AMR - What's New About New Forms

WHATS NEW ABOUT NEW FORMSOF ORGANIZING?

PHANISH PURANAMINSEAD

OLIVER ALEXYTechnische Universitt Mnchen

MARKUS REITZIGUniversity of Vienna

In order to assess whether new theories are necessary to explain new forms oforganizing or existing theories suffice, we must first specify exactly what makes aform of organizing new. We propose clear criteria for making such an assessmentand show how they are useful in assessing if and when new theories of organizingmay truly be needed. We illustrate our arguments by contrasting forms of organizingoften considered novel, such as Linux, Wikipedia, and Oticon, against their traditionalcounterparts. We conclude that even when there may be little that existing theorycannot explain about individual elements in these new forms of organizing, opportu-nities for new theorizing lie in understanding the bundles of co-occurring elementsthat seem to underlie them and why the same bundles occur in widely disparateorganizations.

It has become common to lament that extantorganization theories appear to have little ofvalue to say about new and emerging forms oforganizing. More than two decades ago Daft andLewin expressed regret about the fact that thenew forms of organizing they observed seemedfar removed from academic research (1990: 1).A decade and a half later, Schoonhoven, Meyer,and Walsh (2005a,b) and Dunbar and Starbuck(2006) continued to be sufficiently troubled bythe problem to ask for the development of newtheories of organizing to replace our older theo-

ries, which were inextricably rooted in bygonefeatures of economic production (Walsh, Meyer,& Schoonhoven, 2006: 661). More recently, Miller,Greenwood, and Prakash (2009) and Greenwoodand Miller (2010) renewed the call for newer the-ories of organization design to respond to thenovelty and diversity of contemporary forms oforganizing. The consensus diagnosis seems tosuggest that our existing theories of organizingare too rooted in a context that no longer corre-sponds to present day realityand that theneed of the hour is to craft new theories thatbetter correspond to this new reality.On the one hand, it can be argued that the

cure for the ills caused by the excessive contextdependence of past theories can hardly lie inbuilding new theories that closely fit our currentcontexts; doing so would be merely a form of(un)planned obsolescence since todays contextswould inevitably become yesterdays. To illus-trate this point, consider that when critics spokeof new organizational realities that remainedunacknowledged by traditional theories of orga-nization two decades ago, they were referringprimarily to networked forms of organization,such as Thermo Electron or the Toyota suppliernetwork (Daft & Lewin, 1990); today the com-plaint is usually made in terms of the limited

This article has greatly benefited from the guidance ofPeer Fiss and three reviewers. Phanish Puranam acknowl-edges funding from the European Research Council undergrant no. 241132 for The Foundations of Organization De-sign project. Oliver Alexy is grateful for the support of boththe U.K. Innovation Research Centre (RES/G028591/1; spon-sored by the Economic and Social Research Council; theNational Endowment for Science, Technology and the Arts;the U.K. Department for Business, Innovation and Skills; andthe Technology Strategy Board) and the Engineering andPhysical Sciences Research Council (GR/R95371/01). Eliza-beth Philp and Angela Zhang provided valuable researchassistance. We appreciate valuable comments from SendilEthiraj, Gerry George, Srikanth Kannan, Markus Perkmann,Bart Vanneste, and from seminar participants at ImperialCollege London and Politecnico di Milano. All remainingerrors are, of course, our own.

Academy of Management Review2014, Vol. 39, No. 2, 162180.http://dx.doi.org/10.5465/amr.2011.0436

162Copyright of the Academy of Management, all rights reserved. Contents may not be copied, emailed, posted to a listserv, or otherwise transmitted without the copyrightholders express written permission. Users may print, download, or email articles for individual use only.

representation of information technology (IT)and its catalytic role in organizing as evidencedby, say, Linux or Wikipedia (Zammuto, Griffith,Majchrzak, Dougherty, & Faraj, 2007). In thisview, perhaps what we should aim for are gen-eral theories of organizing that explain and in-deed predict the emergence of new forms, ratherthan remain specifically tied to current contexts.On the other hand, the case has also been

made that it is too ambitious for us at the pres-ent state of development of the field to hope fortheories that are general enough to understandnew forms of organizing across contexts (Davis& Marquis, 2005; Kilduff, Mehra, & Dunn, 2011).Rather than adjudicate between these points ofview, in this article we propose that, regardlessof whether one believes that new theories arenecessary or unnecessary to understand thenewer forms of organizing visible around us to-day, a meaningful debate between these view-points requires us to first answer an importantantecedent question: What, exactly, is a newform of organizingand, in particular, whatqualifies it as new? Our primary goal in thisarticle is to help develop an answer to thisquestion.We approach the problem by conceptualizing

a form of organizing as a set of solutions to fouruniversal problems that all organizations con-front (which we will elaborate on later). A novelform of organizing is one that solves one or moreof these problems in a new manner, relative toexisting forms of organizing that have similarobjectives. We believe that once we are able toprecisely specify in this manner the nature ofnovelty in a form of organizing, we can turn to acareful consideration of whether a new theory oforganizing is indeed necessary in light of thenovelty uncovered, and perhaps outline theshape of such a theory.To illustrate how this may be done, we con-

sider a few examples of both established andrelatively recent forms of organizing. These in-clude stereotypical traditional organiza-tionsas most business firms, government bu-reaucracies, and indeed nonprofit agencies arestill organized todaybut also some forms oforganizing hailed as novel, such as Linux andWikipedia. Furthermore, we consider changeswithin a single organization over time from atraditional to a novel form of organizingOticon. These illustrations are not meant to bean exhaustive list of new forms of organizing;

our objective here is conceptual clarity ratherthan empirical validation, so these exampleshelp primarily to explain our arguments ratherthan offer evidence for them. We conclude bynoting how our approach to analyzing forms oforganizing in terms of a set of universal under-lying problems is useful for understanding deepstructural similarities between apparently dis-parate organizations (also see, for example,Gersick, 1991), as well as for studying organiz-ing as a problem-solving process.

WHAT IS A FORM OF ORGANIZING?

We draw on the conceptualization of an orga-nization offered by March and Simon:

Organizations are systems of coordinated actionamong individuals and groups whose prefer-ences, information, interests or knowledge differ.Organization theories describe the delicate con-version of conflict into cooperation, the mobiliza-tion of resources and the coordination of effortthat facilitate the joint survival of an organizationand its members (1993: 2).

While other analysts have offered variants (e.g.,Aldrich, 1979; Burton & Obel, 1984; Etzioni, 1964;Scott, 1998; Stinchcombe, 1965; Weick, 1969), theirvarious conceptualizations of an organizationhave always preserved some common features;in essence, they portray an organization as (1) amultiagent system with (2) identifiable boundar-ies and (3) system-level goals (purpose) towardwhich (4) the constituent agents efforts are ex-pected to make a contribution (we use the termsactor, agent, and individual synonymously).Each of the four elements in this definition is

critical and reflects widely accepted concep-tions of what kind of a system an organizationis. First, the system in question must have morethan a single agent. Second, the set of agentsunder consideration must be well identifiedthrough the definition of the organizationsboundaries. The boundaries of an organizationdefine membership and are complemented byrules for entry and exit into the organization. Forbusiness organizations such as firms, theboundaries can be sharply defined through as-set ownership and employment contracts, whichhelp define the zone of acceptance of employees(Simon, 1953) and provide freedom from adjudi-cation by a court of law in many situationsthrough the doctrine of forbearance (William-son, 1985). The existence of a boundary does not,

2014 163Puranam, Alexy, and Reitzig

of course, imply that organizations are closedsystems; they may be quite amenable to theinfluence of the environment outside theirboundaries, and vice versa (Scott, 1998).Third, it should be possible to ascribe a goal

to the system. Typically, the goal of an organi-zationeven if defined simply in terms of con-tinued existencemay be understood in termsof exploiting some form of gains from joint ac-tion that cannot be achieved by individual mem-bers acting in isolation. The goal may well beimplicit, and even if explicit may not be identi-cal to the goals of the constituent agents (Scott,1998). For instance, conventional business firmshave fairly explicitly stated goals, which typi-cally include, quite prominently, the pursuit ofprofits. However, for most business organiza-tions the goals of the employees and the orga-nization diverge, because the overall goals ofthe organization may not directly appeal to itsemployees (Simon, 1951). Employees may notnecessarily be motivated by producing the wid-gets that their department specializes in or maynot necessarily care about the final goal of thecompany. Rather, employees may make contri-butions within a zone of indifference in returnfor inducements (Simon, 1955). Finally, it isworth noting that implicit goals for an organiza-tion may develop quite independently of its ex-plicit goals and in a manner that no agent isparticularly conscious of; Selznicks (1957) clas-sic description of the institutionalization of theTennessee Valley Authority offers a striking il-lustration. In essence, we recognize organiza-tions through our understanding of their pur-pose, whether intended or emergent; indeed,this is implicit in the Greek root of the wordorganization.Fourth, the efforts of the agents in the organi-

zation are expected to make a contribution to-ward the organizations goal. This condition is,of course, intuitive if organizations are viewedas intentionally designed arrangements forachieving a shared goal (e.g., Aldrich, 1979;Stinchcombe, 1965), while it also allows for gapsbetween this intention and reality (Scott, 1998);no presumption exists that the efforts are indeedbeing made or are successfully moving the or-ganization toward its goal. However, even ininformal organizations that emerge spontane-ously, as opposed to intentionally designedones, if we recognize a system-level goal at all,it must be because the constituent agents ef-

forts appear to contribute toward this goal. Forinstance, the existence of an emergent informalorganization such as a community of practicemay be recognized by the efforts of the membersto further the goals of the community (while pos-sibly furthering their own goals as well). Suchan emergent informal organization could, ofcourse, coexist within or across intentionally de-signed formal organizations (e.g., Rosenkopf,Metiu, & George, 2001).The conceptualization of an organization we

have adopted precludes neither a dyad nor acorporation from consideration as an organiza-tion, nor the various departments, units, sub-units, and teams within a corporation; each ofthese can be treated as an organization, albeitof widely varying longevity and scale for thepurpose of analysis. Further, both intentionallydesigned and emergent organizations qualify.For designed organizations, we expect thatagents efforts will contribute to organizationalgoals; for emergent organizations, we may infertheir existence because the agents efforts ap-pear to contribute to an organizational goal.However, individuals, goalless networks, andcrowds (but not mobs), or, in general, collectivesin which the constituent agents are neither ex-pected to nor seem to make efforts towardachieving the collectives goals, are ruled out byour definition.In this conceptualization boundaries and

goals jointly identify organizations uniquely.This is because the same set of individuals mayform multiple organizations, each with a differ-ent goalsome implicit and others explicit.Thus, the same set (or partially overlappingsets) of individuals may be colleagues in (a di-vision of) Microsoft as well as cofounders of astart-up or comembers of a country club.

Organizing As Problem Solving

Weick (1969, 1974) persuasively argued thatorganizations and organizing are not the samethings. Organizing is a process, and we believeit is useful to view it as a problem-solving pro-cessas a search through a problem space(Newell & Simon, 1972; also see reviews by Dun-bar, 1995, and Fischer, Greiff, & Funke, 2012). Inthis section we propose a useful set of dimen-sions to understand this problem space.In keeping with most theories of organization,

we take as a given that the agents who make up

164 AprilAcademy of Management Review

human organizations have limited capacity foraccessing and processing information (thebounded rationality assumption) and must becompensated for their efforts (the self-interestassumption), with it being understood that com-pensation need not take only monetary or evenmaterial form (Simon, 1947; Williamson, 1975).Several scholars have remarked that any func-tioning organization comprising agents withthese attributes must have solved two funda-mental and interlinked problems: the division oflabor and the integration of effort (Burton &Obel, 1984; Lawrence & Lorsch, 1967; March &Simon, 1958; Mintzberg, 1979). We develop thisidea by arguing that these problems can be fur-ther decomposed into four universal problemsthat any form of organizing, by definition, mustsolve. The first two problems, task division andtask allocation, we jointly refer to as the divi-sion of labor. The second two, reward provisionand information provision, jointly constitute theproblem of achieving integration of effort (seefor instance Galbraith, 1973: 3; March & Simon,1993: 2; Porter, Lawler, & Hackman, 1975). Weconsider each of these in detail.The division of labor. The division of labor in

organizations refers to the breakdown of the or-ganizations goals into contributory tasks andthe allocation of these tasks to individual mem-bers within the organization.Task division refers to the problem of mapping

the goals of the organization into tasks and sub-tasks. When this is done consciously by humanagents, it typically involves a means-ends de-composition of the goals of the organization tocreate the set of interrelated subtasks and infor-mation/material transfers that are believed tobe necessary for the goals of the organization tobe achieved (Newell & Simon, 1972). Workflowdiagrams, business process mapping, valuechains, and engineering drawings are instancesof artifacts that capture task division in the typ-ical business firm. The knowledge that enablesa human designer to accomplish more effectivetask divisions may be thought of as architec-tural knowledge, which Baldwin defines asknowledge about the components of a complexsystem and how they are related (2010: 3). Thisknowledge is gleaned from experience and ismore likely to exist for well-understood goals(von Hippel, 1990).Task allocation refers to the problem of map-

ping the tasks obtained through task division to

individual agents and groups of agents. In thetraditional business firm, the designation of for-mal roles and the recruitment of individuals intothem are the usual mechanisms to match sub-tasks to members skill profiles. This may ormay not involve assignation of clusters of simi-lar repeatable tasks to an individual (i.e., spe-cialization), although the benefits of specializa-tion in general are considerable (Smith, 1776).Alternatives include task allocation to minimizeinterdependence across agents, to increase di-versity of tasks, and to assign responsibility fortangible outputs rather than intermediate steps(Hackman & Oldham, 1976). Both the specifica-tions of the role and the matching of individualagents to roles are typically conducted within abusiness firm by somebody with the formal au-thority to do so; indeed, the authority to conducttask allocation is one of the key features of anemployment contract (Simon, 1951) and definesthe boundary of the firm.The integration of effort. The integration of

effort within an organization requires the reso-lution of both cooperation and coordinationproblems (Gulati, Lawrence, & Puranam, 2005;Lawrence & Lorsch, 1967). The former is accom-plished through providing the necessary moti-vation, the latter through information. The fail-ure to solve either problem is sufficient to causea breakdown in integration; good solutions toboth problems are necessary to achieve integra-tion (Camerer & Knez, 1996, 1997; Heath &Staudenmayer, 2000).The provision of rewards refers to the problem

of mapping a set of rewards (both monetary andnonmonetary) to the agents in the organiza-tionin order to motivate the agents to cooper-ate by taking costly actions toward executingthe tasks they have been allocated. Any organi-zation must feature either implicit or explicitmechanisms for providing inducementsex-trinsic or intrinsicto its members in order tomotivate entry and continuance of membership,if its constituent agents need to be motivated tocontribute (Simon, 1951). For instance, proce-dures for bestowing monetary compensation asstipulated in the employment contract, and non-monetary compensation in the form of work con-ditions and titles, the choice of colleagues, andadvancement opportunities, are the primary re-ward mechanisms in the traditional businessfirm (Gibbons, 1998; Prendergast, 1999).


The provision of information refers to the prob-lem that an organizations agents have the in-formation needed to execute their tasks and co-ordinate actions with others. Coordinated actionrequires that enough information exist for inter-acting individuals to be able to act as thoughaccurately anticipating each others actions(Schelling, 1960). Students of organizations havepointed out that there are two basic means ofsolving the information provision problem: ei-ther the task division and allocation is such asto reduce the need for such information (e.g.,through the use of directives, schedules, plans,and standardsas long as each of us obeysthem, our actions will in aggregate be coordi-nated), or the channels needed to generate suchinformation can be enriched (e.g., the opportu-nity for rich face-to-face or electronic communi-cation; March & Simon, 1958). As with task allo-cation, authority also plays a critical role in thedesign and implementation of the reward andinformation distribution system of the tradi-tional business firm.

The Universality of the Problems of Organizing

We argue that that the existence of solutionsto the four problems of organizing we high-lighttask division, task allocation, reward pro-vision, and information provisionare individ-ually necessary and collectively sufficient for anorganization to existthat these problems are,in effect, universals of organizing.First, we define what we mean by the exis-

tence of a solution. Since we have definedeach of the problems in terms of choosing amappingfrom goals to tasks (task division),from tasks to agents (task allocation), from re-wards to agents (reward provision), and frominformation to agents (information provision)that is meant to contribute to the goals of theorganization, we say that a solution exists ifsuch a mapping exists. It may thus be seen as achoice about organizing. The value of proposedsolutionsthe magnitude of the contribution ofthat solution toward achieving the goal of theorganizationmay vary significantly. Higher-valued solutions are the ones that bring the or-ganization closer to achieving its goals, andthere may be equifinality in solutions.An organization, given how it is defined and

the behavioral assumptions of bounded rational-ity and self-interest, must necessarily feature

solutions to each of the four basic problems oforganizing. Absent organizational choices thatare expected to retain and motivate individuals,enabling them to undertake their assigned tasks(which aggregate toward the organizationsgoals) in a coordinated manner, one cannot rec-ognize the existence of an organization definedas a multiagent system with identifiable bound-aries and system-level goals toward which theconstituent agents efforts are expected to makea contribution.Further, the existence of solutions to the four

problems is also sufficient for an organization toexist. If the four problems have proposed solu-tions, then we have in effect (a) a set of tasksbelieved to contribute toward a goal (b) as-signed to a set of agents who have been (c)rewarded for and (d) informed about executingthose tasks so that their efforts are expected tocontribute toward the goal of the organization.Thus, for an organization to exist, it is necessaryand sufficient that solutions exist to each of thefour basic problems of organizing.If the set of problems of organizing we have

identified is both parsimonious and complete(necessary and sufficient), then it must be alsobe universal to all organizations. In order to beable to meaningfully analyze differences informs of organizing, we believe it is useful todefine a form of organizing as follows:

Definition 1: A form of organizing is aspecific set of solutions to the four uni-versal problems that any organizationmust address in order to exist.

This definition highlights the many-to-one re-lation between solutions and each problem;while the problems that any form of organizingmust solve in order to attain its goals are uni-versal, the specific ways in which those prob-lems are proposed to be solved can vary consid-erably, and we argue that, indeed, this is whatgenerates novelty in forms of organizing. In thisview a form of organizing can be seen as a set ofproposed solutions or hypotheses about how tosolve the four universal problems of organizing; itcan be viewed as a model of organizing, and thesolutions will generally vary in their fitness orquality (Baldwin & Clark, 2000; Levinthal, 1997).In traditional business organizations it is cer-

tainly true that formal authority (Weber, 1922) isfrequently both the origin of and the basis onwhich the solutions to the problems of organiz-


ing are enforced (Galbraith, 1973; Simon, 1947;Williamson, 1985). However, these solutionsneed not be crafted by an actor with formalauthority, or perhaps by any actor at all. Selec-tion through competition among variations pro-vided by unintentional experimentation mayalso throw up working solutions, with little or noforesight involved (Hodgson & Knudsen, 2010).The existence of emergent informal organiza-tions, often within the context of a formal orga-nizational structure, may share a similar under-lying dynamic; a set of solutions that contributetoward the organizations implicit goal mayhave emerged without conscious design.

IDENTIFYING NOVELTY IN FORMS OFORGANIZING: WHAT MAKES THEM NEW?

To the extent that the set of solutions to thefour universal problems we identified character-izes a form of organizing, it follows that noveltyin forms of organizing can only arise in the formof a novel solution to one of these problems or inthe form of several novel solutions to several ofthese problems. Put simply, a new form of orga-nizing must solve one or more of the problems ofdivision of labor and/or integration of effort in anovel manner. But novel relative to what? Weneed not insist that the standard of novelty ofthe solution requires it to be new to the world;rather, there must at least be novelty relative toa comparable group of organizations.To establish the appropriate comparison

group, we believe it is useful to consider orga-nizations with comparable goals. Thus, one mayobserve a new form of organizing to developsoftware (e.g., Linux) or to generate encyclope-dic content (e.g., Wikipedia), but in each casethe novelty can only be appreciated with respectto existing forms of organizing that achieve thesame goals. Thus, we argue that the more sim-ilar the goals of two organizations, the moreinsightful will be a comparison of their underly-ing forms of organizing. This also implies thatcomparisons of organizations that share similargoals at high levels of generality (e.g., be prof-itable or survive) are less likely to be usefulthan comparisons of organizations that sharegoals at high levels of specificity (e.g., developan operating system or provide encyclopediccontent). Comparing Wikipedia with, say, Mi-crosoft is therefore less likely to be a usefulexercise than comparing it with the Encyclopae-

dia Britannica if the goal is to understand nov-elty in forms of organizing. It may also be in-structive to compare the same organization overtime as its form of organizing (but not its goals)change. We formally define this as follows.

Definition 2: Novelty in a form of orga-nizing can be determined by assessingnovelty in its solutions to the universalproblems of organizingin comparisonwith existing forms of organizing withcomparable goals.

We now discuss some recent and prominentdevelopments in organizing to illustrate howour framework is useful in pinpointing whatmakes them novel. Specifically, we look at threerelatively recent forms of organizing as instan-tiated by Linux, Wikipedia, and Oticon. Each ofthese has already been studied carefully by re-searchers, so we have a wealth of information todraw on. The key dimensions on which each ofthese forms of organizing differ from their tradi-tional counterparts are summarized in Table 1.

Open Source Software Development

Open source software development (OSSD;e.g., what underlies the operating system Linux)has been hailed as a promising new mode oforganization (von Hippel & von Krogh, 2003: 213)and an entirely new model of innovation ofrelevance to many fields beyond open sourcesoftware (von Krogh & von Hippel, 2006: 976). Toillustrate the value of our ideas, it is useful tocontrast OSSD with proprietary closed sourcesoftware development (PCSSD), the in-house de-velopment of software by commercial firms.In both forms of organizing, task division has

at least some elements of centralization, al-though it is much more in evidence in PCSSD,where it is conducted by a designated systemarchitect. In OSSD the initial formulation of theproblem by the founder of a project influencesthe task division (MacCormack, Rusnak, & Bald-win, 2006), but it is elaborated and developedsignificantly as others join and contribute to theproject. Furthermore, whereas task allocation isformally mandated within a firm, OSSD departssignificantly from this because of self-selectionof members into roles within the organization.While software developers inPCSSD are typi-cally salaried employees of the firm, contribu-tions by individuals to OSSD are voluntary and


TABL

E1

Con

trasting

Exam

ples

forSo

lution

sto

theUniversal

Prob

lemsof

Organ

izing

Goa

lSo

ftwareDevelop

men

tProvisionof

Encyclop

edia

Con

tent

Designan

dProd

uction

ofInno

vative

Hea

ring

Aids

PCSS

DOSS

DClassic

Encyclop

edia

Wikiped

iaOticonPre1987

OticonPo

st1989

Task

division

System

architectprovides

task

architecture;o

nly

visible

to(selected)

mem

bersof

firm

Fou

nder

provides

laye

r1

architecture,w

hich

subsequen

tlyev

olve

s;visible

toev

eryo

ne

Editorialteam

provides

task

architecture;

only

visible

to(selected)mem

bers

offirm

Nocentralizedtask

division

Rather

detailed

specification

bymanagem

ent;

competen

cebased

High-lev

elsp

ecification

ofsixstrategic

tasks

bytheCEO;d

etailed

subtaskssp

ecified

byproject

leaders

Task

allocation

Bythefirm

(basedon

skills)

Self-selection(basedon

skills

andpreferences)

Bytheed

itorialteam

(basedon

skills/exp

ertise)

Self-selection

(basedon

preferences)

Authoritative

task

allocation

;assignmen

tof

specialist

tasksto

specialist

dep

artmen

ts

Self-allocation

into

projects;

within-

project

subtask

assignmen

tRew

ard

distribution

Salaries

andbon

usesas

determined

bythefirm

;combination

ofex

trinsicincentive

sandmon

itoring

Universa

lfree

accessto

OSS;intrinsic

motivation

,visibility

within

community;

signalingto

employe

rs

Salaries

asdetermined

bythefirm

,plus

recognition(inthe

case

ofsu

bject

area

experts)

Intrinsic

motivation

,visibilitywithin

community

Fixed

financial

compen

sation

schem

e,as

wellasstatusalonga

seve

n-laye

rhierarchy

Mon

etary

rewards

restructuredto

includestock

option

s;increa

sed

emphasison

intrinsicrewards

Inform

ation

flow

sPhysicalcollocation

and

grouping,p

lus(virtual)

supportinfrastructure

andtools

Virtualsu

pport

infrastructure

and

tools

Physicalcollocation

andgrouping,p

lus

virtualsu

pport

infrastructure

and

tools

Virtualsu

pport

infrastructure

andtools

Formalwritten

communication

;plans,

mem

oranda,e

tc.

Inform

ation

transfer

largelypaperless

(facilitatedbyIT)

may be rewarded by the fulfillment of use-needsand intrinsic motivation (including status andresponsibility; Shah, 2006; von Hippel & vonKrogh, 2003), as well as extrinsic motivation, in-cluding the signaling of competence to potentialemployers (Lakhani & Wolf, 2005; von Krogh,Haefliger, Spaeth, & Wallin, 2012).1

Coordination and information provision is-sues in PCSSD and OSSD are resolved some-what differently. Within the firm, formal organi-zational structure, physical collocation, andshared tenure enable coordination, in additionto technologies for virtual collaboration. InOSSD, coordination through collocation or pre-existing shared experience is typically not anoption; instead, OSSD developers rely morestrongly than their PCSS counterparts on virtualcollaboration technologies and the software ar-tifact itself for coordinationfor instance, by in-cluding more comments in and documentationwith their source code (Cramton, 2001; Hinds &Kiesler, 2002). In this sense it is quite clear thatwithout the affordances provided by IT, particu-larly virtual and mass collaboration, as well asthe ability to visualize the entire work systemthat is inherent in OSSD platforms (such asSourceForge, for instance), this form of organiz-ing could not exist (Zammuto et al., 2007).Unlike in PCSSD, OSSD cannot rely on formal

authority created through employment contractsor property rights to solve any of the problems oforganizing (von Hippel & von Krogh, 2003), yetresearch suggests that certain individuals ac-tively exert authority; for example, founders andadministrators reject contributions made to theproject and exercise some influence over whojoins the project (Dahlander & OMahony, 2011;Lee & Cole, 2003).

Wikipedia

Seeking to provide comprehensive informa-tion, traditional reference works, such as the

famous Encyclopaedia Britannica or the Ger-man Brockhaus, divide the tasks of informationgathering, fact checking, and information pre-sentation across groups of specialists who havemonetary incentives for doing so. The profes-sional editorial team allocates tasks to the dif-ferent experts within the organization, coordi-nates the exchange of information among them,and retains the ultimate power to decide whichcontent enters the encyclopedia and in whatformat. Thus, task division and task allocationare heavily centralized; rewards distributionand information provision follow the template ofany traditional business firm.Free user-generated reference works, the most

popular among them being Wikipedia, havepresented alternative organizational ap-proaches to fulfilling the same goal as conven-tional encyclopedia compilers. Acting as open(online) platforms, they allow any registereduser to create new articles and anyone to editexisting content, without offering their authorsfinancial rewards. Task division in Wikipedia isarguably even less centralized than in OSSDprojects (like Linux), where the founders initialproblem statement may shape how the task ar-chitecture evolves. However, like Linux, contrib-utors self-select into projects they wish to con-tribute to. Thus, no chief editor compiles aprospective table of contents and then proceedsto allocate entry writing to designated experts.The online encyclopedia also offers the commu-nication platform required to pass on relevantinformation between agents.Notably, because neither task division nor

task allocation is centralized, exceptions oftenarise. The first kind includes factual misinfor-mation or duplicationthe inevitable hazardsof a noncentralized division of labor. Here Wiki-pedia relies on the detection and correction ofmisleading material through knowledgeableuser contributors (Gorbotai, 2012). The secondkind of exception elicits a more authoritativeform of intervention; these are the cases of libel-ous or malicious content, in which case thefounder has publicly stated that he feels com-pelled to intervene in an authoritative manner(e.g., TEDGlobal, 2005).

Oticon

In order to compare different forms of organiz-ing, we could also track the way the same entity

1 OSSD and Wikipedia are largely collaborative forms ofengaging autonomous actors in an innovative project. Suchapproaches are in contrast to contest-oriented models thatseek one best answer in response to a stated problem, suchas instantiated in InnoCentive, NineSigma, or TopCoder (forrecent works on this topic, see, for example, Boudreau &Lakhani, 2013, and Guinan, Boudreau, & Lakhani, 2013). Yetsome of the motivations found in the collaborative forms mayalso be present in the contest forms, and vice versa.


changes its form of organizing over time, pro-vided (1) the overall goal of the firm remainstemporally constant and (2) the reorganizing ef-fort is radical in the sense that it leads to afundamentally different way the organizationseeks to solve division of labor and integrationof effort problems.Arguably one of the best documented cases in

the modern management literature of such adramatic change is that of the Danish hearingaids producer Oticon (OKeefe & Lovas, 2002).The critical phase of interest in its corporatehistory was triggered by the successful intro-duction of a radical new technology by one ofOticons competitors in 1987, causing the firmsworldwide market for hearing aids to plummetfrom 14 to 9 percent over a period of roughlyeighteen months and causing the company tolose more than half of its market capitalization.In a dramatic attempt at organizational change,the newly appointed CEO Lars Kolind embarkedon a restructuring of the firms activities, a re-structuring that has since received much atten-tion and is often referred to as the spaghettireorganization (Foss, 2003).Todayas much as in 1904 when it was

founded, and undeniably also in the year 1989when it was reorganizedOticons overarchinggoal is to create and capture rents from thedevelopment, manufacturing, and sales of hear-ing aids worldwide. Yet the ways in which itpursued its aim pre 1989 and between 1989 and,say, 1993 differed considerably. Pre 1989, author-itative task allocation entailed assigning spe-cialist tasks to specialist departmentsmarket-ing and sales, finance, and manufacturing andoperations (OKeefe & Lovas, 2002: 1). Rewardscame in the form of fixed financial compensa-tion, as well as status, allowing executives topeer down a seven-layer hierarchy from theirspacious, plush, and far removed offices(OKeefe & Lovas, 2002: 1). Traditional adminis-trative structures were used to provide informa-tion to employees, including the heavy use offormal (written) communication.Shortly after taking the helm in 1987, Kolind

dramatically changed the way tasks were di-vided and allocated to employees, and he alsoaltered reward distribution and information pro-vision systems. Around 1989 he defined, at avery high level, six tasks that employees wouldhave to be engaged innamely, they would have

(1) to develop and articulate strategic goalswhich defined the strategic intent of the organi-zation; (2) to sponsor strategic initiatives; (3) toallocate financial capital to strategic initiatives;(4) to recruit people to the organization; (5) to takeresponsibility for the development of one area offunctional expertise and knowledge in the orga-nization . . . ; (6) to work on at least two strategicinitiatives at any given point of time (Lovas &Ghoshal, 2000: 888).

Engaging in the first five tasks would be re-served for the firms top managementmem-bers of the higher of only two remaining hierar-chical layers in the firm overall. Task 6 would liewith the other members of the organization.Within this grand allocation scheme, however,employees could rather freely choose how tospend to their work time thenceforthwhichprojects to initiate and how to structure them(conditional on top managements budgetingapprovala process that Kolind intended to befast and informal).Monetary rewards were restructured to in-

clude stock schemes, and self-selection intoprojects allowed for greater self-actualization ofemployees. Information transfer became largelypaperless, with Kolind symbolically shreddingall paperwork, after scanning and saving it inthe firms IT system, and letting it tumblethrough a transparent tube running through thefirms cafeteria. The project-based approach towork from this point on mostly relied on directface-to-face information exchange among mem-bers working within the same projects, with in-dividuals able to switch between projects andwork locations through the simple expedient ofcarrying all their data and files with them on apush cart. IT thus played both an enabling rolein allowing this degree of mobility of employeesacross projects and perhaps a more importantsymbolic role through the initiative to gopaperless.As the comparisons in Table 1 make clear, the

consideration of organizations like Oticon, Li-nux, and Wikipedia suggests that, relative totheir traditional alternatives, these are legiti-mately seen as novel forms of organizing (inde-pendent of their success). Our approach allowsus to point precisely to what is novel about eachof themin what way they solve the universalproblems of organizing in a new way relative tothe traditional alternative (as summarized inTable 1).


DO NEW FORMS OF ORGANIZING REQUIRENEW THEORIZING?

While the preceding account highlights how toidentify novelty in forms of organizing relative totheir traditional alternatives, this still leaves openthe question of what this means for theorizinginother words, whether the mechanisms involvedrequire novel theorizing or can be accommodatedwithin existing theories (also see Felin & Zenger,2011: 169). To examine this issue, we considerwhatprior theories have to say about the mechanismsthat are used to achieve the novel division of laborand the integration of effort in Linux, Wikipedia,and the new Oticon.

Novelty in Task Allocation

Common to the aforementioned examples isthe self-selection of contributors into subtasksthey prefer, rather than assignment to thosetasks by somebody with authority. As Table 1shows, this feature is common to OSSD, Wikipe-dia, and the new Oticon. While this is clearlynovel when compared to the respective tradi-tional counterparts in each case, situationswhere individuals select from among optionsare, of course, staple fare in any theory of deci-sion making (e.g., Kahneman & Tversky, 1979;March & Simon, 1958; Simon, 1947), so we do notthink the choice process itself requires much inthe way of fresh theorizing. However, as we notebelow, this solution to the problem of task alloca-tion gives rise to associated changes in the solu-tions adopted on the other three dimensions.

Novelty in Reward Distribution

The idea that intrinsic motivation stemmingfrom properties of the task plays an importantrole in compensating employees is a well-established one (Hackman & Oldham, 1976).This is certainly a key and indeed explicit prem-ise in each of the examples we have discussed,whether it be voluntary contributors to Wikipe-dia or the self-selection of employees into proj-ects they like in Oticon (also see Lakhani & Wolf[2005] on its importance in open source projects).Also common to the reward distribution rules

in OSSD and Wikipedia (and more subtly so forOticon) is the hazard of free-ridinganyone canbenefit from the achievement of the goals of theorganization (i.e., the creation of the OSS or the

Wiki) without even having to be a member ofthe organization. The potential for free-ridingexists in any situation where an individual canbenefit by withholding contribution, given thatothers are contributing (Hardin, 1968). OSS andWikipedia, by definition, are intended to be pub-lic goods (e.g., Gambardella & Hall, 2006), sofree-riding can impede the provision of the so-cial optimum level of these public goods. In thelimit, we may note that each individual can ben-efit (through free and open access to the soft-ware or encyclopedic content) by not contribut-ing at all. So why do people contribute?The puzzle, however, is not unique to Linux or

Wikipedia; other public goods that can exist de-spite the absence of mandated contribution in-clude blood banks (Titmuss, 1970), the army,and, more generally, charitable and voluntaryservice organizations. Indeed, the mechanismswhereby the free-rider problem is effectivelysuppressed or mitigated in these situations arewell known. These include social norms thatencourage contributions and that sanction non-contribution (e.g., Ostrom, 1990), the existence ofprivileged groups that value the public goodhighly enough to contribute regardless of the(non) contributions of others (Olson, 1971), andthe creation of institutions in which transactioncosts are reduced sufficiently between the po-tential beneficiaries of a public good to be ableto pool their efforts effectively (Coase, 1960).These mechanisms seem also in evidence in

OSSD: social norms that govern fair contributionhave been documented by Lee and Cole (2003),OMahony (2003), and Shah (2006) and noted bythe founder of Wikipedia (TEDGlobal, 2005). Theformal legal structure of Oticon is, of course, aclassic institution in which transaction costs arelimited (i.e., a firm), but online platforms such asSourceForge or Wikipedia also represent insti-tutions that lower transaction costs for contrib-utors. Contributors to such projects also repre-sent privileged group members who gainsufficiently from their own user needs and mo-tivation to initiate and contribute to a projectirrespective of whether others do so (Raymond,1999; von Hippel & von Krogh, 2003). What isinteresting is that free-riding may well exist de-spite the use of the latter two mechanisms in theliteral sense that there will be users of the OSSor Wikipedia who make no contributionbutthe mechanisms noted above ensure that thepublic good is nonetheless created. In sum, the


universal access reward distribution rule inOSSD or Wikipedia may not eliminate free-riding, but the presence of norms, institutions,and heterogeneity in utility allows this publicgood to be created, and these are well-knownmechanisms in other contexts.

Novelty in Information Provision

While less prominent in the case of Oticon, adistinctive feature of OSSD and Wikipedia is thatthe contributing members who self-select into theorganization are rarely (if ever) collocatedthey form a classic distributed organization (Lee& Cole, 2003; Shah, 2006). As we know from theextensive literature on distributed work, whenface-to-face communication channels are un-available, the burden shifts to one of two othermechanisms. The first of these is creating andleveraging common groundknowledge that isshared and known to be sharedto enable co-ordination of activities. IT often provides themeans for creating common ground across loca-tions, by allowing electronic communicationchannels, cross-visibility of actions and context,and the visualization of joint tasks (see Zam-muto et al., 2007, for a discussion of the affor-dances for virtual collaboration provided by IT).The second is creating a modular task architec-ture that enables different individuals to work inparallel without having to explicitly coordinatetheir actions (Hinds & Kiesler, 2002; Srikanth &Puranam, 2011). This distinction mirrors an es-tablished one in organization theory betweencoordination through feedback versus program-ming (Galbraith, 1973; March & Simon, 1958;Thompson, 1967).Modular task architectures are important in

OSSD, although self-selection by contributorsinto tasks within the system may make it hard tomaintain complete independence of actionacross them. The creation of common ground inOSSD relies on many of the same collaborativetechnologies as seen in other forms of organiz-ing distributed work, including the use of email,bulletin boards, and version control software(e.g., Lee & Cole, 2003; Raymond, 1999; Shah,2006)as are seen, for instance, in offshore soft-ware development (Srikanth & Puranam, 2011). Itis undoubtedly also true that OSSD and Wikipe-dia benefit from the fact that in both cases thefinal productthe software or the online ency-clopediais its own representation. To the ex-

tent the code or content (and changes to it) iscommonly visible, coordination across differentagents contributing to it becomes simplified.However, none of these features are unique to

these novel forms of organizing; they are staplefare in more conventional forms of distributedwork and have been theorized about extensively(e.g., Faraj & Sproull, 2000; Hinds & Kiesler, 2002;Wasko & Faraj, 2005).

Novelty in Task Division

In general, there are a large number of waysto conduct task division. Some well-known ob-jectives that may motivate the choice of partic-ular ways to divide tasks include the gains fromspecialization (Smith, 1776), the match to indi-vidual skills (Moreland & Argote, 2003) and pref-erences (Hackman & Oldham, 1976; Madsen, De-sai, Roberts, & Wong, 2006), the gains fromparallel efforts (Baldwin & Clark, 2000), lowermeasurement costs (Barzel, 1982), and enhancedaccountability (Zenger & Hesterly, 1997).While specialization, skill matching, parallel-

ism, and accountability are all well-recognizedcriteria for task division, in each of the newforms of organizing we have discussed, the factthat individuals self-select into tasks (ratherthan are assigned to tasks by a hierarchicalsuperior) suggests another rationale for prefer-ring certain task divisions over others: transpar-ency. As Baldwin and her colleagues have ar-gued, a task division that is transparent (i.e.,makes visible as fine-grained a task structureas possible) allows potential contributors to se-lect specific tasks so as to participate based ontheir personal skills and motivations, which in-creases the likelihood they will choose to contrib-ute instead of free-ride (e.g., Baldwin&Clark, 2006;Colfer & Baldwin, 2010; MacCormack, Baldwin, &Rusnak, 2012; MacCormack et al., 2006). Contribu-torsmay specialize (von Krogh, Spaeth, & Lakhani,2003) or seek variation or opportunities to exerciseparticular competencies (Lakhani & Wolf, 2005).While the criterion of transparency as a basis

for task division is now recognized, further workthat explicitly recognizes the trade-offs betweendifferent criteria for task division will doubtlessbe useful. For instance, we may conjecture thatin OSSD or Wikipedia, transparency overridesparallelism as the key criterion for task division,given that the founder and the core of memberswork across interdependent modules, make de-


cisions about acceptance of contributions, andensure that no essential tasks remain incom-plete or redundancies arise (von Krogh etal., 2003).In profit-motivated business organizations in

general, administrative authority plays a keyrole in the solution to the four basic problems oforganizing, as we have noted. Is it missing inthe new forms of organizing we discuss? Whilein Oticon (both pre and post reorganization) staffis employed through formal employment con-tracts, no explicit contracts exist in OSSD or inWikipedia that could provide a source of formalauthority. However, this does not mean that au-thority relationships are missing.Accounts of open source projects demonstrate

the exercise of authority by the projects found-ers in a number of ways (e.g., Business Week,2004; Raymond, 1999). For example, the assign-ment of incoming reports describing errors inthe software (bug reports) to specialists in acertain area of the software is a form of taskallocation based on the founders authority andis a common practice in many OSSD projects. Inaddition, for a new project on the SourceForgeplatform, contributors interested in the projectcannot self-select into the set of core members(e.g., Lee & Cole, 2003). Rather, these noncoremembers self-select into the periphery and con-tribute as outsiders until core membership iseventually bestowed on them by the founder orhis delegate(s). A similar structure of authoritybased on founders rights appears to be inplace at Wikipedia, as indicated by its founder(TEDGlobal, 2005), although it is exercised onlyfor managing some kinds of exceptionssuchas the addition of libelous or malicious con-tentand as a backup if its other exceptionmanagement mechanism breaks down.This authority differs from that seen in con-

ventional business contexts in that it is notbased on formal contracts and it lays a greateremphasis on the right to accept or reject contri-butions and claims of membership (Dahlander &OMahony, 2011), rather than the right to directan employee within a zone of indifference (Si-mon, 1951). However, the broader notion that au-thority can arise from noncontractual sources isat least familiar to us from the work of Weber, ifnot from earlier work.In sum, novelty in forms of organizing does not

always imply the need for novel theorizing. Thisis because the solutions in question may be

novel relative to other forms of organizing withsimilar goals but not necessarily to the world.They may therefore be very well understood the-oretically in other contexts.

NEW FORMS OF ORGANIZING AS NOVELBUNDLES OF OLD SOLUTIONS

The examples of new forms of organizing wehave considered here are justifiably novel rela-tive to their comparison group of organizationsbecause of novelty in how they attempt to solveone or more of the universal problems of orga-nizing (Table 1). At least in these examples, ITplays an important role in enabling their exis-tence, particularly through the affordances itprovides of visualizing the global task architec-ture and allowing mass and virtual collabora-tion (e.g., Wikipedia, Linux), but also in a sym-bolic form (e.g., the public display of an iconrepresenting paperlessness at Oticon). How-ever, as we have argued, existing theory pro-vides a sound basis for understanding much ofthis novelty, because the solutions are rarelynovel to the world. Of course, this may well be aproperty of the set of illustrations we have con-sidered. What is more interesting in our view isthat our illustrations also offer an unusuallyclear window onto the complementarities be-tween the solutions for each of the four basicproblems of organizing (Milgrom & Roberts,1990, 1995).In the context of the four universal problems of

organizing that we define (i.e., task division,task allocation, information provision, and re-ward provision), we can say that a pair of solu-tions to these problems may be complementswhen adopting one increases the value of theother (this is the usage introduced by Milgromand Roberts [1990] and later shown by them to beanalytically identical to notions of synergy andfit [Milgrom & Roberts, 1995]).Consider task allocation through self-selec-

tiona common solution to task allocation ineach of the new forms of organizing consideredabove (see Table 1)which has implications forthe value of solutions to task division, rewarddistribution, and information provision. Giventask allocation through self-selection, task divi-sion in each case must ideally be transparentenough to enable self-selection (e.g., Baldwin &Clark, 2006; MacCormack et al., 2006). Clearly,this is the case in OSSD as well as Wikipedia;


the switch to paperless administration couldhave facilitated a similar move at Oticon(though we lack the primary data on this toknow for sure). Thus, the value of self-selection(as a solution to the task allocation problem)increases if the solution to the task divisionproblem features a high emphasis on transpar-ency, and vice versa. Self-selection-based taskallocation and transparency-based task divi-sion are therefore complements.In turn, appropriate task allocation through

self-selection may obviate the need for provid-ing extrinsic incentives, as in Wikipedia orOSSD (Hackman & Oldham, 1976; Wasko &Faraj, 2005), and possibly lower them at Oticon.Thus, the value of self-selection (as a solution tothe task allocation problem) must also increasewith the reliance on intrinsic motivators as asolution to the reward distribution problem;these, too, are complements.Furthermore, since task allocation through

self-selection into a transparent task architec-ture may also draw in a diversity of possiblygeographically distributed contributors, the in-formation requirements for coordinated actionswill necessarily look different from those for em-ployees traditionally gathered together by ad-ministrative authority. For geographically dis-tributed actors, as in Wikipedia or OSSD,technologies that enable physically distributedagents to communicate and observe each oth-ers contributions are therefore solutions (to theinformation provision problem) whose valuemay increase with the adoption of self-selectionand transparent task architecture as solutions totask allocation and task division problems(Hinds & Kiesler, 2002). For collocated contribu-tors, as in Oticon, increasing the amount of in-formation about each others actions may occurinstead through greater emphasis on face-to-face communications and rich, regular informa-tion exchange (OKeefe & Lovas, 2002).This pattern of spiraling linkages among the

solutions to the problems of organizingin ourexamples, among self-selection, transparenttask architectures, intrinsic motivation-basedrewards, and virtual collaboration technolo-gieshighlights their complementarities(Drazin & Van de Ven, 1985; Milgrom & Roberts,1990). Because of these linkages across prob-lems, we also conjecture that it will be rare tofind new forms of organizing that display nov-elty in the manner in which only one of the basic

problems of organizing is solved. More typically,we may expect to see frequently occurring clus-ters of solutions across forms of organizing; inthis sense, new forms of organizing may, in fact,be new bundles of old solutions. It is also likelythat organizations will have varying predispo-sitions and levels of aptitude at discovering andimplementing such recombinations (Galunic &Simon, 1998; Kogut & Zander, 1992).

DISCUSSION AND CONCLUSION

We discuss how our work adds to prior workon understanding organizations as complemen-tary bundles of elements, and we offer somedirection for applying our approach. We con-clude by revisiting the question of what makes anew form of organizing new and whether itnecessarily requires new theorizing.

Comparison with Prior Work onComplementarities Among OrganizationalElements

Organization theorists have offered many in-sightful and impactful analyses on commonlyoccurring clusters of organizational attributes.Configuration theories (e.g., Doty, Glick, & Hu-ber, 1993; Fiss, 2007; Meyer, Tsui, & Hinings, 1993)are an important instance. Others include soci-ological theories of organizational forms (e.g.,DiMaggio & Powell, 1991) or theories of organi-zational adaptation on rugged landscapes (e.g.,Levinthal, 1997). We explain below why our uni-versal problems of organizing approach canadd something new to this already extensiveliterature.Configurational approaches to organizations

emphasize commonly occurring clusters of at-tributes of organizational strategies, structures,and processes (Ketchen, Thomas, & Snow, 1993:1278). Such clusters of attributes are either em-pirically determined (e.g., Miller & Friesen, 1984)or theoretically specified (Mintzberg, 1979). Con-sider two important instances. First, in the Milesand Snow (1978) typology, prospectors, defend-ers, and analyzers face different environmentsand, consequently, have differing objectives;Miles and Snow argue that this results in sys-tematic differences along key organizational di-mensions, such as formalization, centralization,size, and vertical and horizontal differentiation.Second, Mintzbergs typology specifies ideal


types of organizationssimple structure, ma-chine bureaucracy, professional bureaucracy,divisionalized form, and adhocracywhich alsorepresent clusters of scores on key organiza-tional dimensions. Unlike Miles and Snow, Mintz-bergs justification for these particular clustersrests on internal coherence rather than externalalignment (see Doty et al., 1993, for a comparativetest of these theories). Nonetheless, the centraltheoretical claims in configurational theories takethe form of explanations of (1) why only certainclusters of attributesconfigurationsshould beobserved and (2) when some are relativelymore orless likely to be observed than others.In our view, the clusters of attributes identified

in configurational theories embody regions in thespace of solutions to a set of underlying universalorganizational problems. Our focus has been onidentifying these problems. We recognize we arenot alone in making such an attempt. In theirinfluential work Miles and Snow (1978: 2223) ex-plained organizational configurations as clustersof solutions to the universal problems organiza-tions confront when adapting to their environ-mentsthe entrepreneurial problem (a choice ofproduct market domain), the engineering problem(the choice of a technology), and the administra-tive problem (reducing uncertainty and creatinginnovation). Quite explicitly, these authors usedorganizational-level constructs, to the pointwhere entire organizations can be portrayed asintegrated wholes in dynamic interaction withtheir environments (1978: 30).Our approach focuses on a different set of

basic problems of organizing that an organiza-tion needs to solve to exist as an organiza-tiona multiagent system with goals. The focustherefore has been on how the actions of indi-vidual agents are aggregated into organiza-tional-level objectives. To do so we argued thatthe four basic problems are task division, taskallocation, information provision, and rewardprovision. Looking at organizations through thetheoretical lenses provided by these two differ-ent problem spaces yields different insights.Even though different organizational configura-tions may embody different solutions to theadaptive problems of entrepreneurship, engi-neering, and administration, they must eachhave solved the underlying problems of divisionof labor and integration of effort, possibly evenin similar ways. For this reason, every new formof organizing is likely to be a distinct configura-

tion, but the converse is not necessarily true. Forinstance, prospectors and defenders may differin their degrees of centralization, formalization,and structural differentiation, but we would hes-itate to think of them as distinct forms of orga-nizing as long as both resemble each other inhow they divide and allocate tasks and informand reward their employees.Conversely, comparing the Encyclopaedia Bri-

tannicawith Wikipedia on the dimensions of cen-tralization, formalization, and structural differen-tiation, or even noting that they seem to embodydifferent solutions to the entrepreneurial, engi-neering, and administrative problems, would notexpose qualitative differences in how they solvethe basic problems of organizingin how theydivide and allocate tasks and reward and informagents. Thus, we conjecture that configurationalapproaches may be more useful than our ap-proach for understanding variations within a formof organizing in terms of how the problems ofadaptation are solved, whereas our approachmaybe more useful for understanding differencesacross forms of organizing in how division of laborand integration of effort are accomplished.We also recognize that institutional sociologists

have offered useful perspectives on the emer-gence of new organizational forms (e.g.,Lounsbury, 2002; Rao & Haveman, 1997). However,we emphasize that what we call a form of orga-nizing is not necessarily the same as an organi-zational form. The latter is used in the sociologi-cal literature on the categorization processes bywhich audiences confer legitimacy on novel cate-gories of organizations. An organizational form istreated as a bundle of attributes, whose presenceor absence influences the process of legitimationby the relevant audience (Hannan & Freeman,1977; Hsu & Hannan, 2005). These features couldalso include the goals of the organization, and,indeed, in empirical practice that is how organi-zational forms are often distinguished. Thus, theconception of an organizational form is primarilyin terms of an object of social evaluation; how theorganization works is of secondary, if any, impor-tance in this conceptualization. In other words,when assessing novelty in organizational forms,the goals are not necessarily held constant; perour analysis, assessing novelty in forms of orga-nizing would be impossible without holding thegoals constant.To take a specific instance, the plethora of

new organizational forms that Lounsbury (2002)


documented in response to a change towardmarket-based logics in the U.S. financial sectordo not seem to display novelty in forms of orga-nizing. Conversely, as long as the novelty withinthe new forms of organizing we describe in ourarticle remains invisible to the audience of eval-uators, they will not be qualified as a new orga-nizational form (Hsu & Hannan, 2005).The literature on organizational adaptation

on rugged NK landscapes has also taken theexistence of complementarities in elementsof organization as a central premise (e.g.,Levinthal, 1997). This approach focuses on de-scribing boundedly rational search processes ina problem space that is characterized bycomplementarities; the key insights that seemrelevant to our conceptualization of organizingas problem solving are (1) because of comple-mentarities between choices on the four prob-lem dimensions, there are likely to be manylocal peaks; (2) for the search process to avoidentrapment on a local peak, it must featuremechanisms for both expanding search effortsbeyond the immediate neighborhood (explora-tion) and preserving a good solution (exploita-tion); and (3) complementarities in the searchspace (i.e., the degree of ruggedness of the land-scape) influence the relative importance ofthese mechanisms (e.g., Rivkin & Siggelkow,2003, 2007; Siggelkow & Levinthal, 2003; Sig-gelkow & Rivkin, 2005). Thus, we would alsoexpect forms of organizing to stabilize aroundclusters of complementary solutions, with thediscovery of new forms of organizing beingfairly rare events triggered by organizationsand individuals with strong motivations to ex-plore. While these are extremely valuable in-sights into the process of search for new forms oforganizing in a problem space characterized bycomplementarities, where those complementari-ties come from and the particular dimensionsinvolved are not the focus for the literature us-ing NK models, whereas they are for us.Thus, while NK modelbased approaches pre-

sume the existence of complementarities, theywould not specifically predict the content of par-ticular bundles of complementary solutionsforinstance, that transparency of task architec-tures, self-selection into tasks, the reliance onintrinsic motivation, and virtual collaborationare likely to be found together, as appears to bethe case in a large number of IT-enabled newforms of organizing.

Directions for Further Research

In our view, the universal problems of orga-nizing approach can be very generative, withmany implications. The first of theseunder-standing novelty in forms of organizingiswhat we have developed in detail in this article,but there are others that also seem fruitful op-portunities other researchers could exploit. Forinstance, the approach we have developed canbe used to understand similarity in organizing;the universal problems approach may also helpscholars classify organizations in novel, fruitfulways (for other approaches to classification, seealso McKelvey, 1982). Rather than cluster orga-nizations by industry, size, or for-profit versusnot-for-profit stauts, or on attributes like central-ization and formalization, scholars could in-stead discover underlying similarities based onhow the universal problems are solved. For in-stance, such an approach may reveal that disas-ter relief organizations have a lot more in com-mon with open source software or Wikipediathan hitherto suspected because of the potentialvalue from using self-selection to solve the taskallocation problem. Conversely, one may beable to design new forms of organizing by bor-rowing from one domain and inserting into an-other; for instance, allowing self-selection intotasks may improve motivation in and contribu-tions to voluntary organizations, improvingpeer-to-peer communication may reduce reli-ance on command and control systems in mili-tary contexts, and so on. Such designed inter-ventions could be rigorously tested throughcontrolled field experiments.Another possible application is the study of

organizing as a problem-solving process. SinceNewell and Simons (1972) work, the process ofhuman problem solving has been modeled as asearch process through a problem space. Anextensive body of literature in cognitive psy-chology suggests that the search behavior of aboundedly rational agent in this sort of problemspace is characterized by a few well-knownsearch strategies, such as local hill climbing,which is the search for local improvements(through change in one dimension at a time)from the current position; analogical reason-ing, where one imports solutions from one prob-lem space into another; and divide and con-quer, where the problem space is partitioned


and searched independently (Dunbar, 1995;Fischer et al., 2012).In order to move beyond the general features

of human problem solving to focus on organiz-ing as a particular kind of problem solving(rather than problem solving in the abstract, asoften represented in NK models), we must firstdefine the dimensions of the problem spacethrough which the process of search progresses(Newell & Simon, 1972). The four universal prob-lems we identify provide the natural set of di-mensions for the problem space of organizing.Thus armed with this conceptualization, schol-ars could study, for instance, how choices aboutorganizing vary when made in a decentralizedmanner (as in self-organizing groups) versusthrough a centralized designer (as in traditionalbureaucracies). Another interesting area ofstudy could be comparing how the problem solv-ing differs when the problem solver is an expertversus a novicewhich has emerged as a stan-dard approach in problem-solving research todocument the existence and nature of expertisein a domain (Ericsson, 2006). Finally, one couldstudy what happens when the solution space istransformed through exogenous shocks alongany one dimension (such as the recent rapiddevelopments in distributed collaboration tech-nologies) to test our conjecturethat no signifi-cant change should occur until complementarysolutions are found on other problem dimen-sions. A related issue pertains to whether denovo entrants or established organizations aremore likely to embrace new forms of organiz-ingwhether, in fact, the Oticons of the worldare rarer than the Wikipedias.Such analyses only become tractable once

we have dimensions on which to compare theproblem-solving efforts and resulting solu-tions, which our approach provides. Put differ-ently, the power of using these four dimen-sions of the problem space for organizing liesin the fact that they are specific to organizingbut not to any particular organization. Usingthem, our analysis makes clearer what consti-tutes novelty in forms of organizing and whenthis may require new theorizing. We hope thatthe approach we depict can help enrich thediscussion of new forms of organizing, beyondthe rallying cry that we need new theory tounderstand them.

Conclusion

Our goal has been to develop a systematicapproach to answering the question of whatmakes a new form of organizing new. We pro-posed that a new form of organizing is a noveland unique set of solutions to the universalproblems of division of labor and integration ofeffort. We also decomposed these basic prob-lems into four subproblems that we think arefundamental to any form of organizing. Noveltyin a form of organizing arises when one or moreof these problems are solved in a novel man-nerat least relative to the class of organiza-tions with comparable goals (i.e., the solutionneed not be new to the world).However, for the same reason, new forms of

organizing do not automatically imply the needfor new theorizing since the solutions may existand be well known in other contexts. Thus, in thecase of each of the new forms of organizing wehave discussed (Table 1), our analysis suggeststhat (1) it is a new form of organizing relative tohow the same goal was being met traditionallyand (2) much of this novelty can be explained byexisting theories, but (3) fruitful avenues for newtheory development may nonetheless lie in under-standing the complementarities across the solu-tions to the problems of organizing that these newforms of organizing embody.

REFERENCES

Aldrich, H. 1979. Organizations and environments. Engle-wood Cliffs, NJ: Prentice Hall.

Baldwin, C. Y. 2010. When open architecture beats closed:The entrepreneurial use of architectural knowledge.http://www.hbs.edu/research/pdf/10-063.pdf; accessedAugust 31, 2011.

Baldwin, C. Y., & Clark, K. B. 2000. Design rules: The power ofmodularity. Cambridge, MA: MIT Press.

Baldwin, C. Y., & Clark, K. B. 2006. The architecture of par-ticipation: Does code architecture mitigate free riding inthe open source development model? Management Sci-ence, 52: 11161127.

Barzel, Y. 1982. Measurement cost and the organization ofmarkets. Journal of Law and Economics, 25: 2748.

Boudreau, K. J., & Lakhani, K. R. 2013. Using the crowd as aninnovation partner. Harvard Business Review, 91(4): 6069.

Burton, R. M., & Obel, B. 1984. Designing efficient organiza-tions: Modelling and experimentation. Amsterdam:North-Holland.

Business Week. 2004. Linus Torvalds benevolent dictator-ship. http://www.businessweek.com/technology/content/


aug2004/tc20040818_1593.htm: August 17. Accessed January7, 2008.

Camerer, C. F., & Knez, M. 1996. Coordination, organizationalboundaries and fads in business practices. Industrialand Corporate Change, 5: 89112.

Camerer, C. F., & Knez, M. 1997. Coordination in organiza-tions: A game-theoretic perspective. In Z. Shapira (Ed.),Organizational decision making: 158188. Cambridge:Cambridge University Press.

Coase, R. H. 1960. The problem of social cost. Journal of Lawand Economics, 3: 144.

Colfer, L., & Baldwin, C. Y. 2010. The mirroring hypothesis:Theory, evidence and exceptions. http://ssrn.com/abstract1539592; accessed February 6, 2011.

Cramton, C. D. 2001. The mutual knowledge problem and itsconsequences for dispersed collaboration. OrganizationScience, 12: 346371.

Daft, R. L., & Lewin, A. Y. 1990. Can organization studiesbegin to break out of the normal science straitjacket? Aneditorial essay. Organization Science, 1: 19.

Dahlander, L., & OMahony, S. 2011. Progressing to the cen-ter: Coordinating project work.Organization Science, 22:961979.

Davis, G. F., & Marquis, C. 2005. Prospects for organizationtheory in the early twenty-first century: Institutionalfields and mechanisms. Organization Science, 16: 332343.

DiMaggio, P., & Powell, W. W. 1991. The new institutionalismin organizational analysis. Chicago: University of Chi-cago Press.

Doty, D. H., Glick, W. H., & Huber, G. P. 1993. Fit, equifinality,and organizational effectiveness: A test of two configura-tional theories. Academy of Management Journal, 36:11961250.

Drazin, R., & Van de Ven, A. H. 1985. Alternative forms of fit incontingency theory. Administrative Science Quarterly,30: 514539.

Dunbar, K. 1995. How scientists really reason: Scientific rea-soning in real-world laboratories. In R. J. Sternberg &J. E. Davidson (Eds.), The nature of insight: 365395. Cam-bridge, MA: MIT Press.

Dunbar, R. L. M., & Starbuck, W. H. 2006. Learning to designorganizations and learning from designing them. Orga-nization Science, 17: 171178.

Ericsson, K. A. 2006. An introduction to the Cambridge hand-book of expertise and expert performance. In K. A. Eric-sson, N. Charness, P. J. Feltovich, & R. R. Hoffman (Eds.),The Cambridge handbook of expertise and expert per-formance: 320. New York: Cambridge University Press.

Etzioni, A. 1964. Modern organizations. Englewood Cliffs, NJ:Prentice Hall.

Faraj, S., & Sproull, L. 2000. Coordinating expertise in soft-ware development teams. Management Science, 46:15541568.

Felin, T., & Zenger, T. R. 2011. Information aggregation,matching and radical markethierarchy hybrids: Impli-

cations for the theory of the firm. Strategic Organization,9: 163173.

Fischer, A., Greiff, S., & Funke, J. 2012. The process of solvingcomplex problems. Journal of Problem Solving, 4: 1942.

Fiss, P. C. 2007. A set-theoretic approach to organizationalconfigurations. Academy of Management Review, 32:11801198.

Foss, N. J. 2003. Selective intervention and internal hybrids:Interpreting and learning from the rise and decline ofthe Oticon spaghetti organization. Organization Sci-ence, 14: 331349.

Galbraith, J. R. 1973. Designing complex organizations. Read-ing, MA: Addison-Wesley.

Galunic, D. C., & Simon, R. 1998. Resource recombinations inthe firm: Knowledge structures and the potential forSchumpeterian innovation. Strategic Management Jour-nal, 19: 11931201.

Gambardella, A., & Hall, B. H. 2006. Proprietary versus publicdomain licensing of software and research products.Research Policy, 35: 875892.

Gersick, C. J. G. 1991. Revolutionary change theories: A mul-tilevel exploration of the punctuated equilibrium para-digm. Academy of Management Review, 16: 1036.

Gibbons, R. 1998. Incentives in organizations. Journal of Eco-nomic Perspectives, 12(4): 115132.

Gorbotai, A. 2012. Aligning collective production with de-mand: Evidence from Wikipedia. Working paper, HaasSchool of Business, Berkeley, CA.

Greenwood, R., & Miller, D. 2010. Tackling design anew:Getting back to the heart of organizational theory. Acad-emy of Management Perspectives, 24(4): 7888.

Guinan, E. C., Boudreau, K. J., & Lakhani, K. R. 2013. Experi-ments in open innovation at Harvard Medical School.MIT Sloan Management Review, 54(3): 4552.

Gulati, R., Lawrence, P. R., & Puranam, P. 2005. Adaptation invertical relationships: Beyond incentive conflict. Strate-gic Management Journal, 26: 415440.

Hackman, J. R., & Oldham, G. R. 1976. Motivation through thedesign of work: Test of a theory. Organizational Behav-ior and Human Performance, 16: 250279.

Hannan, M. T., & Freeman, J. 1977. The population ecology oforganizations. American Journal of Sociology, 82: 929964.

Hardin, G. 1968. The tragedy of the commons. Science, 162:12431248.

Heath, C., & Staudenmayer, N. 2000. Coordination neglect:How lay theories of organizing complicate coordinationin organizations. Research in Organizational Behavior,22: 153191.

Hinds, P. J., & Kiesler, S. 2002. Distributed work. Cambridge,MA: MIT Press.

Hodgson, G. M., & Knudsen, T. 2010. Darwins conjecture: Thesearch for general principles of social and economicevolution. Chicago: University of Chicago Press.

Hsu, G., & Hannan, M. T. 2005. Identities, genres, and orga-nizational forms. Organization Science, 16: 474490.


Kahneman, D., & Tversky, A. 1979. Prospect theory: An anal-ysis of decision under risk. Econometrica, 47: 263292.

Ketchen, D. J., Jr., Thomas, J. B., & Snow, C. C. 1993. Organi-zational configurations and performance: A comparisonof theoretical approaches. Academy of ManagementJournal, 36: 12781313.

Kilduff, M., Mehra, A., & Dunn, M. B. 2011. From blue skyresearch to problem solving: A philosophy of sciencetheory of new knowledge production. Academy of Man-agement Review, 36: 297317.

Kogut, B., & Zander, U. 1992. Knowledge of the firm, combi-native capabilities, and the replication of technology.Organization Science, 3: 383397.

Lakhani, K., & Wolf, B. 2005. Why hackers do what they do:Understanding motivation and effort in free/open sourcesoftware projects. In J. Feller, B. Fitzgerald, S. Hissam, &K. Lakhani (Eds.), Perspectives on free and open sourcesoftware: 321. Cambridge, MA: MIT Press.

Lawrence, P. R., & Lorsch, J. W. 1967. Organization and envi-ronment: Managing differentiation and integration. Bos-ton: Harvard University Press.

Lee, G. K., & Cole, R. E. 2003. From a firm-based to a commu-nity-based model of knowledge creation: The case of theLinux kernel development. Organization Science, 14:633649.

Levinthal, D. A. 1997. Adaptation on rugged landscapes.Management Science, 43: 934950.

Lounsbury, M. 2002. Institutional transformation and statusmobility: The professionalization of the field of finance.Academy of Management Journal, 45: 255266.

Lovas, B., & Ghoshal, S. 2000. Strategy as guided evolution.Strategic Management Journal, 21: 875896.

MacCormack, A., Baldwin, C., & Rusnak, J. 2012. Exploringthe duality between product and organizational archi-tectures: A test of the mirroring hypothesis. ResearchPolicy, 41: 13091324.

MacCormack, A. D., Rusnak, J., & Baldwin, C. Y. 2006. Explor-ing the structure of complex software designs: An em-pirical study of open source and proprietary code. Man-agement Science, 52: 10151030.

Madsen, P., Desai, V., Roberts, K., &Wong, D. 2006. Mitigatinghazards through continuing design: The birth and evo-lution of a pediatric intensive care unit. OrganizationScience, 17: 239248.

March, J. G., & Simon, H. A. 1958. Organizations. New York:Wiley.

March, J. G., & Simon, H. A. 1993. Introduction to the secondedition. In J. G. March & H. A. Simon (Eds.), Organiza-tions (2nd ed.): 119. New York: Wiley.

McKelvey, B. 1982. Organizational systematics: Taxonomy,evolution, classification. Berkeley: University of Califor-nia Press.

Meyer, A. D., Tsui, A. S., & Hinings, C. R. 1993. Configuration-al approaches to organizational analysis. Academy ofManagement Journal, 36: 11751195.

Miles, R. E., & Snow, C. C. 1978. Organizational strategy,

structure, and process (2003 reprinted ed.). New York:McGraw-Hill.

Milgrom, P. R., & Roberts, B. M. 1990. The economics of mod-ern manufacturing technology, strategy and organiza-tion. American Economic Review, 80: 511528.

Milgrom, P. R., & Roberts, B. M. 1995. Complementarities andfit: Strategy, structure and organizational change inmanufacturing. Journal of Accounting and Economics,19: 179208.

Miller, D., & Friesen, P. H. 1984. Organizations: A quantumview. Englewood Cliffs, NJ: Prentice Hall.

Miller, D., Greenwood, R., & Prakash, R. 2009. What hap-pened to organization theory? Journal of ManagementInquiry, 18: 273279.

Mintzberg, H. 1979. The structuring of organizations: A syn-thesis of the research. Englewood Cliffs, NJ: PrenticeHall.

Moreland, R. L., & Argote, L. 2003. Transactive memory indynamic organizations. In R. S. Peterson & E. A. Mannix(Eds.), Leading and managing people in the dynamicorganization: 135162. Mahwah, NJ: Lawrence ErlbaumAssociates.

Newell, A., & Simon, H. A. 1972. Human problem solving.Englewood Cliffs, NJ: Prentice Hall.

OKeefe, B., & Lovas, B. 2002. Oticon A/S. London: LondonBusiness School.

Olson, M. 1971. The logic of collective action: Public goodsand the theory of groups (2nd ed.). Cambridge, MA: Har-vard University Press.

OMahony, S. 2003. Guarding the commons: How communitymanaged software projects protect their work. ResearchPolicy, 32: 11791198.

Ostrom, E. 1990. Governing the commons: The evolution ofinstitutions for collective action. Cambridge: CambridgeUniversity Press.

Porter, L. W., Lawler, E. E. I., & Hackman, J. R. 1975. Behaviorin organizations. New York: McGraw-Hill.

Prendergast, C. 1999. The provision of incentives in firms.Journal of Economic Literature, 37: 763.

Rao, H., & Haveman, H. A. 1997. Structuring a theory of moralsentiments: Institutional and organizational coevolutionin the early thrift industry. American Journal of Sociol-ogy, 102: 16061651.

Raymond, E. S. 1999. The cathedral and the bazaar: Musingson Linux and open source by an accidental revolution-ary. Sebastopol, CA: OReilly.

Rivkin, J. W., & Siggelkow, N. 2003. Balancing search andstability: Interdependencies among elements of organi-zational design. Management Science, 49: 290311.

Rivkin, J. W., & Siggelkow, N. 2007. Patterned interactions incomplex systems: Implications for exploration.Manage-ment Science, 53: 10681085.

Rosenkopf, L., Metiu, A., & George, V. P. 2001. From the bottomup? Technical committee activity and alliance formation.Administrative Science Quarterly, 46: 748772.


Schelling, T. C. 1960. The strategy of conflict. Cambridge,MA: Harvard University Press.

Schoonhoven, C. B., Meyer, A. D., & Walsh, J. P. 2005a. Push-ing back the frontiers of organization science. Organi-zation Science, 16: 327331.

Schoonhoven, C. B., Meyer, A. D., &Walsh, J. P. 2005b. Movingbeyond the frontiers of organization science. Organiza-tion Science, 16: 453455.

Scott, W. R. 1998. Organizations: Rational, natural, and opensystems (4th ed.). Upper Saddle River, NJ: Prentic-Hall.

Selznick, P. 1957. Leadership in administration. New York:Harper & Row.

Shah, S. 2006. Motivation, governance, and the viability ofhybrid forms in open source software development.Management Science, 52: 10001014.

Siggelkow, N., & Levinthal, D. A. 2003. Temporarily divide toconquer: Centralized, decentralized, and reintegratedorganizational approaches to exploration and adapta-tion. Organization Science, 14: 650669.

Siggelkow, N., & Rivkin, J. W. 2005. Speed and search: De-signing organizations for turbulence and complexity.Organization Science, 16: 101122.

Simon, H. A. 1947. Administrative behavior. New York: Mac-millan.

Simon, H. A. 1951. A formal theory of the employment rela-tionship. Econometrica, 19: 293305.

Simon, H. A. 1953. Notes on the observation and measure-ment of political power. Journal of Politics, 15: 500516.

Simon, H. A. 1955. A behavioral model of rational choice.Quarterly Journal of Economics, 69: 99118.

Smith, A. 1776. An inquiry into the nature and causes of thewealth of nations. London: W. Strahan and T. Cadell.

Srikanth, K., & Puranam, P. 2011. Integrating distributedwork: Comparing task design, communication, and tacitcoordination mechanisms. Strategic Management Jour-nal, 32: 849875.

Stinchcombe, A. L. 1965. Social structure and organizations.In J. G. March (Ed.), Handbook of organizations: 142193.Chicago: Rand McNally.

TEDGlobal 2005. Jimmy Wales on the birth of Wikipedia (TEDtalk). http://www.ted.com/talks//eng/jimmy_wales_on_the_birth_of_wikipedia.html; accessed August 31, 2010.

Thompson, J. D. 1967. Organizations in action: Social sciencebases of administrative theory. New York: McGraw-Hill.

Titmuss, R. M. 1970. The gift relationship: From human bloodto social policy. London: Allen & Unwin.

von Hippel, E. 1990. Task partitioning: An innovation processvariable. Research Policy, 19: 407418.

von Hippel, E., & von Krogh, G. 2003. Open source software andthe private-collective innovation model: Issues for orga-nization science. Organization Science, 14: 209233.

von Krogh, G., Haefliger, S., Spaeth, S., & Wallin, M. W. 2012.Carrots and rainbows: Motivation and social practice inopen source software development. MIS Quarterly, 36:649676.

von Krogh, G., Spaeth, S., & Lakhani, K. R. 2003. Community,joining, and specialization in open source software in-novation: A case study. Research Policy, 32: 12171241.

von Krogh, G., & von Hippel, E. 2006. The promise of research onopen source software. Management Science, 52: 975983.

Walsh, J. P., Meyer, A. D., & Schoonhoven, C. B. 2006. A futurefor organization theory: Living in and living with chang-ing organizations. Organization Science, 17: 657671.

Wasko, M. M., & Faraj, S. 2005. Why should I share? Exam-ining social capital and knowledge contribution in elec-tronic networks of practice. MIS Quarterly, 29: 3557.

Weber, M. 1922. Wirtschaft und Gesellschaft. Tbingen, Ger-many: Mohr.

Weick, K. E. 1969. The psychology of organizing. Reading,MA: Addison-Wesley.

Weick, K. E. 1974. Amendments to organizational theorizing.Academy of Management Journal, 17: 487502.

Williamson, O. E. 1975. Markets and hierarchies: Analysisand antitrust implications. New York: Free Press.

Williamson, O. E. 1985. The economic institutions of capital-ism. New York: Free Press.

Zammuto, R. F., Griffith, T. L., Majchrzak, A., Dougherty, D. J., &Faraj, S. 2007. Information technology and the changingfabric of organization. Organization Science, 18: 749762.

Zenger, T. R., & Hesterly, W. S. 1997. The disaggregation ofcorporations: Selective intervention, high-powered incen-tives, and molecular units. Org

Documents

AMR - What's New About New Forms