375; one longitudinal case study.
Miles and Snow 1978. Organizational strategy, structure, and process, McGraw-Hill, New York, 1978; reissued by Stanford University Press, 2003 (Stanford Business Classics).
Mintzberg 1979. The Structuring of Organizations: A Synthesis of the Research, Prentice-Hall, Englewood Cliffs, NJ.
Mintzberg 1980. Structure in 5's: a synthesis of the research on organization design, Management Science 26(3), 322 to 341; a synthesis of earlier research, offering propositions without a new test.
Mohr 1971. Organizational Technology and Organizational Structure, Administrative Science Quarterly 16(4), 444-459; a quantitative field study of technology and structure.
Nobeoka 1996a. Reorganizing for multi-project management: Toyota's new structure of product development centers, MIT Japan Program Working Paper MITJP 96-11, Center for International Studies, MIT (reprinted from the working paper series of the MIT International Motor Vehicle Program); an earlier version is Kobe University RIEB Discussion Paper 48 (1995, revised August 1995); one case study built on interviews.
Nobeoka 1996b. マルチプロジェクト戦略: ポストリーンの製品開発マネジメント (Maruchi purojekuto senryaku: posuto rīn no seihin kaihatsu manejimento; Multi-project strategy: post-lean product development management), Yuhikaku (有斐閣), Tokyo, October 1996, 221 pages, ISBN 4-641-06792-9. No English translation was found. Cusumano and Nobeoka's English-language book Thinking Beyond Lean (Free Press, 1998) is a related work, not a translation as far as could be checked; it was not read and is not recorded..
Nonaka and Takeuchi 1995. The Knowledge-Creating Company: How Japanese Companies Create the Dynamics of Innovation, Oxford University Press, New York; case studies of several Japanese firms by the theory's authors.
Ocasio 1997. Towards an attention-based view of the firm, Strategic Management Journal 18(S1) (summer special issue), 187 to 206.
Ouchi 1980. Markets, bureaucracies, and clans, Administrative Science Quarterly 25(1), 129 to 141; a conceptual article.
Palmer and colleagues 1993. Late adoption of the multidivisional form by large U.S. corporations: institutional, political, and economic accounts, Administrative Science Quarterly 38(1), 100 to 131; a multi-firm statistical study of adoption.
Patanakul and colleagues 2012. Autonomous teams and new product development, Journal of Product Innovation Management 29(5), 734 to 750; a survey of several hundred projects.
Pennings 1975. The relevance of the structural-contingency model for organizational effectiveness, Administrative Science Quarterly 20(3), from p. 393 (September 1975); a quantitative field test of the structural-contingency model.
Pfeffer and Salancik 1978. The external control of organizations: a resource dependence perspective, Harper and Row; reissued by Stanford University Press, 2003.
Podolny and Page 1998. Network forms of organization, Annual Review of Sociology 24, 57 to 76.
Powell 1990. Neither market nor hierarchy: network forms of organization, Research in Organizational Behavior 12, 295 to 336 (JAI Press).
Puranam 2021. Human-AI collaborative decision-making as an organization design problem, Journal of Organization Design 10(2), 75 to 80.
Puranam and colleagues 2014. What's "new" about new forms of organizing?, Academy of Management Review 39(2), 162 to 180; a conceptual argument.
Qiu and Donaldson 2012. Stopford and Wells were right! MNC matrix structures do fit a 'high-high' strategy, Management International Review 52(5), 671 to 689; a preliminary test on a sample of German multinationals.
Reitzig and Maciejovsky 2015. Corporate hierarchy and vertical information flow inside the firm: a behavioral view, Strategic Management Journal 36(13), 1979 to 1999; a field study combined with experiments.
Rumelt 1974. Strategy, Structure, and Economic Performance, Division of Research, Harvard Business School, Boston; a multi-firm study of large corporations.
Schoonhoven 1981. Problems with contingency theory: testing assumptions hidden within the language of contingency 'theory', Administrative Science Quarterly 26(3), 349 to 377; a quantitative study of hospital operating room suites.
Siggelkow and Levinthal 2003. Temporarily divide to conquer: centralized, decentralized, and reintegrated organizational approaches to exploration and adaptation, Organization Science 14(6), 650 to 669; a computer simulation, with no field data.
Simon 1946. The proverbs of administration, Public Administration Review 6(1), from page 53.
Simon 1962. The architecture of complexity, Proceedings of the American Philosophical Society 106(6), 467 to 482 (12 December 1962).
Smith 1967. Field Administration: An Aspect of Decentralisation, Routledge and Kegan Paul, London; Humanities Press, New York (Library of Political Studies; viii and 135 pages).
Stewart 2006. A meta-analytic review of relationships between team design features and team performance, Journal of Management 32(1), 29 to 55; a meta-analysis of 93 studies.
Stopford and Wells 1972. Managing the Multinational Enterprise: Organization of the Firm and Ownership of the Subsidiaries, Basic Books, New York; a study of 187 U.S. multinational enterprises.
Streeck 1984a. Industrial Relations in West Germany: A Case Study of the Car Industry, Heinemann, London; St. Martin's Press, New York (178 pages).
Sy and Côté 2004. Emotional intelligence: a key ability to succeed in the matrix organization, Journal of Management Development 23(5), 437 to 455.
Sytch and colleagues 2018. Collaborative by design? How matrix organizations see/do alliances, Organization Science 29(6), 1130 to 1148; a quantitative study of many firms' alliances, with interviews.
Sytch and Wohlgezogen 2026. Solving the integration challenge: a review of research and best practices for high-performance matrix organisations, Publications Office of the European Union, Luxembourg, JRC145912 (an external study report for the European Commission's Joint Research Centre; Laurine Heris and Paul Smits, editors; ISBN 978-92-68-39416-8); a narrative review of the research.
Thompson 1967. Organizations in action: social science bases of administrative theory, McGraw-Hill; reissued by Routledge, 2017.
Tolbert and Zucker 1983. Institutional sources of change in the formal structure of organizations: the diffusion of civil service reform, 1880 to 1935, Administrative Science Quarterly 28(1), from p. 22.
Trist and Bamforth 1951. Some social and psychological consequences of the longwall method of coal-getting, Human Relations 4(1), 3 to 38; an interview-based field study in one industry.
Turato and colleagues 2024. The insights of allied health professionals transitioning from a matrix structure to a centralized profession-based structure within a public hospital setting, Journal of Organization Design 13(4), 197 to 212; one case study.
Tushman and Nadler 1978. Information processing as an integrating concept in organizational design, Academy of Management Review 3(3), from p. 613.
Volberda and colleagues 2012. Contingency Fit, Institutional Fit, and Firm Performance: A Metafit Approach to Organization-Environment Relationships, Organization Science 23(4), 1040-1054; a survey of 3,259 respondents in 1,904 companies.
Wagner 1994. Participation's effects on performance and satisfaction: a reconsideration of research evidence, Academy of Management Review 19(2), 312 to 330; a meta-analytic reanalysis that compares eleven reviews.
Wagner 2021. Dimensional research on organization structure: meta-analysis and conceptual redirection, Journal of Management & Organization 31(4), 1854 to 1871 (published online December 2021; issue of July 2025); a meta-analysis of 346 correlations from 155 published articles.
Ward 1958. The firm in Illyria: market syndicalism, American Economic Review 48(4) (1958); reprinted in David F. Heathfield (ed.), The Economics of Co-Determination, Palgrave Macmillan, 1977, 1 to 25.
Weber 1978. Economy and Society: An Outline of Interpretive Sociology, Edited by Guenther Roth and Claus Wittich. University of California Press, Berkeley.
Whitley 1999. Divergent capitalisms: the social structuring and change of business systems, Oxford University Press.
Wolf and Egelhoff 2002. A reexamination and extension of international strategy-structure theory, Strategic Management Journal 23(2), 181 to 189; a statistical study of 95 German firms.
Wolf and Egelhoff 2013. An empirical evaluation of conflict in MNC matrix structure firms, International Business Review 22(3), 591 to 601; a survey of 82 German multinationals.
Wood and colleagues 2019. Good gig, bad gig: autonomy and algorithmic control in the global gig economy, Work, Employment and Society 33(1), 56 to 75; interviews in six countries and a survey of 679 workers.
Woodward 1965. Industrial organization: theory and practice, Oxford University Press, London and New York; a survey of a hundred manufacturing firms with case studies.
Worren 2017. The matrix as a transitory form: the evolution of FMC Technologies 2001 to 2016, Journal of Organization Design 6(1), article 13; one case study.
Yoshino 1968. Japan's Managerial System: Tradition and Innovation, MIT Press, Cambridge, MA.
Zhou and Sui 2025. What can the research on Chinese bureaucracy do for organization theory?, Management and Organization Review 21(1), 3 to 20.

On this page, “read” means read by this series: where a line says a work, or part of one, was not read, the series did not read it.