Algorithmically managed organization

Software allocates tasks, sets the pace, and evaluates performance, taking over much of what supervisors and the technostructure did.

What it is

An algorithmically managed organization hands much of what supervisors and the technostructure did to software. In Kellogg, Valentine, and Christin's (2020) review of earlier research through one theoretical lens, with no new data, algorithms direct workers (restricting and recommending), evaluate them (recording and rating), and discipline them (replacing and rewarding).

Algorithmic management hands much of the supervisor's work to software

Select a unit or a letter to read it; the key above the drawing names the lines and letters used. Centralization and decision rights are contested, and the drawing shows their first view, power at the top; the profile below gives the other.

Reports toWStandardization of work processesOStandardization of outputs

Select a unit or a letter to read it. W, standardization of work processes: The content of the work is specified or programmed in advance, usually by analysts. O, standardization of outputs: The results of the work are specified, such as targets or product dimensions, and the method is left to the unit.Management. Layer: Management. Reports to no one. Formal power stays at the top, our judgment after Mintzberg's (1980) Type B. Employers use algorithms to direct, evaluate, and discipline workers (Kellogg, Valentine, and Christin 2020, a review).Algorithm. Layer: Software. Reports to Management. Directs workers by restricting and recommending, evaluates them by recording and rating, and disciplines them by replacing and rewarding (Kellogg and colleagues 2020). It takes over much of what supervisors did: in Lee and colleagues (2015), software assigned, optimized, and evaluated drivers' work. Those who design it gain the power Mintzberg's analysts hold, our judgment.Workers. Layer: Workers. Reports to Algorithm. By function or by product. With no grouping basis of its own, the form puts its layer over the structure it finds; this unit is grouped by function, one of the two bases recorded as judgment. On online labor platforms much of the work is done by people outside the firm's employment, connected through the platform.Workers. Layer: Workers. Reports to Algorithm. By function or by product. A unit grouped by product, the other basis recorded as judgment. Employers who direct their own staff by algorithm keep the boundary they had.W, standardization of work processes. The content of the work is specified or programmed in advance, usually by analysts.O, standardization of outputs. The results of the work are specified, such as targets or product dimensions, and the method is left to the unit.

The structure in words

Management
- Management, reports to no one. Formal power stays at the top, our judgment after Mintzberg's (1980) Type B. Employers use algorithms to direct, evaluate, and discipline workers (Kellogg, Valentine, and Christin 2020, a review).
Software
- Algorithm, reports to Management, coordinated by standardization of work processes and standardization of outputs. Directs workers by restricting and recommending, evaluates them by recording and rating, and disciplines them by replacing and rewarding (Kellogg and colleagues 2020). It takes over much of what supervisors did: in Lee and colleagues (2015), software assigned, optimized, and evaluated drivers' work. Those who design it gain the power Mintzberg's analysts hold, our judgment.
Workers
- Workers, reports to Algorithm, grouped by function or by product. With no grouping basis of its own, the form puts its layer over the structure it finds; this unit is grouped by function, one of the two bases recorded as judgment. On online labor platforms much of the work is done by people outside the firm's employment, connected through the platform.
- Workers, reports to Algorithm, grouped by function or by product. A unit grouped by product, the other basis recorded as judgment. Employers who direct their own staff by algorithm keep the boundary they had.

Software directs workers, evaluates them, and disciplines them, standardizing both their work and its outputs, over whatever grouping the organization already has. Formal power stays at the top, and in our judgment those who design the algorithms gain the power Mintzberg's analysts hold.

Source: Kellogg and colleagues 2020, a review of earlier research through one theoretical lens, with no new data; Lee and colleagues 2015b, a qualitative study of drivers at two ride-sharing services; Mintzberg 1980, a synthesis of earlier research, offering propositions without a new test.

Its profile, our judgment unless marked:

Basis of grouping: spans function and product; it lays its layer over the structure it finds.
Specialization: medium.
Formalization: high, as theory (Kellogg and colleagues 2020).
Centralization: contested. Centralized, as theory (Kellogg and colleagues 2020); or mixed to decentralized, since on online labor platforms it tended to give workers high flexibility and autonomy (Wood and colleagues 2019, interviews in six countries and a survey of 679 workers).
Layers and span: flat; software does much of the supervisor's assigning and evaluating (Lee and colleagues 2015, a qualitative study of drivers at two ride-sharing services).
Decision rights: contested. Limited horizontal decentralization (Type B in Mintzberg 1980, a synthesis of earlier research, offering propositions without a new test): formal power stays at the top while the algorithm's designers gain his analysts' power. Or, our reading, selective to full decentralization: platform workers' autonomy (Wood and colleagues) and ride-hailing's frequent, narrow choices (Cameron 2024, a seven-year qualitative study of one industry's workers).
Coordinating mechanism: spans standardization of work processes and of outputs, as theory.
Lateral mechanisms: hierarchy, rules, and goals only.
Differentiation against integration: low; the platform, not integrators, joins the workers.
Mode of exchange: spans market and bureaucracy. Uber drivers responded in market-like and organization-like ways (Möhlmann and colleagues 2021, one case study of a single platform). Vallas and Schor (2020), a review essay, dispute the scale: platforms are a governance mechanism different from markets, hierarchies, or networks.
Boundary permeability: spans selective and permeable; much platform work is done by people outside the firm's employment.

Where the algorithmically managed 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: Function or process and product or purpose. Basis: Judgment. The form sets no grouping basis of its own and lays its defining layer over whatever structure it finds; the two bases of a single-line hierarchy are recorded as judgment so a reader's profile can still match. Sources: Kellogg and colleagues 2020.Specialization: Medium. Basis: Judgment. Workers on online platforms reported task variety and complexity, which points away from narrow jobs, but nothing read settles the band; judgment. Sources: Wood and colleagues 2019.Standardization and formalization: High. Basis: Theory. Algorithms restrict and recommend what workers do and record and rate how they do it, a rational control distinct from earlier forms. Sources: Kellogg and colleagues 2020.Centralization: Centralized. Basis: Theory. Employers use algorithms to direct, evaluate, and discipline workers. Sources: Kellogg and colleagues 2020. View 2: Mixed to decentralized. Algorithmic management on online labor platforms tended to give workers high flexibility and autonomy, alongside low pay, isolation, and overwork. Sources: Wood and colleagues 2019.Layers and span of control: Flat. Basis: Judgment. Software takes over much of what supervisors did, assigning, optimizing, and evaluating work; judgment for the band. Sources: Kellogg and colleagues 2020; Lee and colleagues 2015b.Decision rights and the locus of authority: Limited horizontal decentralization. Basis: Judgment. Formal power stays at the top while those who design the algorithms gain the power Mintzberg's analysts hold in Type B; judgment. Sources: Kellogg and colleagues 2020; Mintzberg 1980; Wood and colleagues 2019; Cameron 2024. View 2: Selective vertical and horizontal decentralization to vertical and horizontal decentralization. On online labor platforms algorithmic management tended to give workers high flexibility and autonomy, and in ride-hailing it split the work into many interactions that allow frequent, narrow choices; placing that on the decentralization types is judgment. Sources: Wood and colleagues 2019; Cameron 2024.Prime coordinating mechanism: Standardization of work processes and standardization of outputs. Basis: Theory. Standardization of work and outputs by algorithm, with real-time direction and evaluation. Sources: Kellogg and colleagues 2020; Lee and colleagues 2015b.Lateral mechanisms: Hierarchy, rules, and goals only. Basis: Judgment. Coordination runs through the system rather than designed links among workers; judgment. Sources: Kellogg and colleagues 2020.Differentiation against integration: Low differentiation, light integration. Basis: Judgment. Workers are joined by the platform rather than by integrators; judgment. Sources: Kellogg and colleagues 2020; Lawrence and Lorsch 1967a.Mode of exchange and control: Market and bureaucracy. Basis: Judgment. Platforms use algorithms for matching and control, and Uber drivers met them with both market-like and organization-like responses; employers who direct their own staff by algorithm sit at bureaucracy; the span for the form is judgment. Sources: Möhlmann and colleagues 2021; Kellogg and colleagues 2020.Boundary permeability: Selective to permeable. Basis: Judgment. On online labor platforms much of the work is done by people outside the firm's employment, connected through the platform, while employers who use algorithms to direct their own staff keep the boundary they had; judgment for the band. Sources: Kellogg and colleagues 2020; Möhlmann and colleagues 2021.

Three levels rest on theory, each from Kellogg, Valentine, and Christin's (2020) review: high formalization, centralized decisions, and standardized work and outputs. The other eight are judgment. Two rows are contested, centralization and decision rights: a second view, drawn from studies of platform workers (Wood and colleagues 2019; Cameron 2024), places them toward decentralization, and for decision rights that placement is our reading.

Source: Kellogg and colleagues 2020, a review of earlier research through one theoretical lens, with no new data; Wood and colleagues 2019, interviews in six countries and a survey of 679 workers; Lee and colleagues 2015b, a qualitative study of drivers at two ride-sharing services; Mintzberg 1980, a synthesis of earlier research, offering propositions without a new test; Cameron 2024, a seven-year qualitative study of one industry's workers; Lawrence and Lorsch 1967a, a multi-firm comparative field study; Möhlmann and colleagues 2021, one case study of a single platform.

Where it came from

Min Kyung Lee and colleagues (2015), in an early study of Uber and Lyft drivers, called their subject algorithmic, data-driven management: software assigned, optimized, and evaluated the work.

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: Labor process and critical theory; the algorithmic-management literature.

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.

The problem: how to direct, evaluate, and discipline work when software, not managers, does it, and what this does to control and contest. Sector of origin: ride-hailing, then online labor platforms and other digital services.

The antecedents are older: Taylor's (1911) separation of planning from doing and Mintzberg's (1980) machine bureaucracy. Kellogg and colleagues (2020) argue that algorithms facilitate a rational control distinct from the technical and bureaucratic control Edwards (1979) traced. Noponen and colleagues (2024), a qualitative systematic review of 172 published articles, many of them case studies, with no pooled analysis, find that it creates a complex digital version of Taylorism. Steinberg (2022), a historical and conceptual argument, traces the platform concept and data-driven production practice to the Toyota Production System, and cautions against assuming that new technologies require new organizational forms.

Where it fits

For this form, no study found tests fit to a contingency factor. Eleven cells are our judgment; one, knowledge intensity, rests on studies that found the form operating; technology is contested. The strategy premise is Burton and Obel (2004), the most developed scholarly fit model.

The algorithmically managed organization against the twelve contingency factors

Each factor at its three levels, Fits, Partial, or Strain; select a row for its reason, sources, and grade.

Select a factor to read its reason, sources, basis, and grade.Size. Software that allocates, paces, and rates work pays where there are many similar tasks and workers; a small organization has too few for it to learn from or justify. Basis: Judgment. Sources: Kellogg and colleagues 2020.Age. Young platform firms are built around the algorithm; in established organizations incumbents may comply only in appearance, though in one case the way the organization knew its customers changed anyway. Basis: Judgment. Sources: Pachidi and colleagues 2021; Kellogg and colleagues 2020.Environmental uncertainty (unpredictability, dynamism). Software can reassign routine tasks in real time as demand shifts, but its rules come from past patterns, so unforeseen change is where its direction fails. Basis: Judgment. Sources: Kellogg and colleagues 2020; Valentine and colleagues 2017; Lebovitz and colleagues 2022; Pennings 1975.Environmental complexity. Software that directs and rates work is easiest to build where few, similar factors decide the task; complex work has been structured by algorithm, but only in a few cases. Basis: Judgment. Sources: Mintzberg 1980; Valentine and colleagues 2017; Pennings 1975.Munificence and hostility. Software control can cut supervisory cost under pressure, but nothing read ties algorithmic management to munificence. Basis: Judgment. Sources: Kellogg and colleagues 2020.Diversity of products and markets. Software can direct and rate work in any business mix; nothing read ties it to diversity. Basis: Judgment. Sources: Kellogg and colleagues 2020.Strategy type. Software that allocates, paces, and rates work standardizes and measures it, which serves efficiency in a known domain; its high formalization and central control are the opposite of the low formalization and centralization Burton and Obel give a prospector. Basis: Judgment. Sources: Kellogg and colleagues 2020; Burton and Obel 2004.Technology and routineness. Algorithms can direct and rate work whose steps, timing, and measures can be specified, as systematized programs for repetitive tasks are; with no known procedure there is nothing for the software to specify. Basis: Judgment. Sources: Kellogg and colleagues 2020; Van de Ven and colleagues 2013; Perrow 1967; Wood and colleagues 2019; Hickson and colleagues 1969; Mohr 1971. View 1: Routine fits, Mixed partial, Nonroutine strain. Algorithms can direct and rate work whose steps, timing, and measures can be specified, as systematized programs for repetitive tasks are; with no known procedure there is nothing for the software to specify. Sources: Kellogg and colleagues 2020; Van de Ven and colleagues 2013; Perrow 1967. View 2: Routine fits, Mixed fits, Nonroutine partial. Across varying countries and types of remote work, algorithmic control was central to how the platforms operated, and it tended to give workers high task variety and complexity, so the form runs more than repetitive work. Sources: Wood and colleagues 2019.Interdependence. Software coordinates workers who each draw on the platform and can sequence handoffs, but reciprocal adjustment among workers is not something it allocates or rates. Basis: Judgment. Sources: Kellogg and colleagues 2020; Thompson 1967.Knowledge intensity and professionalization. Software that assigns, paces, and rates work ran ride-hailing drivers' work and a wide variety of remote digital services, including varied and complex tasks; professionals given opaque algorithmic outputs often did not take them up. Basis: Evidence. Grade: a multi-firm empirical study or several independent case studies by scholars. Sources: Kellogg and colleagues 2020; Lee and colleagues 2015b; Wood and colleagues 2019; Cameron 2024; Lebovitz and colleagues 2022.Global integration pressure. Platforms direct workers in many countries at once, but no work read sets algorithmic management against integration pressure across countries. Basis: Judgment. Sources: Kellogg and colleagues 2020; Wood and colleagues 2019.Local responsiveness pressure. No work read sets algorithmic management against responsiveness to countries, so this factor gives no reason for or against. Basis: Judgment. Sources: Kellogg and colleagues 2020.

Counting the one contested cell, technology, by its first view, the 36 settings read Fits at 13, Partial at 20, and Strain at 3, the strains at a prospector strategy, nonroutine technology, and reciprocal interdependence. Eleven of the twelve cells rest on judgment.

Source: Kellogg and colleagues 2020, a review of earlier research through one theoretical lens, with no new data; Pachidi and colleagues 2021, one field study of a single organization; Valentine and colleagues 2017, a system deployment reported by its designers; Lebovitz and colleagues 2022, one field study in a single hospital; Pennings 1975, a quantitative field test of the structural-contingency model; Mintzberg 1980, a synthesis of earlier research, offering propositions without a new test; Burton and Obel 2004; Van de Ven and colleagues 2013; Perrow 1967; Wood and colleagues 2019, interviews in six countries and a survey of 679 workers; Hickson and colleagues 1969, a multi-organization comparative survey; Mohr 1971, a quantitative field study of technology and structure; Thompson 1967; Lee and colleagues 2015b, a qualitative study of drivers at two ride-sharing services; Cameron 2024, a seven-year qualitative study of one industry's workers.

Your own situation against the algorithmically managed 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

Flagged (its sources disagree): A prospector bound by heavy rules. Disputed: Burton and Obel 2004; against it, Sine and colleagues 2006; Adler and Borys 1996.
Flagged (its sources disagree): A prospector that decides everything at the top. Disputed: Burton and Obel 2004; against it, Burton and Obel 2004.
Flagged (its sources disagree): A defender with authority spread wide. Disputed: Burton and Obel 2004; against it, Mintzberg 1980.
Flagged (it meets the form only through part of its profile): An analyzer with authority spread wide. Source: Burton and Obel 2004.
Flagged (its sources disagree): A large organization that decides everything at the top. Disputed: Burton and Obel 2004; against it, Mintzberg 1980.
Flagged (its sources disagree): A small organization with a large one's rules. Disputed: Burton and Obel 2004; against it, Sine and colleagues 2006.
Flagged (its sources disagree): Nonroutine work bound by procedures. Disputed: Burton and Obel 2004; against it, Adler and colleagues 1999; Adler and Borys 1996; against it, Hickson and colleagues 1969; Child and Mansfield 1972.
Flagged (its sources disagree): Nonroutine work decided at the top. Disputed: Burton and Obel 2004; against it, Hickson and colleagues 1969; Child and Mansfield 1972; Mohr 1971.
Flagged (its sources disagree): Heavy rules in an unpredictable environment. Disputed: Burton and Obel 2004; against it, Sine and colleagues 2006; Bigley and Roberts 2001; Adler and Borys 1996.
Flagged (it meets the form only through part of its profile): Central decisions in a turbulent environment. Source: Burton and Obel 2004.
Flagged (its sources disagree): Reactor: a prospector's strategy on a defender's structure. Disputed: Miles and Snow 1978; Burton and Obel 2004; against it, Burton and colleagues 2015b; Snow and Hrebiniak 1980.
Flagged (its sources disagree): Reactor: a defender's structure kept after the environment changed. Disputed: Miles and Snow 1978; Burton and Obel 2004; against it, Sine and colleagues 2006; Bigley and Roberts 2001; Adler and Borys 1996.
Flagged (it meets the form only through part of its profile): One way of working for very different parts of the environment. Source: Lawrence and Lorsch 1967a; Lawrence and Lorsch 1967b.
Flagged (its sources disagree): The hierarchy alone carries an unpredictable environment. Disputed: Galbraith 1974; Galbraith 1973; against it, Burton and Obel 2004; Mintzberg 1980.
Flagged (its sources disagree): The hierarchy alone carries nonroutine work. Disputed: Galbraith 1974; Galbraith 1973; against it, Schoonhoven 1981.
Flagged (it meets the form only through part of its profile): The hierarchy alone carries reciprocal work. Source: Galbraith 1973; Galbraith 1974; Thompson 1967.
Flagged (it meets the form only through part of its profile): Reciprocal work coordinated by standards. Source: Thompson 1967; Van de Ven and colleagues 1976.

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 prospector bound by heavy rules fires when Strategy type at Prospector (exploration): flagged (its sources disagree). Disputed: Burton and Obel 2004; against it, Sine and colleagues 2006; Adler and Borys 1996.
A prospector that decides everything at the top fires when Strategy type at Prospector (exploration): flagged (its sources disagree). Disputed: Burton and Obel 2004; against it, Burton and Obel 2004.
A defender with authority spread wide fires when Strategy type at Defender (exploitation): flagged (its sources disagree). Disputed: Burton and Obel 2004; against it, Mintzberg 1980.
An analyzer with authority spread wide fires when Strategy type at Analyzer (exploitation and exploration): flagged (it meets the form only through part of its profile). Source: Burton and Obel 2004.
A large organization that decides everything at the top fires when Size at Large: flagged (its sources disagree). Disputed: Burton and Obel 2004; against it, Mintzberg 1980.
A small organization with a large one's rules fires when Size at Small: flagged (its sources disagree). Disputed: Burton and Obel 2004; against it, Sine and colleagues 2006.
Nonroutine work bound by procedures fires when Technology and routineness at Nonroutine: flagged (its sources disagree). Disputed: Burton and Obel 2004; against it, Adler and colleagues 1999; Adler and Borys 1996; against it, Hickson and colleagues 1969; Child and Mansfield 1972.
Nonroutine work decided at the top fires when Technology and routineness at Nonroutine: flagged (its sources disagree). Disputed: Burton and Obel 2004; against it, Hickson and colleagues 1969; Child and Mansfield 1972; Mohr 1971.
Heavy rules in an unpredictable environment fires when Environmental uncertainty (unpredictability, dynamism) at Unpredictable: flagged (its sources disagree). Disputed: Burton and Obel 2004; against it, Sine and colleagues 2006; Bigley and Roberts 2001; Adler and Borys 1996.
Central decisions in a turbulent environment fires when Environmental uncertainty (unpredictability, dynamism) at Unpredictable and Environmental complexity at Complex: flagged (it meets the form only through part of its profile). Source: Burton and Obel 2004.
Reactor: a prospector's strategy on a defender's structure fires when Strategy type at Prospector (exploration): flagged (its sources disagree). Disputed: Miles and Snow 1978; Burton and Obel 2004; against it, Burton and colleagues 2015b; Snow and Hrebiniak 1980.
Reactor: a defender's structure kept after the environment changed fires when Strategy type at Defender (exploitation) and Environmental uncertainty (unpredictability, dynamism) at Unpredictable: flagged (its sources disagree). Disputed: Miles and Snow 1978; Burton and Obel 2004; against it, Sine and colleagues 2006; Bigley and Roberts 2001; Adler and Borys 1996.
One way of working for very different parts of the environment fires when Environmental complexity at Complex: flagged (it meets the form only through part of its profile). Source: Lawrence and Lorsch 1967a; Lawrence and Lorsch 1967b.
The hierarchy alone carries an unpredictable environment fires when Environmental uncertainty (unpredictability, dynamism) at Unpredictable: flagged (its sources disagree). Disputed: Galbraith 1974; Galbraith 1973; against it, Burton and Obel 2004; Mintzberg 1980.
The hierarchy alone carries nonroutine work fires when Technology and routineness at Nonroutine: flagged (its sources disagree). Disputed: Galbraith 1974; Galbraith 1973; against it, Schoonhoven 1981.
The hierarchy alone carries reciprocal work fires when Interdependence at Reciprocal: flagged (it meets the form only through part of its profile). Source: Galbraith 1973; Galbraith 1974; Thompson 1967.
Reciprocal work coordinated by standards fires when Interdependence at Reciprocal: flagged (it meets the form only through part of its profile). Source: Thompson 1967; Van de Ven and colleagues 1976.

Seventeen of the 39 contingency misfit rules can fire for this form, and each would only flag it. The five that meet its high formalization in full are disputed, and the other twelve meet only one end of a span or one view of its contested centralization.

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

Size. Partial small, fits medium and large: the software pays where many similar tasks give it something to learn from (premise: Kellogg and colleagues 2020).
Age. Fits young and established, partial old, our reading: platform firms are built around the algorithm, while incumbents may comply only in appearance, though in Pachidi and colleagues (2021), one field study of a single organization, knowing changed anyway.
Environmental uncertainty. Fits stable and shifting, partial unpredictable: its rules come from past patterns, though flash organizations, reconfigured continuously (Valentine and colleagues 2017, a system deployment reported by its designers), keep unpredictable at partial.
Environmental complexity. Fits simple, partial moderate and complex, on Mintzberg's (1980) premise that standardization goes with simple environments; flash organizations keep complex at partial. Limit on both environment bullets: Pennings (1975), a quantitative field test of the structural-contingency model, found environment-structure fit failed to explain effectiveness, a null Donaldson (2001) lays to method.
Munificence and hostility; diversity; global integration and local responsiveness. Partial throughout: no work read ties the form to them.
Strategy. Fits defender, partial analyzer, strain prospector: measured, standardized work serves efficiency in a known domain, and high formalization and central control oppose what Burton and Obel give a prospector.
Technology. Contested. Fits routine, partial mixed, strain nonroutine: software directs only work whose steps, timing, and measures can be specified (Kellogg and colleagues 2020; Van de Ven, Ganco, and Hinings 2013; Perrow 1967). Or fits routine and mixed, partial nonroutine: algorithmic control ran remote work of high variety and complexity (Wood and colleagues 2019); placing that at mixed is our reading, since variety and complexity are not Perrow's unanalyzable exceptions. Limits on both: Hickson, Pugh, and Pheysey (1969), a multi-organization comparative survey, found the technological imperative generally not supported, and Mohr (1971), a quantitative field study of technology and structure, no support for effectiveness depending on their match, a null Donaldson (2001) also lays to method, citing later positive tests of job unpredictability.
Interdependence. Fits pooled and sequential, strain reciprocal, on Thompson's (1967) propositions: the software can sequence handoffs but does not allocate or rate reciprocal adjustment.
Knowledge intensity. Evidence: fits low and moderate, partial high. Lee and colleagues (2015) and Cameron (2024) found it running ride-hailing work, learned on the job, and Wood and colleagues across many digital services, moderate in our reading. Lebovitz and colleagues' radiologists (one field study in a single hospital) did not take up opaque AI results in two of three departments, but that is AI for professional judgment, so high is partial.

What it does well

It holds a dispersed workforce to a standard. Uber's app held its drivers to a relatively standardized passenger experience (Rosenblat and Stark 2016, the case below).

Wood and colleagues (2019) found it tends to give remote gig workers high flexibility, autonomy, task variety, and complexity.

Noponen and colleagues (2024) found that it at once restrains and enables autonomy, while dependence on platform income pushes workers into inflexible practices. Companies have mostly chosen to use it to control; enabling uses are rare in the literature, which its search terms may undercount.

What it costs

The mechanisms that freed Wood and colleagues' gig workers can also bring low pay, social isolation, irregular hours, overwork, sleep deprivation, and exhaustion.

Mariscal de Gante Martin and Villani (2026), a working paper surveying employees in the 27 European Union states, report the form associated with less autonomy, fewer breaks, and more stress; direction of tasks and pace shows the strongest links to reduced discretion and work intensification. The links are cross-sectional associations, not effects, the authors' weighted estimates included.

How it fails

How the algorithmically managed 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.

Automation surprise and mode confusion Covert automation authority Silent monitor failure Trust miscalibration Grounding failure Silent exit

How many ties the failure has. Highest: Workers cannot see the rules that judge themPlaying the algorithm wins consent that hides the problemsOpaque tools are refused or used blindCoercive use of the tools backfiresWorkers with little power walk awayAutomation surprise and mode confusionCovert automation authoritySilent monitor failureTrust miscalibrationGrounding failureSilent exit

Select a failure to read its place in the network.Workers cannot see the rules that judge them. Degree 2Playing the algorithm wins consent that hides the problems. Degree 1Opaque tools are refused or used blind. Degree 1Coercive use of the tools backfires. Degree 1Workers with little power walk away. Degree 1Automation surprise and mode confusion. Degree 1Covert automation authority. Degree 1Silent monitor failure. Degree 1Trust miscalibration. Degree 1Grounding failure. Degree 1Silent exit. Degree 1

Failure	Degree
Workers cannot see the rules that judge them	2
Playing the algorithm wins consent that hides the problems	1
Opaque tools are refused or used blind	1
Coercive use of the tools backfires	1
Workers with little power walk away	1
Automation surprise and mode confusion	1
Covert automation authority	1
Silent monitor failure	1
Trust miscalibration	1
Grounding failure	1
Silent exit	1

Five failure modes name six patterns: workers who cannot see the rules that judge them name two, automation surprise and covert automation authority, and each other mode names one. Each tie is a likeness the entry draws, not a finding of the studies it cites.

Source: Rahman 2021, one field study of a single platform; Cameron 2022, a four-year qualitative study of workers in one sector; Cameron 2024, a seven-year qualitative study of one industry's workers; Lebovitz and colleagues 2022, one field study in a single hospital; Anthony 2021, an ethnography of four groups in one bank; Rahman and Valentine 2021, one field study of a single platform; Burbano and colleagues 2026, two field experiments with one platform.

Workers cannot see the rules that judge them. In Rahman (2021), one field study of a single platform, an opaque evaluation algorithm left freelancers unable to align their work with criteria they could not identify; in our reading, automation surprise and covert automation authority.

Playing the algorithm wins consent that hides the problems. Drivers in Cameron (2022, a four-year qualitative study of workers in one sector) turned the work into games. In Cameron (2024) they followed the nudges or manipulated their inputs, and both elicited consent that, the author argues, can mask structurally problematic features. The masking is close, in our reading, to silent monitor failure.

Opaque tools are refused or used blind. Lebovitz and colleagues' radiologists mostly did not take up AI results. In Anthony (2021), an ethnography of four groups in one bank, groups splitting validation between junior and senior roles used their tools routinely but black boxed the algorithms. Both concern professionals' analytical tools, not algorithmic management, and sit close to trust miscalibration.

Coercive use of the tools backfires. In Rahman and Valentine (2021), one field study of a single platform, coercion contributed to uncompleted projects, while repairing misaligned interpretations got them done, a pattern close to grounding failure.

Workers with little power walk away. In Burbano, Koo, and Luo (2026), two field experiments with one platform, workers told of a proposed rating system disengaged completely, which the authors attribute to the power differential: close to silent exit.

On the profile's first views, fifteen misfit rules fire, eleven propositions and four our judgment, none tested singly; Galbraith's three, Thompson's, and Lawrence and Lorsch's fire only at one end of a span: grouping by function, or standardized work processes. Limits: Burton, Lauridsen, and Obel (2002), a quantitative study of 224 Danish small and medium-sized firms, tested the 1998 version of Burton and Obel's rules only as a set, finding lower return on assets with misfits; one factor paired with one structure, as most rules are, is the interaction form of fit, unsupported in a quantitative study of 629 work units in two states' employment security agencies (Drazin and Van de Ven 1985); contingency and institutional fit explained performance together, yet firms in perfect institutional fit slightly lost performance striving for contingency fit (Volberda and colleagues 2012, a survey of 3,259 respondents in 1,904 companies); and Pennings's (1975) null on environment fit, disputed by Donaldson (2001).

Each of the four undisputed rules fires on only part of the profile, so each flags and none sinks: central decisions amid many dissimilar, unpredictable factors (Burton and Obel 2004), on the first centralization view; reciprocal work left to the hierarchy (Galbraith 1973, 1974) or, a judgment, to standards (Thompson 1967); and, another judgment, one way of working for very different parts of the environment, premised on Lawrence and Lorsch (1967), a multi-firm comparative field study of functionally organized firms only; the rest at the span ends above.

The other eleven are disputed, so each flags. Against heavy rules: formalized founding teams that outperformed (Sine and colleagues 2006, a quantitative study of new ventures in one sector), enabling formalization (Adler and Borys 1996, a conceptual argument with no new test), bureaucracies that handled volatile or nonroutine tasks (Bigley and Roberts 2001; Adler and colleagues 1999; each one case study), and, for nonroutine work, the Aston reappraisal (Hickson and colleagues 1969; Child and Mansfield 1972, a multi-firm comparative study). Against central decisions: for nonroutine work, the same reappraisal and Mohr; at large size, Mintzberg's (1980) centralized machine bureaucracy, though Doty and colleagues (1993), a quantitative test across a sample of organizations, did not find that resembling his types predicted effectiveness, a null Krabberød (2015, ungraded here) argues is not shown exhaustive, while Ketchen and colleagues (1997), a meta-analysis of 40 tests pooling many strategy types, found configurations partly explaining performance; and the simple configuration that Burton and Obel's own book gives a prospector. Against Galbraith's overload: Burton and Obel's own simple configuration for simple, uncertain environments (or divisional at low equivocality, ad hoc at high), and Mintzberg's simple structure; for exceptional tasks, Schoonhoven (1981), a quantitative study of hospital operating room suites, where hypotheses drawn from Galbraith went unsupported and more precise interactive ones did better. Against the reactor rules, both our judgment: on a prospector's strategy, Burton, Obel, and Håkonsson's later book (2015, ungraded here) fits a reactor to a calm environment, and Snow and Hrebiniak (1980), a study of top managers' perceptions in four industries, found other types outperforming reactors in competitive industries but not a highly regulated one, on data they call only suggestive, so where competition is weak the pairing need not cost; on a defender's rules kept after the environment turned unpredictable (Miles and Snow 1978), the dispute over heavy rules.

On the second centralization view, the four central-decision rules and the prospector reactor rule drop out; at its decentralized end two Burton and Obel rules fire, against wide authority in an analyzer, undisputed but flagging, since only that end meets it, and in a defender, disputed by Mintzberg's (1980) decentralized professional bureaucracy in complex but stable environments.

What the research found

The evidence is mostly qualitative and set in platform work, and Noponen and colleagues (2024) found it mostly Western, with developers' perspectives understudied.

Kadolkar, Kepes, and Subramony (2025), a topic-model review of the literature across disciplines on gig workers, find it fragmented, with scholars disagreeing on how to conceptualize and measure the form there. Control against autonomy is unsettled: Kellogg and colleagues and Rosenblat and Stark stress control; Wood and colleagues (2019) found autonomy and strain together; Cameron (2024) found workers liking it under a consent that, the author argues, can mask structural problems. Mariscal de Gante Martin and Villani (2026) find the form is not one phenomenon: its links to job quality vary by practice and by country.

None of these studies measures the form's effect on an organization's performance, and no replication failure is recorded. Burbano and colleagues and Mariscal de Gante Martin and Villani (2026) are the most recent, the second not peer reviewed.

The evidence ledger holds 20 studies filed under this form.

In practice: one published case

Uber's drivers are the subject of Rosenblat and Stark (2016), one case study of a single firm's drivers.

Uber managed a large, disaggregated workforce through its ride-hail platform, delivering a relatively standardized experience to passengers while promoting its drivers as entrepreneurs whose work is marked by freedom, flexibility, and independence.

Over nine months the authors analyzed about 1,350 archival items from five online forums of Uber drivers, posted from December 2014 to September 2015, and in-depth interviews with seven drivers, alongside a design critique of the app and a discursive critique of Uber's corporate communications.

The authors found that Uber nonetheless exerted significant indirect control over how drivers did their jobs, through the information and power asymmetries the app produced.

The study concluded that the information and power asymmetries of the application are fundamental to Uber's ability to structure control over its workers; rhetorical appeals to technology and algorithms are used to structure asymmetric relationships to labor; and the role of platform disintermediation in shaping power between employers and workers needs more attention.

The limits are the authors' own: one firm, written from the drivers' side, with forum participants and interviewees who are not necessarily representative of Uber's drivers.

Source: Rosenblat and Stark 2016. Algorithmic labor and information asymmetries: a case study of Uber's drivers, International Journal of Communication 10, 3758 to 3784; one case study of a single firm's drivers.

Variants and near relatives

Platform-mediated gig work (Rosenblat and Stark 2016; Lee and colleagues 2015): independent workers directed through an app's dispatch, pricing, and rating systems. Platform here means the app that directs workers, not Ciborra's platform organization, platform orchestration, the middle platform, Team Topologies' platform team, or the worker-owned platform cooperative.

Human-AI decision structures (Puranam 2021, a short conceptual essay; Raisch and Krakowski 2021, conceptual): designs that divide decisions between people and learning algorithms, treated as an organization design problem; the software shares decisions and does not direct workers.

Flash organizations (Valentine and colleagues 2017): software organizes crowd workers into computationally represented roles, teams, and hierarchies and reconfigures the structure as the work proceeds. A crowd asked only for ideas or solutions, without roles, is not this variant.

Burton, Obel, and Håkonsson's (2020) digital organization, a property, not a form here, has two pieces taken up under this form: agents that are human or intelligent robots, and the algorithmic manager of ride-hailing drivers.

Here software performs direction, evaluation, and discipline continuously and at scale; the machine bureaucracy, its nearest neighbor, relies on analysts' rules applied by human supervisors.

Algorithmic management beside its nearest neighbor, the machine bureaucracy

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

Software performs direction, evaluation, and discipline continuously and at scale, where the machine bureaucracy relies on analysts' rules applied by human supervisors.

Algorithmically managed organization

Reports toWStandardization of work processesOStandardization of outputs

Machine bureaucracy

Reports toAdvises, or reports indirectlyWStandardization of work processes

Software directs, evaluates, and disciplines the workers itself, where the machine bureaucracy's analysts design the rules beside the line and its function managers apply them to the operators.
The algorithm standardizes outputs as well as the work; the machine bureaucracy's one prime mechanism is standardizing the work process.
Algorithmic management puts its layer over whatever grouping it finds, by function or by product, whereas the machine bureaucracy groups its operators by function throughout.

Differs on 6 of 11 dimensions

Basis of grouping Function or process and product or purpose; Machine bureaucracy: Function or process
Specialization Medium; Machine bureaucracy: High
Layers and span of control Flat; Machine bureaucracy: Tall
Prime coordinating mechanism Standardization of work processes and standardization of outputs; Machine bureaucracy: Standardization of work processes
Mode of exchange and control Market and bureaucracy; Machine bureaucracy: Bureaucracy
Boundary permeability Selective to permeable; Machine bureaucracy: Closed

Each standardizes the work and, on the first view of algorithmic management's contested decision rights, keeps formal power at the top. The machine bureaucracy's analysts design the rules and its human supervisors apply them; algorithmic management hands much of that work to software.

Moving into or out of it

Moving in, in our reading, lays software over the structure already there. Pachidi and colleagues (2021) studied one introduction: salespeople pretended to comply while avoiding change, technologists advocated symbolically, and together they unintendedly moved knowing from customer contact to model predictions. A proposal can drive people off (Burbano and colleagues). No study found traces a move out of the form.

Running the reorganization is a question for Change & Transformation.

The thinking behind it

Katherine C. Kellogg, Melissa A. Valentine, and Angèle Christin, "Algorithms at work" (2020): the six mechanisms of algorithmic control and workers' algoactivism, the form's anatomy. Valentine also built flash organizations (2017), a variant, and with Rahman found collaborative repair (2021) holding control better than coercive tool use. Min Kyung Lee (2015): ride-sharing work assigned and evaluated by algorithm.

Richard Edwards, Contested Terrain (1979): simple, technical, and bureaucratic control, the last two set apart from this form's. Frederick Winslow Taylor (1911): scientific management, separating planning from doing, found again in digital form. Harry Braverman (1974): deskilling, in the labor process theory Kellogg and colleagues apply. Henry Mintzberg (1980): the machine bureaucracy, the nearest neighbor, and Type B, a decision-rights view.

Alex Rosenblat (2016): information and power asymmetries as control, the case's mechanism. Alex J. Wood (2019): good gig, bad gig, autonomy beside strain, the second view of centralization. Lindsey D. Cameron (2024): the good bad job, consent that can mask structural problems. Hatim A. Rahman (2021): the invisible cage, control by criteria workers cannot identify. Steven P. Vallas (2020): platforms as permissive potentates, the dissent on mode of exchange.

Stella Pachidi (2021): symbolic conformity that still changed knowing, how an introduction plays out. Sebastian Raisch (2021): the automation-augmentation paradox, behind the human-AI variant. Phanish Puranam (2021): the human-algorithm division of labor as design, the human-AI variant's source.

Richard M. Burton, with Børge Obel (2004): multi-contingency fit and misfit rules, the strategy premise and many firing rules. James D. Thompson (1967): pooled, sequential, and reciprocal interdependence, the interdependence premise. Charles Perrow (1967): technology by exceptions and analyzability, the first view of technology. Paul S. Adler: enabling and coercive bureaucracy (1996), a counter-view against heavy rules, and a critique of labor process theory disputing deskilling as the trend (2007).

Also called

Algorithmic management; algorithmic and data-driven management (Lee and colleagues 2015); algorithmic control (Kellogg and colleagues 2020); humans managed by AI (Hillebrand, Raisch, and Schad 2025, a review).

Works cited

Adler 2007. The future of critical management studies: a paleo-Marxist critique of labour process theory, Organization Studies 28(9), 1313 to 1345.
Adler and Borys 1996. Two types of bureaucracy: enabling and coercive, Administrative Science Quarterly 41(1), from p. 61; a conceptual argument with no new test.
Adler and colleagues 1999. Flexibility versus efficiency? A case study of model changeovers in the Toyota production system, Organization Science 10(1), 43 to 68; one case study of a single plant.
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.
Anthony 2021. When knowledge work and analytical technologies collide: the practices and consequences of black boxing algorithmic technologies, Administrative Science Quarterly 66(4), 1173 to 1212; an ethnography of four groups in one bank.
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.
Bigley and Roberts 2001. The incident command system: high-reliability organizing for complex and volatile task environments, Academy of Management Journal 44(6), 1281 to 1299; one inductive case study.
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.
Burbano and colleagues 2026. A platform rating system and vulnerable workers: evidence from field experiments in Singapore, Organization Science 37(4), 1511 to 1532; two field experiments with one platform.
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.
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 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).
Cameron 2022. "Making out" while driving: relational and efficiency games in the gig economy, Organization Science 33(1), 231 to 252; a four-year qualitative study of workers in one sector.
Cameron 2024. The making of the "good bad" job: how algorithmic management manufactures consent through constant and confined choices, Administrative Science Quarterly 69(2), 458 to 514; a seven-year qualitative study of one industry's workers.
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.
Child and Mansfield 1972. Technology, size, and organization structure, Sociology 6(3), 369 to 393; a multi-firm comparative study.
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 and Lengel 1986. Organizational information requirements, media richness and structural design, Management Science 32(5), 554 to 571.
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.
Donaldson 2001. The contingency theory of organizations, SAGE Publications, Thousand Oaks, CA (Foundations for Organizational Science).
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.
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.
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 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.
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.
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.
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.
Hillebrand and colleagues 2025. Managing with artificial intelligence: an integrative framework, Academy of Management Annals 19(1), 343 to 375.
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.
Kadolkar and colleagues 2025. Algorithmic management in the gig economy: a systematic review and research integration, Journal of Organizational Behavior 46(7), 1057 to 1080 (online 5 September 2024); a topic-model review of the literature across disciplines.
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.
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.
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.
Lawrence and Lorsch 1967b. Differentiation and integration in complex organizations, Administrative Science Quarterly 12(1), 1 to 47.
Lebovitz and colleagues 2022. To engage or not to engage with AI for critical judgments: how professionals deal with opacity when using AI for medical diagnosis, Organization Science 33(1), 126 to 148; one field study in a single hospital.
Lee and colleagues 2015b. Working with machines: the impact of algorithmic and data-driven management on human workers, Proceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems (CHI 2015), 1603 to 1612; a qualitative study of drivers at two ride-sharing services.
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.
Martin and Villani 2026. Algorithmic management and working conditions in Europe: evidence from the AIM-WORK Survey, JRC Working Papers Series on Labour, Education and Technology 2026/05, European Commission, Joint Research Centre, Seville (report JRC147505); a survey of workers across many firms in 27 countries.
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.
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.
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.
Möhlmann and colleagues 2021. Algorithmic management of work on online labor platforms: when matching meets control, MIS Quarterly 45(4), 1999 to 2022; one case study of a single platform.
Mohr 1971. Organizational Technology and Organizational Structure, Administrative Science Quarterly 16(4), 444-459; a quantitative field study of technology and structure.
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.
Noponen and colleagues 2024. Taylorism on steroids or enabling autonomy? A systematic review of algorithmic management, Management Review Quarterly 74(3), 1695 to 1721 (online 5 April 2023); a qualitative systematic review of 172 published articles, many of them case studies, with no pooled analysis.
Pachidi and colleagues 2021. Make way for the algorithms: symbolic actions and change in a regime of knowing, Organization Science 32(1), 18 to 41; one field study of a single organization.
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.
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.
Perrow 1967. A framework for the comparative analysis of organizations, American Sociological Review 32(2), from p. 194.
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.
Rahman 2021. The invisible cage: workers' reactivity to opaque algorithmic evaluations, Administrative Science Quarterly 66(4), 945 to 988; one field study of a single platform.
Rahman and Valentine 2021. How managers maintain control through collaborative repair: evidence from platform-mediated gigs, Organization Science 32(5), 1300 to 1326; one field study of a single platform.
Raisch and Krakowski 2021. Artificial intelligence and management: the automation-augmentation paradox, Academy of Management Review 46(1), 192 to 210.
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.
Rosenblat and Stark 2016. Algorithmic labor and information asymmetries: a case study of Uber's drivers, International Journal of Communication 10, 3758 to 3784; one case study of a single firm's drivers.
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.
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).
Sine and colleagues 2006. Revisiting Burns and Stalker: formal structure and new venture performance in emerging economic sectors, Academy of Management Journal 49(1), 121 to 132; a quantitative study of new ventures in one sector.
Snow and Hrebiniak 1980. Strategy, Distinctive Competence, and Organizational Performance, Administrative Science Quarterly 25(2), from p. 317; a study of top managers' perceptions in four industries.
Steinberg 2022. From automobile capitalism to platform capitalism: Toyotism as a prehistory of digital platforms, Organization Studies 43(7), 1069 to 1090 (online July 2021); a historical and conceptual argument.
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.
Streeck 1984a. Industrial Relations in West Germany: A Case Study of the Car Industry, Heinemann, London; St. Martin's Press, New York (178 pages).
Taylor 1911. The Principles of Scientific Management, Harper & Brothers, New York and London.
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.
Tushman and Nadler 1978. Information processing as an integrating concept in organizational design, Academy of Management Review 3(3), from p. 613.
Valentine and colleagues 2017. Flash organizations: crowdsourcing complex work by structuring crowds as organizations, Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems, 3523 to 3537; a system deployment reported by its designers.
Vallas and Schor 2020. What do platforms do? Understanding the gig economy, Annual Review of Sociology 46, 273 to 294.
Van de Ven and colleagues 1976. Determinants of coordination modes within organizations, American Sociological Review 41(2), from p. 322; a quantitative study of 197 work units in one agency.
Van de Ven and colleagues 2013. Returning to the frontier of contingency theory of organizational and institutional designs, Academy of Management Annals 7(1), 393 to 440 (391 to 438 in the original Taylor and Francis printing).
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.
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.
Yoshino 1968. Japan's Managerial System: Tradition and Innovation, MIT Press, Cambridge, MA.

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.

Algorith­mically managed organization

What it is

Where it came from

Where it fits

What it does well

What it costs

How it fails

What the research found

In practice: one published case

Variants and near relatives

Moving into or out of it

The thinking behind it

Also called

Works cited

Algorithmically managed organization