Social Media, Knowledge Sharing, and Innovation: Toward a ... · shape knowledge sharing in...

INFORMS

holdsco

pyrightto

this

artic

lean

ddistrib

uted

this

copy

asaco

urtesy

totheau

thor(s).

Add

ition

alinform

ation,

includ

ingrig

htsan

dpe

rmission

policies,

isav

ailableat

https://p

ubso

nline.inform

s.org/.

Information Systems ResearchVol. 25, No. 4, December 2014, pp. 796–816ISSN 1047-7047 (print) � ISSN 1526-5536 (online) http://dx.doi.org/10.1287/isre.2014.0536

© 2014 INFORMS

Social Media, Knowledge Sharing, and Innovation:Toward a Theory of Communication Visibility

Paul M. LeonardiTechnology Management Program, University of California, Santa Barbara,

Santa Barbara, California 93106, [email protected]

This paper offers a theory of communication visibility based on a field study of the implementation of anew enterprise social networking site in a large financial services organization. The emerging theory sug-

gests that once invisible communication occurring between others in the organization becomes visible for thirdparties, those third parties could improve their metaknowledge (i.e., knowledge of who knows what and whoknows whom). Communication visibility, in this case made possible by the enterprise social networking site,leads to enhanced awareness of who knows what and whom through two interrelated mechanisms: messagetransparency and network translucence. Seeing the contents of other’s messages helps third-party observersmake inferences about coworkers’ knowledge. Tangentially, seeing the structure of coworkers’ communicationnetworks helps third-party observers make inferences about those with whom coworkers regularly communi-cate. The emerging theory further suggests that enhanced metaknowledge can lead to more innovative productsand services and less knowledge duplication if employees learn to work in new ways. By learning vicariouslyrather than through experience, workers can more effectively recombine existing ideas into new ideas and avoidduplicating work. Moreover, they can begin to proactively aggregate information perceived daily rather thanengaging in reactive search after confronting a problem. I discuss the important implications of this emergingtheory of communication visibility for work in the knowledge economy.

Keywords : social networking; innovation; knowledge sharing; metaknowledge; computer-mediatedcommunication and collaboration; knowledge management; ethnographic research

History : Chris Forman, John Leslie King, Kalle Lyytinen, Senior Editors; Jonny Holmström, Associate Editor.This paper was received July 1, 2013, and was with the author 4 months for 1 revision. Published online inArticles in Advance October 3, 2014.

IntroductionMost of the work conducted in today’s knowledge-intensive organizations is largely invisible. Workerssit at computers typing reports, performing analyses,writing copy, and performing other tasks that are dif-ficult for observers to discern. As scholars have noted,work in the postindustrial economy is increasinglydisembodied such that there are few physical mani-festations of activity for others to observe (Suchman2007). Work is made even more invisible as organiza-tions increasingly rationalize it into smaller task units,spreading them across workers in various groups anddepartments, perhaps in different geographic loca-tions. Even if there were some physical manifesta-tion of work, there would be few relevant cowork-ers around to see it done. Consequently, Nardi andEngeström (1999, p. 2) suggest that in a knowledgeeconomy “work is, in a sense, always invisible toeveryone but its own practitioners.”

Work invisibility can have a number of negativeconsequences for organizations. Research has shownthat when people cannot see what others do, orwhen, with whom, and how they do it, there is

less interpersonal trust among coworkers (Cramtonet al. 2007). In addition, coordination suffers (Dey andde Guzman 2006), work duplication occurs (Lapréand Van Wassenhove 2001), and product and pro-cess innovation are less likely (Majchrzak et al. 2004).Because of these negative consequences, scholars havesearched for technologies that might provide workerssome visibility into the work of their colleagues. Tech-nologies that make workflows (Kinnaird et al. 2012),activity status (Dourish and Bellotti 1992), and evenworker location (Erickson 2010) visible to others pro-vide a means to learn about the work of colleaguesand, hopefully, to avoid some of the problems thatoccur when work is largely invisible to others.

The past few decades have seen an increase in theuse of technology to make communication visible asa work activity. Workplace communication by tele-phone or face-to-face encounters is largely invisibleto all but the parties involved or those who maybe in proximity. The number of people who mightoverhear an interpersonal conversation (even spokenloudly) is small relative to the size of most teams,departments, and organizations. More contemporary

796

INFORMS

holdsco

pyrightto

this

artic

lean

ddistrib

uted

this

copy

asaco

urtesy

totheau

thor(s).

Add

ition

alinform

ation,

includ

ingrig

htsan

dpe

rmission

policies,

isav

ailableat

https://p

ubso

nline.inform

s.org/.

Leonardi: Social Media, Knowledge Sharing, and InnovationInformation Systems Research 25(4), pp. 796–816, © 2014 INFORMS 797

communication (e.g., email and instant messaging) iseasily shared with others, and thus visible to a wideraudience. Collaboration tools, decision support soft-ware, and worker databases have further increasedcommunication visibility by loosening the require-ment to select a target audience through email car-bon copy features or the instant messaging forwardfeature. Instead, the sender can create a public com-munication platform that can be accessed by teammembers. Still even shared communication platformslimit visibility because membership in groupwareand other collaboration tools are normally limited tospecific team members. Shared databases and otherrepositories are often password protected and storedin a location unknown to most of the organization.

The introduction of social media tools, includingsocial networking sites, blogs, wikis, and microblogs,into organizational contexts continues a long trend ofmaking workplace communication visible. However,as Treem and Leonardi (2012) show in their detailedreview, social media are making routine communica-tions between coworkers even more visible to thirdparties than the many preceding communication tech-nologies. For example, the exchanges between twopeople on an enterprise social networking site oftenappear on the wall or newsfeed of a third party notinvolved in the communication (Hampton et al. 2011).That people can articulate their social networks andtag documents and images produced by coworkerswithin social networking sites gives outsiders furthervisibility into the communication partners of theirpeers (Keitzmann et al. 2011). As some commenta-tors suggest, social media are so quickening the paceby which technologies afford communication visibil-ity that we are entering an era of “hypervisbility”where anyone can easily see with ease what any otherperson said and to whom (Keen 2012).

This paper will explore how the increasing visibil-ity of communication might, as several researchershave suggested (Gibbs et al. 2013, Jarrahi and Sawyer2013, Leonardi et al. 2013, Treem and Leonardi 2012),shape knowledge sharing in organizations, and therole that these implications for knowledge sharingmight play in modern work. I conduct a qualitativestudy of the implementation of a new social network-ing site into a large financial services organization.This study takes advantage of a comparative analysisin which only one of two matched-sample groups inthe same organization was given access to the socialnetworking site for six months; the other group wasnot. Interviews with members of both groups beforeimplementation of the social networking site and sixmonths after its introduction provided a compara-tive basis for identifying the advantages the technol-ogy provided for staff work and the social dynamicsthat emerged as communication became more visible.

The findings demonstrate that the increased com-munication visibility associated with the social net-working site enabled users to become more aware ofwhat and whom their coworkers knew. This enhancedawareness allowed staff to avoid work duplicationand to develop more innovative ideas for productsand services. To realize these positive benefits, usershad to enact fundamental changes in the way theyworked. I conclude by discussing the implications ofthis emerging theory of communication visibility forresearch in information systems, management, andcommunication.

Theoretical BackgroundLearning About Knowledge ThroughCommunicationNumerous studies conducted over the past twodecades have shown that successful organizations cre-ate the conditions under which employees can effec-tively share knowledge with one another (Argoteet al. 2000, Hansen 1999, Tortoriello et al. 2012).Most research on organizational knowledge sharinghas focused on when, in the course of their work,researchers should look for new knowledge (Katilaand Chen 2008, March 1991) and what conditionsare likely to ease knowledge transfer (Reagans andMcEvily 2003, Szulanski and Jensen 2006). Clearly,timing and knowledge transfer issues are important.However, the decision on when to acquire knowledgeand how it can be applied to completion of tasks andproblem solving would likely be impossible in theabsence of accurate organizational metaknowledge.Organizational metaknowledge refers to knowledgeabout who knows what and who knows whom withinthe organization. Before workers can acquire instru-mental knowledge, they must know where to get itand how to access other workers, files, or databaseswhere that knowledge can be found.

Researchers who study transactive memory sys-tems have, so far, been among the organizational the-orists to explicitly consider the concept of metaknowl-edge. A transactive memory system is a collectivesystem that individuals in close relationships use toencode, store, and retrieve knowledge (Hollingshead1998a, Ren et al. 2006). A transactive memory sys-tem has two components. The first is a transac-tive memory, or as Ren and Argote (2011, p. 192)call it, “metaknowledge, which helps individuals toretrieve information from external sources.” The sec-ond component is a set of processes for updatingand using that metaknowledge. Wegner (1995) sug-gested three such processes related to the develop-ment and application of metaknowledge: directorydevelopment (producing accurate metaknowledge);

INFORMS

holdsco

pyrightto

this

artic

lean

ddistrib

uted

this

copy

asaco

urtesy

totheau

thor(s).

Add

ition

alinform

ation,

includ

ingrig

htsan

dpe

rmission

policies,

isav

ailableat

https://p

ubso

nline.inform

s.org/.

Leonardi: Social Media, Knowledge Sharing, and Innovation798 Information Systems Research 25(4), pp. 796–816, © 2014 INFORMS

information allocation (new knowledge communi-cated to the person whose expertise will facilitateits storage); and retrieval coordination (knowledgeon any topic retrieved based on knowledge of whoknows what in the system). A transactive memorysystem allows each group member to share accurateorganizational metaknowledge. They can then routenew knowledge to those who can best store it, andwho can teach others in times of need (Brandon andHollingshead 2004).

Despite the central role that accurate metaknowl-edge plays in effective transactive memory systems(without it, coordination among members of the sys-tem would be difficult [Austin 2003, Yuan et al.2007]), most researchers have focused on informa-tion allocation processes (e.g., Jackson and Klobas2008, Yuan et al. 2007) and retrieval coordination (e.g.,Hollingshead 1998b, Jarvenpaa and Majchrzak 2008)at the expense of understanding how accurate meta-knowledge is achieved (see Ren and Argote 2011).Instead, scholars have largely inferred the processesby which accurate metaknowledge is developed andshared among members of a group. For example,Liang et al. (1995) found that group members whounderwent training together had more accurate meta-knowledge that was common to the team than didgroup members who were trained individually. Otherstudies have shown that groups constructed of mem-bers who have had the occasion to see each otherdirectly performing work tasks (Heald et al. 1998) andof members who had previously worked jointly onprojects (Reagans et al. 2005) tend to have more accu-rate and more widely shared metaknowledge. Theinferences drawn from these studies and others likethem is that seeing each other learning to conducttasks and having experience seeing others actuallyconducting tasks leads to more accurate and morewidely shared metaknowledge.

The assumption that workers will develop accu-rate and shared metaknowledge naturally throughdirect observation of work and interaction becomesless tenable when one raises the level of analysisfrom the group to the organization. The departmen-talized structure of most organizations reduces thechances that people from different units will worktogether, even though they may have complemen-tary knowledge. Also, the geographic dispersion ofworkers in different departments and business unitsreduces opportunities to see what others do whenthey work, and with whom. More commonly, work-ers learn by talking with others or hearing second-hand reports about how tasks were conducted andwith whom. In his initial theoretical formulation ofthe transactive memory perspective, Wegner (1987)acknowledged that, in real world settings (versus thelaboratory) talking with others about their tasks and

communication partners might be a primary waythrough which accurate metaknowledge was devel-oped: “The history of conversation about who hasdone what and heard what and been where and stud-ied what and with whom and under what circum-stance could be exceedingly rich, and so allows mem-bers to discern with much greater precision just whois expert in each of a variety of information domains”(p. 191). Recently, Palazzolo et al. (2006, p. 245) sug-gested that informal “water cooler communication”in which workers talked about their tasks and com-munication partners could facilitate development of atransactive memory.

There are a number of ways in which workersmight learn what and whom others know. The dis-tinction between experiential and vicarious learningoften discussed in the literature on organizationallearning demarcates two distinct strategies (Gioia andManz 1985, Kim and Miner 2007). A person canlearn experientially by communicating directly witha coworker, i.e., by asking questions and listeningto answers. Workers can also acquire metaknowl-edge vicariously by watching others communicate.Through direct communication a worker can learnwhat and whom coworkers know even if that wasnot the explicit goal of the communication. Sometimesthe focal actor is looking for specific, necessary instru-mental knowledge from coworkers and, for this rea-son, asks them to recount tasks they have conductedor the people they know (March 1991). Yet workerscan also learn through talking with coworkers thoughthey may have no specific or immediate need for theknowledge they hope to obtain (Newell et al. 2009).

Like experiential learning from active communica-tion, vicarious learning by watching the communica-tion of others can occur when attention is focusedor divided. Some third-party observation of commu-nication occurring among others is focused in thesense that the focal actor is looking to find specific,needed instrumental knowledge from targets and, forthis reason, watches to see what people are sayingto others, to whom they are talking, and who theymention they have talked with before (Liebeskind1996). Workers can also be exposed to communica-tions between others even when they are not focusedon trying to learn anything. In this way, they dividetheir attention among many communication eventsand file vicariously learned metaknowledge for futureuse (Weick 1995).

To the extent that the organizational environ-ment makes it difficult for workers to develop accu-rate metaknowledge by observing what others doand with whom, it is unlikely that the environ-ment will foster experiential learning through direct,active communication. Studies of co-located cross-functional teams and geographically distributed func-tional teams find that the departmental, temporal,

INFORMS

holdsco

pyrightto

this

artic

lean

ddistrib

uted

this

copy

asaco

urtesy

totheau

thor(s).

Add

ition

alinform

ation,

includ

ingrig

htsan

dpe

rmission

policies,

isav

ailableat

https://p

ubso

nline.inform

s.org/.


and linguistic boundaries between workers make itdifficult for them to talk directly with each other(Cummings et al. 2009). Other studies show thatactive communication among coworkers who are noton teams, but who work in the same organizationand do not sit near each other is relatively nonex-istent (Allen 1977). Consequently, when consideringthe development of metaknowledge across the orga-nization, as opposed to on small teams characterizedby high degrees of task interdependence, experien-tial learning through direct communication with oth-ers, whether one’s attention is focused or divided,may not be much more effective than direct observa-tion of others’ work. However, given recent advancesin communication technologies within organizations,improvements in the accuracy of metaknowledgethrough vicarious learning, which can occur by sim-ply watching or eavesdropping on others communica-tion, may be not only feasible, but also effective (Renand Argote 2011).

Enterprise Social Networking Technologies andCommunication VisibilityMany organizations, large and small, are beginning touse enterprise social networking sites. According toLeonardi et al. (2013, p. 2), enterprise social network-ing technologies allow workers to: (1) communicatemessages with specific coworkers or broadcast mes-sages to everyone in the organization; (2) explicitlyindicate or implicitly reveal particular coworkers ascommunication partners; (3) post, edit, and sort textand files linked to themselves or others; and (4) viewthe messages, connections, text, and files communi-cated, posted, edited, and sorted by others in the orga-nization at any time. Significantly, enterprise socialnetworking sites also provide a forum for public com-munication among employees (DiMicco et al. 2008).When two people share messages with one another,those messages appear on the communicators’ per-sonal profile pages, or “news feeds.” On many socialnetworking sites sold specifically for enterprise use,or enterprise applications that are built in-house, com-munications between coworkers can be seen by all oftheir respective contacts, and, if the appropriate set-tings are applied, by everyone in the organization.

The little empirical research on the use of enterprisesocial networking sites suggests that, unlike publicsocial networking sites such as myspace or Facebookwhere a user’s online connections are strongly cor-related with his or her offline social networks (seefor example, Lampe et al. 2006), employees who useenterprise social networking sites tend to maintainconnections with coworkers whom they do not knowand with whom they do not regularly interact offline

(DiMicco et al. 2008). Given the access to others’ com-munications they provide, enterprise social network-ing sites may enable vicarious learning through pas-sive exposure to communications between others thatextend observers’ reach beyond their immediate workgroup members.

If enterprise social networking sites are indeed use-ful in helping workers to develop broader and moreaccurate metaknowledge through vicarious engage-ment in others’ communications than they would oth-erwise develop through direct experiential involve-ment in communication with others, they are likelyto do so only very slowly. It would be unlikelythat a routine communication between two cowork-ers occurring on an enterprise social networking sitewould contain a concise description of someone’sknowledge that would be useful at some other timeby a casual observer. Instead, it is more probable thatroutine communications seen by third-party observerscontain some bits of information that can only beturned into metaknowledge when they are assembledwith other bits of information from different observedcommunications. Even if this were the case, the mech-anisms through which communication on an enter-prise social networking site would become visible tothird-party observers are unclear. Also, the kinds ofbehavioral or work-based changes necessary to takeadvantage of this communication visibility, and thegoals of that visibility, are unclear. To explore theseissues, I turn to the study described below.

MethodsResearch Setting and DesignThe data for this study were collected at a largefinancial services firm headquartered in the Midwest-ern United States. American Financial (a pseudonym,as are all names in this article) is a major directbanking and payment services company. In late 2010,the company’s Director of Internal Communicationbegan working with the external software vendor,Jive, to customize an enterprise social networkingsite for internal communication among employees.The social networking site, called “A-Life” (short for“American Financial Life”), seemed nearly identicalto publicly available social networking sites, such asmyspace and Facebook. It contained profile pages,news feeds, and algorithms for suggesting new con-tacts. The default settings for the site, which wereunchanged, allowed anyone who used it to view any-one else’s profile and to see communications betweenany other users on the site. The site also containeda shared document repository where items could betagged and linked to a user’s profile page or newsfeed. The system used employment data to create alist of a user’s work group members and display that

INFORMS

holdsco

pyrightto

this

artic

lean

ddistrib

uted

this

copy

asaco

urtesy

totheau

thor(s).

Add

ition

alinform

ation,

includ

ingrig

htsan

dpe

rmission

policies,

isav

ailableat

https://p

ubso

nline.inform

s.org/.


list on his or her profile page. A-Life also allowedemployees to post queries (much like a discussionboard) and to blog about ideas.

In late 2011, the IT team responsible for A-Life’simplementation selected 20 groups, at random, fromacross the company to participate in a pilot studyto ensure the new technology was working prop-erly before it was rolled out across the entire com-pany. One of these groups was a management leader-ship program in the company’s Marketing Division.The program consisted of employees from variousdepartments within the Marketing Division. Employ-ees were selected for the program after a compet-itive admissions process during their first year ofemployment and remained in the program through-out their tenure at the company. Membership in theprogram provided employees with regular access tocompany executives, extensive professional develop-ment classes, and regularly scheduled workshops andspeaker events specifically designed for the program.In short, the program was a community of practicewithin the Marketing Division. At the time of thisstudy, there were 44 program members. Their tenurewith the company ranged from six months to 11years and spanned entry-level positions through vice-president.

Because the Marketing group had been selectedas a user community for the A-Life pilot study, buthad not yet seen or begun to use the technology, apre-/post-implementation research design could beused to assess what types of changes to work wereassociated with implementation and use of the socialnetworking site by comparing how the individualsworked before as opposed to after their use of the sys-tem. Explaining whether and how A-Life increasedcommunication visibility, and exploring changes inawareness and work practice over time required morethan a simple pre/post research design. I needed todemonstrate that the dynamics resulting from A-Lifeuse were not just attributable to the normal passageof time (Benbasat et al. 1987). To address this issue,we were permitted to examine the work of staff ina similar management leadership program group inthe Operations Division over the same time period.The Marketing and Operations groups representeda matched sample in that the two leadership pro-grams had nearly identical distributions of demo-graphic profiles as to age, gender, ethnicity, tenure atthe company, hierarchical level, and job performanceratings. There were 50 employees in the Operationsgroup. None of them communicated with informantsin the Marketing group. This reduced the possibilitythat changes occurring in one group might dependon or be affected by changes in the other. A-Life wasimplemented in Marketing in early January 2012.

Data Collection and AnalysisIn December 2011, semi-structured interviews wereconducted with 16 employees from Marketing and18 from the Operations leadership program groups(hereafter, simply Marketing and Operations) toexamine their normal patterns of communication andthe various technologies they used to share knowl-edge. Each interview lasted between 40 and 70 min-utes, with an average length of 50 minutes. Theinterviews followed the same semistandard protocolcomprised of four major sections. In the first sec-tion, informants were asked general questions abouttheir background and past work experience. Ques-tions such as “What do you do in your job?” and“What does it take to be successful in your currentrole?” were designed to solicit opinions on the skillsnecessary to do their work. In the second section, weasked specific questions about information and dataused at work. Questions such as, “What type of infor-mation or data do you think is most important forbeing a good (insert job title)?” and “Are some peopleat American Financial more knowledgeable in certainareas than others?” were meant to encourage infor-mants to discuss their needs for information and datato complete their work. The third section asked themabout the kinds of people they interacted with daily,how they interacted with them, and how they knewwhich coworkers to contact. Questions such as “Howdo you decide who you will talk with when you’reworking on a project?”; “What kinds of technolo-gies do you use to interact with others?”; and “Whatkinds of technologies do you use to learn aboutthe interactions occurring among your coworkers?”were intended to capture strategies for communica-tion, interaction, and learning. The fourth section con-tained specific questions about various technologiesused at home and at work, their general perceptionsof social media technologies, and their perceptions ofthe role of technology in successfully executing theirwork.

Six months after Marketing began using A-Life,a second round of interviews was conducted withthe same 16 employees from Marketing and 18 fromOperations. The same interview protocol used in sec-tion one was used again for all employees; how-ever, this round of interviews included questions onspecific similarities and differences in communicationand knowledge sharing since the last interview. Mar-keting participants were also asked specifically howthey used A-Life over the previous six months. Intotal, 68 interviews were conducted across the tworounds with both groups. With the consent of theinformants, all of the interviews were audio-recordedand later transcribed verbatim. In total, 473 pages ofsingle-spaced interview transcripts were produced toserve as the raw data for analysis in this study.

INFORMS

holdsco

pyrightto

this

artic

lean

ddistrib

uted

this

copy

asaco

urtesy

totheau

thor(s).

Add

ition

alinform

ation,

includ

ingrig

htsan

dpe

rmission

policies,

isav

ailableat

https://p

ubso

nline.inform

s.org/.


The logic underlying this research design is the abil-ity to isolate the effects of communication visibilityenabled by social media use in Marketing througha two-way comparison (e.g., Miles and Huberman1994). The diachronic comparison (i.e., comparisonover time) facilitated understanding of changes towork occurred in Marketing after the implementationof the social networking site. The synchronic com-parison (comparison with another group at the sametime) between Marketing and Operations both beforethe implementation of the technology and after sixmonths of use facilitated an understanding of whatchanges in Marketing resulted from enterprise socialnetworking site use as opposed to simply broaderorganizational changes or changes due to naturalincreases in learning amongst coworkers due to sixadditional months spent working together. In otherwords, comparing the experience of staff in Marketingand Operations at two points in time facilitated focuson the nature of changes in Marketing, the only groupto use the enterprise social networking site. Data anal-ysis proceeded through three sequential phases (seeTable 1).

The first phase of data analysis was designed touncover the mechanisms by which use of the socialnetworking technology made communication amongcoworkers visible. I began the analysis with the pro-cess of theoretical coding. Glaser (1978) suggests thattheoretical coding is a useful strategy for the ana-lyst who wishes to study a concept by examiningthe data. Accordingly, I began by examining the Mar-keting interview transcripts only from the interviewsconducted six months after A-Life was implemented.I identified each instance in which informants talkedabout vicarious exposure to coworkers’ communica-tions. I applied codes to each instance that indicatedwhat type of communication was made visible, howit was made visible, what the informant learned fromthe visible communication, and how that new knowl-edge was used in their work. In total, 49 distinctcodes were produced in these various categories. Thenext step of analysis was axial coding. This involvedreassembling the coded data in new ways by group-ing codes that were conceptually similar (Strauss andCorbin 1998). Axial coding resulted in the reclassi-fication into two main codes indicating what infor-mants became aware of due to increased communica-tion visibility (knowing what other people knew andknowing who other people knew), two main codesindicating how use of the technology prompted thisawareness (making it easy to see the content of mes-sages exchanged and making it easy to see who peo-ple sent messages to or indicated as a frequent con-tact), and four main codes indicating changes to theirwork that arose from this new awareness (expanded

their personal networks, avoided duplicating previ-ously completed work, finished projects faster, andcombined ideas from various sources).

The second data analysis phase was designed toconfirm that the codes about what workers learnedfrom using the new technology and the outcomesthat resulted from its use were specific to Market-ing at Time 2. Following the analytical inductionprocess (Glaser 1965), I examined all of the othertranscripts (Marketing before implementation of A-Life and Operations at both time periods) to ascer-tain whether the activities that Marketing partici-pants performed after using A-Life were similar toor different from the activities before A-Life andwhether the same was true for Operations informantsat either time. Specifically, I used my understandingof what Marketing participants learned through vicar-ious exposure to visible communication on the enter-prise social networking technology, and what they didwith this knowledge, to uncover whether this kind ofvicarious learning and knowledge use was discussedin any of the other interview transcripts. To verifythat Operations was an acceptable comparative set toMarketing, I compared the codes generated for thetwo groups from the interviews done before A-Lifewas implemented. The codes were indeed quite sim-ilar both in placement and frequency. This providedevidence that Operations represented a reliable com-parative set and that I could compare changes thatoccurred naturally over time in Operations to changesthat occurred over time in Marketing. I could thensafely argue that those changes in Marketing weredifferent than would be expected given the results inOperations.

To verify that the four changes to the way peo-ple in Marketing worked (expanded their personalnetworks, avoided duplicating previously completedwork, finished projects faster, and combined ideasfrom various sources) were confined, specifically, tothe period of use following the implementation ofthe social networking technology, I followed the sameprocess outlined earlier by comparing these codes tocodes generated in the other interview transcripts. Iused a conservative estimate to determine whetherchanges uncovered through this coding process werespecific to the Marketing group. Specifically, I soughtto determine whether discussions of these four out-comes were present in less than 5% of the inter-views conducted in Marketing before implementa-tion of the social networking tool, or in Operationsat either point in time. Two of the three outcomes(i.e., expanded their personal networks and finishedprojects faster) were mentioned in 12% of interviewsin Operations at Time 2. (They were excluded as find-ings, leaving only two main outcomes linked to theuse of the new technology in Marketing.)

INFORMS

holdsco

pyrightto

this

artic

lean

ddistrib

uted

this

copy

asaco

urtesy

totheau

thor(s).

Add

ition

alinform

ation,

includ

ingrig

htsan

dpe

rmission

policies,

isav

ailableat

https://p

ubso

nline.inform

s.org/.


Table 1 Phases of Data Analysis

Coding procedures Comparison procedures

Phase Goal Data Theoretical Axial Analytic induction 5% threshold analysis Finding

1 To uncovermechanisms bywhich socialnetworkingtechnologymakescommunicationamongcoworkersvisible

—16 interviewswith Marketingemployees aftersix months oftech use

—49 codesdescribing whattype ofcommunicationwas madevisible, how thatcommunicationwas madevisible, what thefocal individuallearned from thevisiblecommunication,and how theyused that newknowledge intheir work

—2 codesindicating whatinformantsbecame aware ofdue to increasedcommunicationvisibility

—2 codesindicating howuse of thetechnologyprompted thisawareness

—4 codesindicatingchanges to workthat resultedfrom newawareness

N.A. N.A. —Knowing whatcoworkersknew andknowing whocoworkersknew)

—Making iteasy to seethe content ofmessagesexchangedand making iteasy to seewhocoworkerssentmessages to

2 To confirm thatfindings fromPhase 1 (whatworkers learnedfrom using thenew technologyas well as theoutcomes thatarose from itsuse) werelimited to thosewho used newtechnology

—32 interviewswith Marketingemployees(16 before and16 afterimplementation)

—36 interviewswith Operationsemployees(18 before and18 afterimplementation)

N.A. N.A. —Comparedplacement andfrequency ofcodes given toMarketing andOperations frominterviewsbefore techimplementation

—Examined 4codes indicatingchanges to workthat arose fromnew awarenessfrom Marketingin Time 2 andcompared toappearance inall othertranscripts

—Avoidedduplicatingpreviouslycompletedwork andcombinedideas fromvarioussources

3 To learn whatkinds ofbehaviors wereneeded tochange work inways found inPhase 1 andverified inPhase 2

—16 interviewswith Marketingemployees aftersix months oftech use

—18 codesindicating newways of workingfollowingimplementationof new tech

—2 codesindicatingbehavioralchangesnecessary toachieve benefitsfound in Phases1 and 2

N.A. N.A. —Learningwithoutcontext andstoringknowledge forpossible lateruse

The final phase of the analysis was designed todetermine what kinds of behaviors were needed tochange work in ways uncovered in the previousphase. To execute this phase, I isolated all instances inwhich Marketing informants talked about new waysof working following the implementation of A-Life. Ithen repeated the steps summarized above—engagingin theoretical and axial coding to narrow and connectphenomena in the data and then comparing thosecodes to the other interview transcripts in an attemptto isolate these emerging findings to the work of thosewho used A-Life and, consequently, improved theirawareness by being exposed to visible communica-tions occurring amongst coworkers. Through this pro-cess, I identified two behavioral changes needed to

support new ways of working (i.e., learning withoutcontext and storing knowledge for future use).

I used Atlas.ti software to keep track of all the-oretical, axial, and selective codes in each round ofcoding described above. To verify the accuracy ofthese codes, two research assistants repeated the cod-ing procedure used by the author, using the detailedcodebook (i.e., descriptions of each code) created inthe Atlas.ti software. To ensure consistency, the authorinitially reviewed the codes applied by the two assis-tants after they had coded five randomly assignedtranscripts. This helped the researchers to reach a con-sensus on the proper interpretation and application ofeach code. The researchers then recoded the five ini-tial transcripts and all subsequent transcripts. At the

INFORMS

holdsco

pyrightto

this

artic

lean

ddistrib

uted

this

copy

asaco

urtesy

totheau

thor(s).

Add

ition

alinform

ation,

includ

ingrig

htsan

dpe

rmission

policies,

isav

ailableat

https://p

ubso

nline.inform

s.org/.


end of this process, a Cohen’s Kappa was calculatedto determine the extent of agreement between thethree coders (i.e., the author and the two researchers).The final Cohen’s Kappa, 0.87, was very acceptable(Fleiss 1971).

Next I present data primarily from the 16 Market-ing interviews conducted after implementation of theenterprise social networking technology. As describedabove, the additional 52 interviews that were ana-lyzed allowed me to say with certain confidence thatthe practices and changes that occurred within Mar-keting were indeed related to the way individualsin this group used the new enterprise social net-working site because they represented changes to theinformants’ own work before the implementation ofthe technology, as well as changes from the normaldynamics of work in Operations occurring during thesame six month window.

FindingsThe results of the analyses are presented in three sec-tions. First, I discuss how specific advantages cre-ated by use of the enterprise social networking siteenabled third-party observers to become aware of“who knows what” and “who knows whom” in theorganization. I then show how this newly held knowl-edge about what and whom others knew (i.e., meta-knowledge) facilitated two changes to the way mar-keting informants worked. First, they avoided taskduplication through more effective reuse of knowl-edge. This increased the speed at which they workedand allowed them to move on to tasks that addedmore value to the organization. Second, they recom-bined ideas culled from various coworkers into newideas that helped them solve problems faster andmore thoroughly. Finally, I show how the ability todevelop sufficiently accurate metaknowledge fromvisible communication that could lead to recombi-nant innovation and avoidance of task duplicationrequired at least two significant behavioral changes:Shifting from problemistic search to proactive aggre-gation of content and doing something else. I presentevidence for these claims below.

Communication Visibility and the Developmentof MetaknowledgeLike workers in many large organizations, employ-ees at American Financial typically had a good ideaof what their coworkers (team members, other staff,or those they saw daily) were doing. Yet their under-standing of other coworkers’ activities in differentdepartments (or geographic locations) was not exten-sive. As one Marketing informant observed, “This isa big company. I don’t really know what other peopledo outside of my team. I definitely don’t know whatthey do all across Marketing. My guess is that some

people do really cool things that would be really use-ful for my projects, but I just don’t know what that isor who those people are.”

The primary reason that most workers knew lit-tle about what their coworkers were doing is lim-ited communication. Marketing employees consis-tently indicated that they learned about coworkers’daily activities from exposure to their communica-tions. Because the actual conduct of one’s work wasinvisible to others, people typically learned what heor she did by either asking them directly about itor by seeing or hearing communications they hadwith others that revealed the nature of their tasks.For example, when asked during the first round ofinterviews with whom she typically interacted in theworkplace, one informant answered:

I typically interact with people I know somethingabout—know what they do. That mostly means I com-municate with people on my immediate team 0 0 0 I don’tever really see those people actually doing work. Imean I see them sitting at their computers, but I don’tknow what they’re doing. They could be writing areport, checking email, or just reading the news. I don’tknow. But I normally get a sense of what they dobecause they tell me about it when we talk, or I getcopied on an email they send to someone else and Iread what they say or I overhear them talking aboutsome aspect of their project with someone who comesto their desk. So you just sort of get a sense of whatpeople do because you’re getting to see what they’resaying or hearing them talk about it. If you don’t havethat, it’s hard to know what they do.

As indicated, much of her awareness about hercoworkers’ tasks was generated through exposure totheir communications. However, as she implied, she,like most other employees at American Financial,only had visibility into the communications of thoseon her team or those sitting around her, and visibil-ity into even those communications was often onlyserendipitous.

Comparing the Marketing communication environ-ment before and after implementation of the enter-prise social networking site, as well as to the dynam-ics of the Operations communication environmentover time, showed that workers used the new tech-nology to enhance their awareness of who amongtheir coworkers had what kind of knowledge, aswell as to learn which coworkers knew each other.Increased awareness of who knows what and whoknows whom improved the accuracy of their meta-knowledge. Two advantages fostered by the new tech-nology helped improve the accuracy of metaknowl-edge: the transparency of coworkers’ messages andthe translucence of others’ networks.

INFORMS

holdsco

pyrightto

this

artic

lean

ddistrib

uted

this

copy

asaco

urtesy

totheau

thor(s).

Add

ition

alinform

ation,

includ

ingrig

htsan

dpe

rmission

policies,

isav

ailableat

https://p

ubso

nline.inform

s.org/.


Message Transparency and Awareness of “WhoKnows What.” As demonstrated above, a typicalproblem stemming from the invisibility of workand communication at American Financial was thatemployees did not know what their coworkers acrossthe organization knew. Employees recognized thatunder normal conditions communication betweencoworkers was largely invisible to everyone but thesender and the receiver. As one informant com-mented, “I send emails all day. What I say in themcould maybe be useful to someone else, but I don’tknow that so I don’t know who to copy on it. Some-one could probably learn a lot if they browsed aroundinside my inbox for a day just like I could if I couldsee inside other peoples’ emails.” Emails were notthe only culprit. Even messages exchanged in face-to-face encounters were often indiscernible by third-party observers. As another informant commented:“Sometimes I see people talking a couple desks away.I can’t hear what they’re saying but they look reallyinto the conversation and it looks important. It proba-bly is, but how can I know unless I go over there andeavesdrop.”

Use of the enterprise social networking site for rou-tine Marketing communication made coworkers mes-sages more transparent than they might have beenusing different technologies. Message transparencymeans that people can literally see what others aresaying to one another. On A-Life, a message sentbetween coworkers appears on both the sender andthe receiver’s "wall." If another person who was nei-ther the sender nor the receiver had accepted one ofthose two communicators into their social networkon the site, that message would appear on his orher news feed, essentially making that individual athird-party observer to the communication among theothers.

As an increasing number of employees began com-municating routine messages on the enterprise socialnetworking site, those who were neither the originalsender nor the recipient began to learn who knowswhat by looking at the messages. One informant suc-cinctly discussed this type of awareness:

I saw some messages exchanged between two guysin another department saying about how they deter-mined the appropriate rate for a consultant. I didn’tknow how to do that and I thought it would be goodto know so I kind of made a mental note that theseguys knew that so I could go ask them in the future.That would really be helpful knowledge to have forsome upcoming projects 0 0 0 I’m glad they sent that mes-sage through A-Life ‘cause that meant I got to see it. Ifthey did it through email I wouldn’t have ever knownthat message was sent and I couldn’t have seen it soI wouldn’t have learned that these guys know aboutrate determination.

Most of the messages communicated across theorganization were opaque to would-be third-partyobservers for two reasons. First, individuals simplydid not know that other people were communicat-ing with each other when those messages were sentvia email, IM, or some other electronic form of writ-ten communication. If they did not know that a mes-sage was being communicated, would-be third-partyobservers could not try to position themselves in aplace where they could view it. Second, even if athird-party could see that communication was occur-ring amongst others (for example, seeing two peopletalking across the hall or seeing someone pick up thephone to call someone else) the contents of those mes-sages were often inaccessible. Messages exchangedthrough the enterprise social networking site were notsubject to these two limitations, as perceived by third-party observers. Not only were the messages senttransparent such that a third-party observer could seeinto them, but they remained transparent over time:

A great thing about seeing people send stuff backand forth on A-Life is that it’s like a conversationalthread. So you can get to the conversation and scrolldown to see what they said before easily because theyrespond specifically to that thread and all the commu-nication’s there in one place 0 0 0Like if someone postedtheir emails to a website. Well, the email chain mightcontain some of the information, but if one of the peo-ple didn’t reply to original message and started a newmessage instead then you’d miss half of the originalmessage. That happens all the time. On A-Life youdon’t have that problem.

Consequently, messages that were communicatedthrough A-Life could be viewed by third partiesin ways that provided an aggregate understandingof the conversation, which helped them to becomeaware of who had what knowledge, even if the con-versation was no longer active.

Message transparency gave individuals access tothe contents of communication occurring among oth-ers. Of course, reading content did not translateinto an immediate awareness of “who knows what.”Instead, third-party observers had to rely on mes-sage content to make inferences about what kinds ofknowledge others in the organization had. Sometimesthe inferences made by third-party observers wereincorrect. But informants overwhelmingly trustedtheir own inferences (based on actual work-relatedinteractions among people) over someone’s proclama-tion about his or her own area of expertise:

A lot of people like to look or act like they know some-thing they don’t. They tell you, “I know how to dothat.” But you ask them later more about it and it turnsout they don’t. I find it’s better if you actually readsomething they wrote—like a note or a report—andyou can compare that to what someone else wrote and

INFORMS

holdsco

pyrightto

this

artic

lean

ddistrib

uted

this

copy

asaco

urtesy

totheau

thor(s).

Add

ition

alinform

ation,

includ

ingrig

htsan

dpe

rmission

policies,

isav

ailableat

https://p

ubso

nline.inform

s.org/.


then really assess what they know. It’s like going onevidence rather than hearsay.

It’s like a mix of things you do to figure out who’sgot the knowledge you need. You might get someadvice, but you want to do some homework anddo some comparison to see who is really an expertbecause there’s usually more than one person youcould ask. So if you see what people say to each otherthat’s a pretty good indicator you can use to decide.

What did informants do with the awareness of“who knows what” that was generated through expo-sure to messages communicated by other people onthe enterprise social networking site? The data sug-gest that they decided to whom they should go foradvice or whom they should approach to ask fortransfer of the needed knowledge. The top half ofTable 2 provides five representative examples, pulledfrom the interview transcripts, indicating how indi-viduals used their newly acquired awareness of “whoknows what,” which was gleaned from their use ofthe enterprise social networking site. In each case,informants indicated that they noticed the commu-nications occurring between others and attended tothem because, “what they were saying seemed relatedto something I was doing or something I knew Ihad to do.” Thus, potential future applicability of theknowledge that person had directed the observer’sattention to further read his or her transparent mes-sages. As the final column in the table indicates,observers did not normally need the knowledge thatthey became aware that someone else had immedi-ately. Instead, they projected a need for it sometimein the future. Out of all 33 instances identified (ofwhich only five are summarized in Table 2) the short-est time frame in which an observer projected needingthe knowledge they became aware that someone hadwas two weeks.

Network Translucence and Awareness of “WhoKnows Whom.” Whereas reading the content of themessages other people communicated on the enter-prise social networking technology enabled employ-ees to become aware of who knows what, the factthat they could see with whom others communi-cated and who those others signaled were their work-related colleagues led to an awareness of who knowswhom. The social networks of others, as perceived bythird-party observers, revealed actual and potentialcommunication partners. However, what third-partyobservers often could not discern was the nature ofthe communication or the scope of the relationship.For example, a third-party observer might see thattwo coworkers exchanged three messages. Yet theycould not know whether those were the only mes-sages they exchanged through any medium (indi-cating a weak tie) or whether they also exchangedmany more messages in person or through some other

technology such as email (indicating a strong tie).Similarly, if someone indicated that another Ameri-can Financial employee was his or her coworker (inthe way one would accept a “friend” in a popularsocial networking site like Facebook), the third-partyobserver could not know what topics these peoplecommunicated about (e.g., exchanging work-relatedadvice or coordinating a meeting time for lunch).Thus, I suggest that the social networks of othersmade visible through the enterprise social networkingsite are only translucent, i.e., clear enough to showthat a relationship exists, but not clear enough toshow the scope and/or nature of that relationship.Network translucence enabled by use of the enter-prise social networking site was often revealing forMarketing employees. Consider the following com-mon responses given by informants to the question,“What was the most surprising thing you learnedfrom using A-Life?”

I guess I didn’t realize how many people there are inMarketing that I don’t know, who know people I actu-ally do know.

I thought for sure no one I work with really talkedto people on the debit side of the business. But it turnsout I actually know some people who talk to thosefolks because I see they’re friends on A-Life.

I had no clue that Molly and Javier knew each other.But I see them send messages to each other from timeto time.

As informants typically indicated, it was difficultto know whom other people knew in such a largecompany. The default assumption enacted by manyindividuals was that the people they knew typicallyknew people they knew. In other words, a logic ofhomophily drove most people’s conceptualizations ofthe underlying social dynamics at American Finan-cial. But by the translucent networks they perceivedon the enterprise social networking technology beliedthis belief. Instead, informants began to learn thattheir communication partner’s communication part-ners were actually quite diverse.

There were several ways that individuals becameaware of whom their coworkers knew. The most com-mon way was by seeing the list of names of “con-nections” that was linked to someone’s profile page.Just like the “friends” list on Facebook, employees atAmerican Financial could formally request someoneat the company to be a connection and accept connec-tion requests from others. This list of cultivated con-nections was displayed in association with a person’spersonal profile. This list of “connections” providedan easy way for individuals to become aware of whoknew whom:

Sometimes I just kind of look at someone’s list ofconnections really quick just to see what people theyknow. That’s always interesting because you get a

INFORMS

holdsco

pyrightto

this

artic

lean

ddistrib

uted

this

copy

asaco

urtesy

totheau

thor(s).

Add

ition

alinform

ation,

includ

ingrig

htsan

dpe

rmission

policies,

isav

ailableat

https://p

ubso

nline.inform

s.org/.


Table 2 Examples of Metaknowledge Learned as a Result of Increased Communication Visibility

How metaknowledge was Time framelearned on enterprise social How metaknowledge for projected

Type Knowledge needed networking site was/will be used use

Who knows what How to make spendingprojections for customeracquisitions

Reading message one employee sent to otherto ask for help

Ask the message’s recipient to teachemployee to make projections

24 weeks

How to determine date forpromotional mailings

Reviewed a template document sent alongwith a message from one employee toanother

Ask the template’s creator aboutfeasibility of altering date

12 weeks

How to perform search engineoptimization

Saw at least four messages among differentpeople where sender asked recipient for thename of an expert who could help on just thistask. One expert was consistently mentioned

Approach employee who isconsistently mentioned and ask to betaught task

2 weeks

How to track tasks in a projectmanagement environment

Noticed template sent by another employeein a message

Ask poster of template for tutorial inproject tracking

12 weeks

How to forecast revenue for newmarket segment

Saw the title of a document that coworkersent, which indicated that the documentcontained a report in which forecast wasmade

Ask poster of document for a lessonon how to conduct forecast

9 weeks

Who knows whom Who has a contact with someoneat the Datacenter who can extractpromotional histories

Reviewing a message response fromemployee commenting on a meeting he wasabout to attend

Ask sender for introduction toemployee at Datacenter

6 weeks

Who to ask about the best personfrom whom to learn how to writerewards language compliant withNAD rulings

Noticing that a close colleague had aconversation with someone in legal

Ask the colleague to ask his or herfriend in legal who is the appropriateperson to contact

3 weeks

Who knows someone to contactin the Finance department toexpedite review of impacts of newpromotional program

Noticed that someone had asked a colleaguea question on his wall about who is the bestperson to contact in the Finance department.Employee figured the person asked thequestion because the person of whom it wasasked knew the answer

Ask the answerer of the question whois the best person to contact in theFinance department and advice forthe best way to approach this person

5 weeks

Who to ask for a contact who hasrun a social media contest

Saw a message sent by a coworker whotalked about a focus group she had done atwork to give ideas for a social mediapromotion

Ask coworker for name of focusgroup leader and for introduction tohim or her

3 weeks

Needed help writing a script toautomate data entry. Heard ofsomeone who knew how to dothis in IT, but heard that personwas busy.

Saw a picture included in a message from acolleague in which the colleague was talkingat a company party with the person from IT.Asked the colleague for an introduction to theperson from IT

Ask colleague to persuade person inIT to help employee with her request.Personal favor needed because it isoutside of IT’s job description

4 weeks

sense of who they talk with and a lot of times they’repeople you wouldn’t expect.

Although perusing the list of one’s displayed con-nections was certainly the easiest way to get a senseof their work-related contacts, it was also the mostunsatisfactory because it provided the leanest infor-mation. After reviewing this list, the observer stillwould not know the content that typically flowedalong the tie between the two individuals nor wouldthey know how frequently those two individualscommunicated.

Consequently, a more robust way of learning “whoknows whom” was to pay attention to the messagespeople exchanged with one another. Informants tookas evidence of a real and substantial relationship the

observation of coworkers sending messages to eachother:

On A-Life you can see what people are sending mes-sages to other people. So you see that message and itslike, “OK, they know each other, great.” And you cankind of just remember that piece of information. Andsometimes you see that they send a lot of messages toeach other so you say, “Well, I guess they really knoweach other pretty well.” So that’s all helpful informa-tion to have.

The advantage of building an awareness of “whoknows whom” by attending to actual messagesexchanged was that the nature and intensity of the tiewere easier to discern than building such an aware-ness simply by reviewing one’s list of contacts. Toacquire this more granular information, informants

INFORMS

holdsco

pyrightto

this

artic

lean

ddistrib

uted

this

copy

asaco

urtesy

totheau

thor(s).

Add

ition

alinform

ation,

includ

ingrig

htsan

dpe

rmission

policies,

isav

ailableat

https://p

ubso

nline.inform

s.org/.


would regularly check their own news feeds on theenterprise social networking site to see if conversa-tions among their connections appeared there. Theywould also regularly peruse the walls of people theyknew with the explicit goal of learning who theircommunication partners were.

What did informants do with the awareness of“who knows whom” that was generated throughexposure to translucent networks on the enterprisesocial networking site? The data suggest that theydecided which of their coworkers could help them togain an introduction to others at the organization orendorse them for others from whom they wanted toask a favor:

Sometimes you want to talk to someone, but you don’tknow them. So you think, “who do I know that mightknow them?” and then you go ask that person if theycan make a connection for you, like call them or some-thing. Or, you might know that you want to ask some-one you don’t really know well for something, like togive you a bunch of data. So to make that easier youask someone you know to put in a good word for youto that person, like “Jim’s not a bad guy it would begreat you’d help him out” and that can sort of greasethe skids.

Another way in which informants indicated thatthey used their knowledge of “who knows whom”gained through use of the enterprise social network-ing site, was to try to discern “who knows who knowswhom.” For example, if a person knew that he or shewanted to talk to a specific person in accounting, butdidn’t know that person directly, they might see onA-Life that one of her coworkers knows someone inaccounting and would ask that person to ask the per-son in accounting if she knows the person in question.The bottom half of Table 2 provides five representa-tive examples, pulled from the interview transcripts,indicating how individuals used their newly acquiredawareness of “who knows whom.” As the examplesillustrate, individuals often had either specific (byname) or general (by position) people they wanted totalk with across the organization and used the enter-prise social networking site to help them figure outwho amongst those that they did know could helpthem achieve the goals summarized above. As wasthe case with knowledge of “who knows what,” infor-mants indicated that they did not normally activatetheir new knowledge of “who knows whom” imme-diately after acquiring it. Out of all 41 instances iden-tified (of which only five are summarized in Table 2)the shortest time frame in which an observer pro-jected needing the knowledge they became aware thatsomeone had was three weeks.

New Ways of Working Arising from ImprovedMetaknowledgeThe increased awareness of who knows what andwho knows whom enabled by seeing coworkers’ com-munications in the enterprise social networking sitehad at least two important consequences for the Mar-keting staff. First, enhanced metaknowledge helpedworkers to avoid doing jobs that had already beendone and to avoid taking the time to learn somethingthat a coworker had already learned and could share.In short, improved metaknowledge helped workersavoid duplicating previous work. Second, enhancedmetaknowledge helped workers to combine disparateideas from across the organization into potential newideas for specific projects. This recombinant innova-tion became feasible when workers improved theirawareness of both who knows what and who knowswhom in the organization.

Duplication Avoidance. As in many large organi-zations, work duplication was a common problemat American Financial. Table 3 provides examples ofwork duplication that occurred in Marketing prior toimplementation of the enterprise social networkingsite. As the table indicates, when staff were assignedto work on a particular task they often began to learnthe knowledge necessary to complete the task with-out learning whether someone else in the organiza-tion had either already completed the same task andcould simply share their results, or whether some-one else in the organization already had the knowl-edge necessary to do the task and could help the per-son in question do it without that person having toengage in a lengthy learning process. Typically, thefact that knowledge and work were duplicated wasdiscovered serendipitously after the duplication hadoccurred and American Financial had incurred theassociated costs.

The improved metaknowledge enabled by use ofthe enterprise social networking site helped Market-ing staff avoid these kinds of duplication. Havingan awareness of who knows what gave informantsa window into the kinds of tasks or jobs coworkerswere likely to conduct. This insight made it easy forthem to ask others about particular tasks they wereworking on to ensure that knowledge or work wouldnot be duplicated. As one informant summarized:

My manager asked me to put together a report sum-marizing some of our competitor’s positioning andbranding strategies for a particular product. I was like,“Ok, I’ll do that.” I had a lot of other things going onand this was going to take two or three days to do anddelay me getting to those other things. I told her [themanager] that and she said it was ok. But as I started todo it I remember that I saw that Jenny in the ConsumerInsights department had sent a message on A-Life tosomeone about some work she was doing with this

INFORMS

holdsco

pyrightto

this

artic

lean

ddistrib

uted

this

copy

asaco

urtesy

totheau

thor(s).

Add

ition

alinform

ation,

includ

ingrig

htsan

dpe

rmission

policies,

isav

ailableat

https://p

ubso

nline.inform

s.org/.


Table 3 Examples of Knowledge Duplication Problems in Marketing Division

Knowledge duplication Reason problem How existing knowledge Consequence arisingTask problem occurred was discovered from duplication

Assigned to createreport on consumerdefault behavior

After creating report,employee discovered thatsomeone had alreadycreated an identical reportsix months earlier

Manager assigned task, butforgot he had assigned thesame task to a person sixmonths before. Employeedid not know that someoneelse had created it

During small talk at companyholiday party employeesdiscussed past projects

Employee delayed other worktasks for four days whilecreating report. Misseddeadline assigned by anothermanager

Assigned to createbanner ad forholiday credit cardoffer

Researched types of graphicsthat capture attention.Hired outside consultant tocreate art. Discovered laterthat graphic could not beused for legal reasons

Employee was unaware of aposition in the company,occupied by anotheremployee, dedicated tocreating banner ads

During review of banner ad bylegal department, lawyerasked why employee hadnot utilized the dedicatedbanner ad position

Employee lost two weeks ofproductivity and spent $15,000on a consultant who created abanner ad that could not beused

Assigned to place asmart button inemail so that whenmobile phone userclicked the link, itwould redirect themto their App store

Looked up logic to insert linkfor appropriate device(iPhone vs. Android).Applied various solutionsfound through Googlesearch. Discovered laterperson in IT who knewsolution.

Employee thought it was sucha rare instance that hedidn’t think to go aroundand ask for someone whoknew (employee didn’tthink anyone had thatexpertise)

Finally connected with ateammate of employee’s“go to” person in IT (afterthree weeks of emails tovarious people), who had abetter solution to theproblem

Wasted time and made himselflook incompetent (when hissolutions found on Googlefailed)

Assigned to create adashboard tosummarize theprogress/successmetrics of variousteam projects

Similar dashboards existed,but needed to makechanges that reflected newmetrics being tracked. Didnot have skills to createdashboard, but tried to byworking two weeksnonstop. Later, found acolleague who knew how tomake the update.

Employee did not think thatanyone else in her grouphad the skills to make theupdate and she didn’t knowanyone in another groupwho might have them

While training a colleaguewho was moving over fromanother department thetrainee mentioned that hepossessed the skills toimprove the dashboardand auto-populate thenecessary data

Employee lost two weeks of worktime trying to update thedashboard. In addition, threeother projects were stalledbecause workers lacked datathat they could have seen if thedashboard had worked.

Needed to changecopy (wording) on aletter going to cardmembers aboutrewards

Information about rewardschanged because of NADrulings. Did not learn thatwording had changed untilafter the letters weremailed.

Employee’s department didnot normally work withlegal so, consequently, theywere not on legal’s regularupdate emails

While waiting for a staffmeeting to begin, employeetalked with coworker aboutthe rewards program thecoworker initiated and thecoworked mentionedchanges from legal

Company had to retract the firstletter and send out a secondletter with revised awards.Also, purchase credits had tobe given to customers whomade purchases under falseassumptions about rewards.Overall, more than $60,000was spent to fix the problem.

particular competitor. So I called her up and told herwhat my manager wanted me to do and before I couldfinish she told me that she had already done that as aby-product of something she was working on and shewould send me the file. It was exactly what I neededso I didn’t have to do the work. It was great 0 0 0 If Ihadn’t seen her talk about that on A-Life I wouldn’thave ever known to ask her and it would have wastedthose couple of days and pushed back the other work.

Awareness of “who knows whom” was also a cru-cial ingredient for duplication avoidance in manycases because even if, through awareness of “whoknows what,” someone learned that someone elsehad completed a particular task or had certain nec-essary knowledge, it was not always straightforwardthat the person who had done the task or hadthe knowledge would share. Sometimes, knowing

someone who knew that person and could provideentrée to him or her was needed:

So it turned out that this guy in Business Developmenthad put together this fair market value tool for payingconsultants, which would say what their rate shouldbe. I needed to hire a consultant for my project and Ineeded to figure out the rate when I found out this guyhad this tool. So I emailed him to ask if he could tell mewhat information he used to make the tool, I wasn’teven asking for the tool itself, but he never emailed meback. Maybe he didn’t want to give it up or something.But I happened to see one day when I was on A-Lifethat he was friends with this other guy I know reallywell. Well, I mean they were listed as “connections”and my friend didn’t have a lot of connections. So Iasked my friend if he knew the guy and would ask himabout the tool for me. He said, “sure” and he copied

INFORMS

holdsco

pyrightto

this

artic

lean

ddistrib

uted

this

copy

asaco

urtesy

totheau

thor(s).

Add

ition

alinform

ation,

includ

ingrig

htsan

dpe

rmission

policies,

isav

ailableat

https://p

ubso

nline.inform

s.org/.


the guy and me on an email together. The next daythe guy sends me his tool. That was great, saved me abunch of time.

Avoiding the duplication of knowledge and workwas a tremendous benefit for employees, their man-agers, and the organization as a whole. As one man-ager noted, “We hire smart people and we want themto share what they know. We’re a better companywhen that happens.” Use of the enterprise social net-working site helped individuals in Marketing to avoidduplication by helping them to become aware of whomight have done something similar already or hadsimilar knowledge and who could help them get tothose people and their products. As one informantnoted, “It’s amazing how much more efficient you canactually be when you have a good lay of the land inthe company. I never realized [before using A-Life]how much I was missing.”

Recombinant Innovation. Innovation in commer-cial products and internal processes was highly val-ued at American Financial. Employees were formallyrecognized by management for developing new waysof working more efficiently or for generating goodideas that translated into new product offerings. Butdeveloping innovative processes and products wasdifficult. An increase in the accuracy of one’s meta-knowledge, which was enabled by use of the enter-prise social networking site, allowed informants tomore effectively engage in what innovation schol-ars refer to as “recombinant innovation” (Hargadon2002). Recombinant innovation refers to innovationsthat arise when knowledge that already exists in theorganizations is re-used in new combinations thatresult in novel and useful changes in products, pro-cesses, or services (Majchrzak et al. 2004). Innovatingthrough recombination of knowledge requires indi-viduals to be aware of what knowledge exists in theorganization that can be combined with other knowl-edge. By enabling better awareness of who knowswhat and whom, use of the enterprise social net-working site allowed some employees in Marketingto meet this requirement.

For example, Marta, who worked in the Card-member Marketing department, had spent severalmonths conducting research on why consumers werelikely to choose one credit card brand over another.Her research uncovered that, within the demographicfor which she was interested, consumers made deci-sions largely based on the availability of rewardsprograms and, more specifically, rewards that couldbe redeemed for cash or credit on a card statement.Marta tried, unsuccessfully, for several weeks to fig-ure out a strategic plan based on her findings. Oneday, however, Marta had a breakthrough:

I was racking my brain about it when I suddenlyremembered that I’d seen a communication exchangedbetween these two guys on A-Life in which one ofthem, Joe Franklin who’s in the Pricing and Ana-lytics department, mentioned something—I forget thecontext—about interest rate variation depending onspending habits. That caught my eye so I read thehistory of conversation he had and it turns out thatthere was some flexibility in rate assignment based onspending class. When I remembered I had seen this itoccurred to me that we might offer a cash-back bonusor something similar for specific consumers in thatspending class if we kept below our rate assignmentthreshold. So I sent him an email to see if I could learnmore about it. It made sense so I developed the pro-gram around it and it has been pretty successful so far.I am proud of the innovativeness as well.

As this example illustrates, Marta combined herknowledge about consumer preferences with JoeFranklin’s knowledge about interest rate variationto produce a novel and useful new marketing pro-gram. Marta produced this innovation by recombin-ing knowledge in new ways. Importantly, her abilityto do so was enabled by her awareness of what Joeknew, which she had because of the visibility into hiscommunication provided by the enterprise social net-working site.

Informants from Marketing discussed numerousother cases similar to Marta’s, most of which weremuch smaller in scope and significance, in whichthey combined knowledge that already existed withinAmerican Financial in new ways because use of theenterprise social networking site allowed them to seewhat knowledge other people had, as well as to seewho they would need to talk with to access thatknowledge. As one informant suggested:

One thing A-Life gives you is a better vantage point.You can simply see more of what other people aredoing by getting involved in their communications. Soyou start to realize what they know. Also, even thoughyou might not understand it completely, you can thinkto yourself, “Hey, I didn’t even realize someone hadthis knowledge” and then you can start to think howthat knowledge might complement some knowledgeeither you already have or some other knowledge thatyou saw someone else had. That is pretty cool andhelps you to be more innovative in your job by lever-aging off of what others know.

The metaknowledge enabled by use of the enter-prise social networking site did not appear to makeanyone inherently more creative. But, it did appear togive people a vision advantage such that they couldsee what knowledge existed within the organizationto be had and to acquire it from those who held itsuch that they could combine it with other pieces ofknowledge to innovate in their work.

INFORMS

holdsco

pyrightto

this

artic

lean

ddistrib

uted

this

copy

asaco

urtesy

totheau

thor(s).

Add

ition

alinform

ation,

includ

ingrig

htsan

dpe

rmission

policies,

isav

ailableat

https://p

ubso

nline.inform

s.org/.


Behavioral Changes Enabling New Ways ofWorkingAlthough the data suggest that duplication avoidanceand recombinant innovations were two new ways ofworking that arose because enterprise social network-ing site use enabled staff to develop more accuratemetaknowledge, not everyone could engage in thesenew behaviors. In fact, the analysis showed that ofthe 16 Marketing enterprise social networking siteusers interviewed at the end of this study, only eight(i.e., exactly half) reported that use of A-Life allowedthem to avoid knowledge duplication or to recombineknowledge in ways that produced product or processinnovation. Interestingly, seven of these eight usersdid both. To learn how these eight users worked inthese new ways, I compared their responses to theresponses of the other eight informants who did notachieve these two benefits, as well as to the Opera-tions informants who also did not report working inthese new ways. This subsequent analysis revealedthat informants worked in new ways. Avoiding dupli-cation and recombining knowledge for innovationwere two important behavioral changes that allowedthem to take advantage of the enhanced awarenessof who knows what and who knows whom affordedby the new technology: They shifted from engagingexperiential learning to vicarious learning and movedfrom conducting reactive search practices to proac-tively aggregating knowledge.

From Experiential Learning to Vicarious Learn-ing. The first behavioral change to adopt the newways described above was to shift the way the workerlearned from others. Instead of learning through per-sonal experience, he or she learned vicariously byobserving the communication between a coworkerand someone else. Marketing and Operations infor-mants suggested that experiential learning was typ-ically the way they developed an understanding ofwhat and whom coworkers knew, and that this hap-pened primarily through communication:

You get to know about other people by talking withthem directly and asking questions and hearing theiranswers. That’s how you get to know what projectsthey’ve worked on and get a sense about the kinds ofthings they have expertise about.

Communicating with other people is one of the mostimportant things you can do to be successful in yourjob. You’ve got to make sure you’re learning all thetime about what other people can do and you canonly learn that by being good at engaging them inconversation.

When you’re talking to someone else you can reallyget a good grip on the kinds of things they knowabout. It’s just the way they respond to you—likehow fast or what they’re body language is like—ishow you can get a sense about it. You know, when

you’re immersed in conversation and you just kind ofabsorb it.

Learning through direct communication with oth-ers was also common because interpersonal exchangeforced individual attention. In other words, in addi-tion to the benefit of asking clarifying questions tosustain a conversation (whether synchronous or asyn-chronous) workers had to focus on the communica-tion partner’s words to respond. This forced attentionhelped them store the learning about what and whomthat person knew in memory. Repeating it in conver-sation helped to solidify it as actual knowledge.

Yet workers effectively avoided duplication ofknowledge, and combined knowledge in ways thatproduced innovative outputs. Not only did they learnthrough their own experience communicating withothers, but also by observing other communications inwhich they were not themselves involved. Althoughthis behavioral change seems somewhat subtle, infor-mants stated that it actually presented substantialcognitive challenges:

The major thing I’ve been finding as I’ve been tryingto figure out the best way to use A-Life is I get a lotmore out of it by being a lurker than I do from talkingto people, personally. I mean the people that I wouldsend a message to on A-Life are probably the samepeople I would talk to offline or on email. But whenI’m on A-Life I can see what a bunch of other peo-ple that I wouldn’t normally talk to are saying to oneanother and all that conversation has useful tidbits init. But I’ve got to make it a process to read those com-munications or at least look at them quickly and I haveto reflect on them a bit more than I would if I talkedto either of those people directly because I’ve got totry to pick up the context of the comments. It’s like,in school, if you ask the teacher something directlyyou know how the answer fits into the bigger picturebecause you asked about it. But if you overhear theteacher talking to another student you can still learn alot, but you’ve got to figure out how to contextualizeit. That’s hard, but it can be real valuable.

Learning vicariously through observation of com-munications between others represented an importantbehavioral change for those who used the enterprisesocial networking site. Conceptualizing user commen-tary to infer knowledge was particularly important;the sustained attention necessary to cull informa-tion from coworkers’ communications required effortand control. Significantly, informants did not aban-don experiential learning entirely. They continued tolearn through direct conversation with others in theirwork groups and from those who sat near them. Thus,vicarious learning took place most often through useof the enterprise social networking site. Informantshad to balance these two modes of learning and movequickly and adeptly between them. This entailedanother unique set of cognitive competencies.

INFORMS

holdsco

pyrightto

this

artic

lean

ddistrib

uted

this

copy

asaco

urtesy

totheau

thor(s).

Add

ition

alinform

ation,

includ

ingrig

htsan

dpe

rmission

policies,

isav

ailableat

https://p

ubso

nline.inform

s.org/.


From Reactive Search to Proactive Aggregation.The second behavioral change that workers madeto avoid work duplication and achieve recombi-nant innovation was in their approach to the tim-ing and process of finding knowledge. As Marketingand Operations informants indicated, they typicallyacquired new knowledge in the process of trying tosolve a newly encountered problem.

I guess what normally happens is I run into some issuein my work that I can’t figure out. So I go out askingpeople if they have some solution. Hopefully I findsomeone. I’ve gotten better about searching for what Ineed over time.

If I have some problem that I can’t solve I look forsome kind of knowledge or information that will helpme solve it. That is probably the instance where I learnthe most. You know, you learn when you don’t under-stand something; it’s just the time when that normallyhappens.

The times when I really find out what people knoware when I’m stuck in my own work. I get into somekind of jam and there is some problem so I search for asolution. I ask around and around to people till I comeup with something that will help. Then, it’s like, “Ohyeah, now I just learned something new."

Put in more formal terms, the typical way in whichpeople dealt with problems encountered in their workwas to reactively search for knowledge that wouldhelp to solve them. The process could be character-ized as reactive because people did not think aboutacquiring new knowledge, or the metaknowledge thatwould help them get that knowledge, until the prob-lem arose. Moreover, the process could be character-ized as search because they were actively looking ina focused and directed way for some specific type ofknowledge.

Note that those who learned the new ways ofworking described above did not reactively searchfor knowledge when they encountered a problem.Instead, they proactively aggregated the metaknowl-edge they acquired daily through the communicationvisibility made possible by the enterprise social net-working technology. In other words, they stumbledinto knowledge of who knows what or who knowswhom at random intervals. Although they had no usefor that metaknowledge at that moment, they held itin passive memory (along with other pieces of meta-knowledge) for future use. Informants noted that thisbehavioral change was profound.

The biggest change I have seen from using A-Life isthat when I encounter a problem I automatically knowwho can help me solve it. It is unusual. Earlier, when Iencountered a problem, I would think, "Damn! Wheream I going to get help for this?" Now, I know imme-diately who to contact even though I may have nevermet that person.

Although this shift in behavior from waiting untilknowledge was needed to learn who might haveit or who might know the person who had it, todeveloping metaknowledge in advance, in anticipa-tion of some future need state, was somewhat drastic,it could pay large dividends. To accumulate those div-idends required that informants be more proactive inthe purposeful development of metaknowledge:

I find myself doing something new now when I’m onA-Life. I see all these things that people are saying andI get a sense for what they know. It’s a lot of noise, butit’s useful noise. So when I see something and realizethat John knows about consumer promotion rates I sortof take a breath and tell myself that I should rememberthat and I file it away. That’s a pretty big change—tosee something and try to absorb it so you can use itlater when you don’t even know if you’ll ever need touse it.

Although use of the enterprise social network-ing site made it easier to learn who knows whatand who knows whom, remembering that knowledgeappeared to take some willful determination.

DiscussionThis paper offers a grounded theory of communi-cation visibility. The emerging theory suggests thatonce invisible communication between others in theorganization becomes visible for third parties, thosethird parties may improve their metaknowledge (i.e.,knowledge of who knows what and who knowswhom in the organization). Communication visibil-ity leads to enhanced awareness of who knowswhat and whom through two interrelated mecha-nisms: message transparency and network translu-cence. Seeing coworkers’ messages helps third-partyobservers make inferences about others’ knowledge.Seeing the structure of coworkers’ communicationnetworks helps third-party observers make inferencesabout who coworkers talk with somewhat regularly.The emerging theory further suggests that enhancedmetaknowledge can lead to more innovative productsand services and less knowledge duplication if stafflearn to work in new ways. Specifically, users canrecombine existing ideas into new ideas more effec-tively and avoid duplicating work if they switch fromlearning through experience to learning vicariouslyand if they can begin to proactively aggregate infor-mation they are perceiving daily rather than engagein reactive search after they have run up against aproblem they cannot solve.

Enterprise social media plays a central role in thistheory because it makes heretofore invisible commu-nication between coworkers visible to others in theorganization. Specifically, the enterprise social net-working site used among employees in this paper

INFORMS

holdsco

pyrightto

this

artic

lean

ddistrib

uted

this

copy

asaco

urtesy

totheau

thor(s).

Add

ition

alinform

ation,

includ

ingrig

htsan

dpe

rmission

policies,

isav

ailableat

https://p

ubso

nline.inform

s.org/.


enabled them to see into their coworkers’ mes-sages (message transparency) and view coworkers’communication networks (network translucence). Byenabling message transparency and network translu-cence, a social media tool such as an enterprise socialnetworking site fosters conditions for the dynamicssummarized above. However, the mechanisms thatdrive these dynamics are themselves message trans-parency and network translucence. Enterprise socialnetworking sites are not the only information tech-nologies that can make messages transparent andnetworks translucent. The notion of visibility is tiedto the amount of effort necessary to locate infor-mation in workers’ communications. If workers per-ceive that information is difficult to access, or if theydo not know what information is accessible, theyare unlikely to seek that information (Brown andDuguid 2000). In short, information may be avail-able for examination, but unless workers know whereit is, it will remain invisible. Clearly, effort can helpmake communications visible. Still, busy workers areunlikely to make that effort if access to those com-munications is too difficult. Thus, any technology thathelps to make coworkers’ communications more eas-ily accessible will make that information visible acrossthe organization. As these findings show, an enter-prise social networking site accomplished this in waysthat other technologies available to informants couldnot, thus providing them access to new and poten-tially valuable knowledge.

This emerging theory of communication visibilityhas several important implications for the nature ofwork in the knowledge economy. As clearly shownby the data presented, communication visibility andexpanded metaknowledge can streamline work acrossan organization, while also fostering improvementsin innovativeness. By learning who knows what andwho knows whom employees at American Finan-cial saved themselves and the company time andmoney by avoiding duplication of knowledge andwork. Those who were best equipped to solve prob-lems could do so. This, of course, is the goal ofwork specialization and encouraging intraorganiza-tional collaboration (Ren and Argote 2011). In addi-tion, by recombining existing knowledge with newideas, employees bring new levels of innovation toorganization products and services. Burt (2004, p. 351)suggests that those who are adept at innovatingthrough recombination of knowledge have a “visionadvantage” that others in the organization do not.He also suggests that this vision advantage comesfrom bridging across structural holes in an organi-zation’s communication network. Yet the findings ofthis study suggest that technologies such as enterprisesocial networking sites make communication amongcoworkers universally visible, thus giving the vision

advantage to all employees, regardless of networkposition.

The findings also showed that although commu-nication between coworkers was visible to all usersof the enterprise social networking site, not all userschanged their behavior to avoid duplication and toinnovate through recombination. Communication vis-ibility thus appears to be a necessary but insufficientcondition for work to change in these ways. Thoseusers who successfully changed their work devel-oped specific new behaviors of vicarious learningand proactive knowledge aggregation. Most organiza-tion research on learning processes heralds experien-tial learning or (i.e., learning by doing) (Von Hippeland Tyre 1995) rather than vicarious learning (i.e.,learning by watching) to acquire knowledge. Thefindings presented here suggest that in the modernworkplace vicarious learning may have as many ifnot more advantages than experiential learning inthe development of metaknowledge. Still commu-nications must be visible if vicarious learning (i.e.,observing coworkers’ communications) is to succeed.Also, organizational scholars typically herald prob-lemistic search (Cyert and March 1963) as the vehi-cle for workers to effectively acquire new knowl-edge, i.e., when they encounter a problem, they searchfor a solution. Yet our findings suggest that proac-tively aggregating knowledge by observing cowork-ers’ visible communications may be more effectivethan the reactive problemistic search. To successfullymake this switch, workers would hone and retaintheir metaknowledge daily so that it could be recalledwhen needed. Developing metaknowledge in thisway would require individuals to attend to bits ofinformation that may seem trivial or useless in themoment and later recall those bits and assemble themwith other bits to arrive at an accurate understandingof where knowledge is located and who can provideaccess to it. Thus, these two behavioral changes rep-resent new ways of working that employees wouldlikely have to adopt to take advantage of the commu-nication visibility provided by new technologies suchas enterprise social media.

Although the outcomes of communication visibil-ity in this particular study were quite positive, intro-duction of social networking technologies to the orga-nization could be problematic. For example, whenonce invisible behavioral information becomes visi-ble, workers may become fearful of surveillance bymanagement and coworkers (Ciborra 1997, Zuboff1988). Moreover, visible communications may fosterself-preservation behaviors by which workers do notcommunicate the true nature of their work, but ratherwhat they believe others think they do and know.(Leonardi and Treem 2012). Thus, the openness fos-tered by communication visibility could backfire to

INFORMS

holdsco

pyrightto

this

artic

lean

ddistrib

uted

this

copy

asaco

urtesy

totheau

thor(s).

Add

ition

alinform

ation,

includ

ingrig

htsan

dpe

rmission

policies,

isav

ailableat

https://p

ubso

nline.inform

s.org/.


create unproductive organizational behaviors includ-ing intentional ambiguity (Gibbs et al. 2013). Clearly,these responses run counter to the tenets of effec-tive knowledge sharing. Over time, if communica-tions through the enterprise social networking tech-nology stagnate, the result could be smaller increasesin the accuracy of metaknowledge. In addition, moti-vation to use the enterprise social networking sitemay be undermined by the visibility of too manycommunications between too many workers. Vicari-ous learning may then be stifled when workers’ atten-tion is divided across too many messages. Workersmay begin to focus instead on those messages thatdirectly relate to their immediate task, and in so doingartificially limit the advantages of the organization-wide reach for which the enterprise social networkingsite was designed.

In addition to the broad implications for organiza-tion theory discussed above, the findings presentedin this paper contribute to the more specific researchon transactive memory. Transactive memory theo-rists argue that accurate metaknowledge is necessaryfor a successful transactive memory system. Yet (tomy knowledge) there has been no in-depth explo-ration of how such accuracy is achieved or improved(Peltokorpi 2008, Ren and Argote 2011). Our find-ings demonstrate how enterprise social networkingsites can foster the development and improvement ofaccurate organizational metaknowledge by enablingawareness of ambient communication. Significantly,the findings show that user inferences about whoknows what and who knows whom, gleaned by atten-tion to coworker’s visible communications, are reli-able and produce accurate cognitive knowledge andsocial structures.

Transactive memory researchers have shown thatorganizational technologies such as databases anddirectories are useful for helping teams allocate andretrieve knowledge from others (Choi et al. 2010).These findings demonstrate that technologies, such asenterprise social networking sites, which make com-munication visible and thus enable ambient aware-ness, can also help workers maintain accurate directo-ries of who knows what and who knows whom. Thefindings also show that use of such technologies canincrease the similarity between coworkers’ cognitiveknowledge and organizational social structures. Theproduction of shared cognition is possible throughthe use of enterprise social networking sites becauseworkers draw from the same set of cues (transparentcommunications and translucent networks) to formperceptions of “who knows what” and “who knowswhom.” Consequently, using enterprise social net-working technologies can lead to metaknowledge thatis not only more accurate, but also more similar acrosscoworkers. Transactive memory theorists argue that

this improved metaknowledge provides the basis forjointly coordinated action and more economic distri-bution of effort (Austin 2003).

To help employees share their knowledge when car-rying out tasks, many organizations implement tech-nologies such as intranets, enterprise resource plan-ning systems, shared data repositories, group supportsystems, groupware and customer relationship man-agement systems, etc. Indeed, years of research ondecision making and information sharing in experi-mental and organizational teams shows that workersoften fail to integrate their knowledge with cowork-ers because they do not know what others know(Argote et al. 2000). Despite the fact that an increas-ing number of firms are implementing knowledgemanagement technologies, few organizations reportdramatic improvements in the exploitation of cumu-lative expertise (Kankanhalli et al. 2005). Scholarshave suggested that many organizations using knowl-edge management systems do not reap their benefitsbecause workers simply do not enter into the systeminformation that accurately reflects their knowledge(Wasko and Faraj 2005). There are multiple hypothe-ses as to why this might be so: Perhaps workers arenot proficient in using the new technology (Yuan et al.2005); perhaps electronic documentation has not yetproliferated as a team norm (Walsham 2002); perhapsworkers do not think the information they have isimportant for their coworkers (Cress et al. 2006).

The largest impediment to the success of knowl-edge management technologies for organization-wideknowledge sharing is that documenting the knowl-edge required to complete one’s work is not a nat-ural or routine part of most tasks. For example, aworker who writes marketing plans is typically eval-uated on the success of those plans. Documentinghow the material for the plan was assembled is not astep in the plan itself. Thus, it may be difficult for aworker to show how they used information technol-ogy to enhance their knowledge and improve knowl-edge management efforts. This is not a normal partof the worker’s task (Bowker et al. 1995). It may evendiminish the worker’s ability to successfully completefuture assignments.

Goffman (1959) referred to this problem of actionversus representation of action in what he called theparadox of representation. He suggested that thosewho are most skilled at particular tasks spend so muchtime on those tasks that they do not have time to letothers know how they completed them. He furtherargued that because most work is done in private (i.e.,away from coworkers), workers must “dramatize”their task to make it visible: “While in the presenceof others, the individual typically infuses his activ-ity with signs which dramatically highlight and por-tray confirmatory factors that might otherwise remain

INFORMS

holdsco

pyrightto

this

artic

lean

ddistrib

uted

this

copy

asaco

urtesy

totheau

thor(s).

Add

ition

alinform

ation,

includ

ingrig

htsan

dpe

rmission

policies,

isav

ailableat

https://p

ubso

nline.inform

s.org/.


unapparent or obscure” (1959, p. 33). This insight sug-gests those who take the time to work well may nothave the time to explain to others how they work well.In other words, Goffman suggested: “The problem ofdramatizing one’s work involves 0 0 0making invisible[actions] visible. The work that must be done by thosewho fill certain statuses is often so poorly designed asan expression of desired meaning, that if the incum-bent would dramatize the character of his role, hemust divert an appreciable amount of his energy to doso” (p. 32). As a consequence of this paradox, manyworkers do not take steps to make their behaviorspublic. Thus, the knowledge accumulated in the pro-cess of generating their work is unavailable to othersin the organization (Gardner 1992).

A major implication of our findings is that ratherthan encouraging (or forcing) workers to documenttheir knowledge for others, a more fruitful strat-egy may be to have them switch their channels ofcommunication to technologies that help make theircommunication visible to others. Developing accu-rate metaknowledge through exposure to ambientcommunications requires little work on the part ofsources. Tangentially, use of an enterprise social net-working site requires minimal effort on the partof knowledge sources. Sources need only encourageworkers to shift their communication activity fromprivate channels (e.g., phone, email, instant messag-ing, etc.) to an enterprise social networking site wherecommunications are public and visible to others in theorganization. Because they are simply communicatingas they normally do, there is a larger body of informa-tion available to their coworkers. From this coworkerscan develop their own metaknowledge.

In conclusion, communication visibility may havemany important consequences for the way work isconducted in the next several decades. Today, enter-prise social media tools enable communication visi-bility at unprecedented levels. As technology design-ers and developers continue to improve these tools,develop streamlining applications, and devise algo-rithms that make them more efficient, we are likely tosee even more communication visibility in the future.This paper has developed a tentative theory of com-munication visibility from a qualitative field studyon enterprise social networking site use in one largeorganization. A good deal of work is needed to refinethis theory, introduce scope, and test its predictionsin varied organizational contexts. Moving the theoryforward in this way will foster a greater understand-ing of how the increasingly public and visible natureof our communication is shifting the ways in whichpeople work.

AcknowledgmentsThe author thanks Jonny Holmström, Kalle Lyytinen,and three anonymous reviewers for comments that have

improved this manuscript, as well as Jeffrey Treem, CaseyPierce, and Stephanie Dailey who helped with data collec-tion. Generous funding for this study was provided by agrant from the National Science Foundation (SES-1331492).

ReferencesAllen TJ (1977) Managing the Flow of Technology (MIT Press, Cam-

bridge, MA).Argote L, Gruenfeld DH, Naquin C (2000) Group learning in orga-

nizations. Turner M, ed. Groups at Work (Lawrence Erlbaum,Hillsdale, NJ), 369–411.

Argote L, Ingram P, Levine JM, Moreland RL (2000) Knowledgetransfer in organizations: Learning from the experience of oth-ers. Organ. Behav. Human Decision Processes 82(1):1–8.

Austin JR (2003) Transactive memory in organizational groups: Theeffects of content, consensus, specialization, and accuracy ongroup performance. J. Applied Psych. 88(5):866–878.

Benbasat I, Goldstein DK, Mead M (1987) The case research strategyin studies of information systems. MIS Quart. 11(3):369–386.

Bowker G, Timmermans S, Star SL (1995) Infrastructure and organi-zational transformation: Classifying nurses’ work. OrlikowskiW, Walsham G, Jones M, DeGross JD, eds. Information Tech-nology and Changes in Organizational Work (Chapman and Hall,London), 344–370.

Brandon DP, Hollingshead AB (2004) Transactive memory systemsin organizations: Matching tasks, expertise, and people. Organ.Sci. 15(6):633–644.

Brown JS, Duguid P (2000) The Social Life of Information (HarvardBusiness School Press, Cambridge, MA).

Burt RS (2004) Structural holes and good ideas. Amer. J. Sociol.110(2):349–399.

Choi SY, Lee H, Yoo Y (2010) The impact of information technol-ogy and transactive memory systems on knowledge sharing,application, and team performance: A field study. MIS Quart.34(4):855–870.

Ciborra C (1997) De profundis? Deconstructing the concept ofstrategic alignment. Scandinavian J. Inform. Systems 9(1):67–82.

Cramton CD, Orvis KL, Wilson JM (2007) Situation invisibilityand attribution in distributed collaborations. J. Management33(4):525–546.

Cress U, Kimmerle J, Hesse FW (2006) Information exchange withshared databases as a social dilemma: The effect of metaknowl-edge, bonus systems, and costs. Comm. Res. 33(5):370–390.

Cummings JN, Espinosa JA, Pickering CK (2009) Crossing spa-tial and temporal boundaries in globally distributed projects:A relational model of coordination delay. Inform. Systems Res.20(3):420–439.

Cyert RM, March JG (1963) A Behavioral Theory of the Firm, 2nd ed.(Prentice Hall, New York).

Dey AK, de Guzman E (2006) From awareness to connectedness:The design and deployment of presence displays. Grinter R,Rodden T, Aoki P, Cutrell E, Jeffries R, Olson G, eds. Proc.Sigchi Conf. Human Factors Comput. Systems (ACM Press, NewYork), 899–908.

DiMicco J, Millen DR, Geyer W, Dugan C, Brownholtz B, MullerM (2008) Motivations for social networking at work. Proc. 2008ACM Conf. Comput. Supported Cooperative Work (ACM Press,New York), 711–720.

Dourish P, Bellotti V (1992) Awareness and coordination in sharedworkspaces. Turner J, Kraut RE, eds. CSCW’92: Proc. Conf.Comput.-Supported Cooperative Work (ACM Press, New York),107–114.

Erickson I (2010) Geography and community: New forms ofinteraction among people and places. Amer. Behav. Scientist53(8):1194–1207.

INFORMS

holdsco

pyrightto

this

artic

lean

ddistrib

uted

this

copy

asaco

urtesy

totheau

thor(s).

Add

ition

alinform

ation,

includ

ingrig

htsan

dpe

rmission

policies,

isav

ailableat

https://p

ubso

nline.inform

s.org/.


Fleiss JL (1971) Measuring nominal scale agreement among manyraters. Psych. Bull. 76(5):378–382.

Gardner W (1992) Lessons in organizational dramaturgy: The art ofimpression management. Organ. Dynam. 21(1):33–46.

Gibbs JL, Rozaidi NA, Eisenberg J (2013) Overcoming the “ideol-ogy of openness”: Probing the affordances of social media fororganizational knowledge sharing. J. Comput. Mediated Comm.19(1):102–120.

Gioia DA, Manz CC (1985) Linking cognition and behavior: A scriptprocessing interpretation of vicarious learning. Acad. Manage-ment Rev. 10(3):527–539.

Glaser B (1965) The constant comparative method of qualitativeanalysis. Soc. Problems 12(4):436–445.

Glaser B (1978) Theoretical Sensitivity (The Sociological Press, MillValley, CA).

Goffman E (1959) The Presentation of Self in Everyday Life (DoubledayAnchor, Garden City, NY).

Hampton KN, Lee C-J, Her EJ (2011) How new media affords net-work diversity: Direct and mediated access to social capitalthrough participation in local social settings. New Media Soc.13(7):1031–1049.

Hansen MT (1999) The search-transfer problem: The role of weakties in sharing knowledge across organization subunits. Admin.Sci. Quart. 44(1):82–111.

Hargadon AB (2002) Brokering knowledge: Linking learning andinnovation. Res. Organ. Behav. 24:41–85.

Heald MR, Contractor N, Koehly LM, Wasserman S (1998) For-mal and emergent predictors of coworkers’ perceptual congru-ence on an organization’s social structure. Human Comm. Res.24(4):536–563.

Hollingshead AB (1998a) Communication, learning, and retrievalin transactive memory systems. J. Experimental Soc. Psych.34(5):423–442.

Hollingshead AB (1998b) Retrieval processes in transactive memorysystems. J. Personality Soc. Psych. 74(3):659–671.

Jackson P, Klobas J (2008) Transactive memory systems in organiza-tions: Implications for knowledge directories. Decision SupportSystems 44(2):409–424.

Jarrahi MH, Sawyer S (2013) Social technologies, informal knowl-edge practices, and the enterprise. J. Organ. Comput. ElectronicCommerce 23(1):110–137.

Jarvenpaa SL, Majchrzak A (2008) Knowledge collaboration amongprofessionals protecting nation security: Role of transactivememories in ego-centered knowledge networks. Organ. Sci.19(2):260–276.

Kankanhalli A, Tan BCY, Wei K-K (2005) Contributing knowl-edge to electronic repositories: An empirical investigation. MISQuart. 29(1):113–143.

Katila R, Chen E (2008) Effects of search timing on innovation:The value of not being in sync with rivals. Admin. Sci. Quart.53(4):593–625.

Keen A (2012) Digital Vertigo (St. Martin’s Press, New York).Keitzmann JH, Hermlens K, McCarthy IP, Silvester BS (2011) Social

media? Get serious! Understanding the functional buildingblocks of social media. Bus. Horizons 54(3):241–251.

Kim J-Y, Miner AS (2007) Vicarious learning from the failures andnear-failures of others: Evidence from the U.S. commercialbanking industry. Acad. Management J. 60(3):687–714.

Kinnaird P, Dabbish L, Kiesler S (2012) Workflow transparency ina microtask marketplace. Proc. Group ’12 (ACM Press, NewYork), 281–284.

Lampe C, Ellison N, Steinfield C (2006) A face (book) in the crowd:Social searching vs. social browsing. Proc. Annual Conf. Comput.Supported Cooperative Work (ACM Press, New York), 167–170.

Lapré MA, Van Wassenhove LN (2001) Creating and transferringknowledge for productivity improvement in factories. Manage-ment Sci. 47(10):1311–1325.

Leonardi PM, Treem JW (2012) Knowledge management technol-ogy as a stage for strategic self-presentation: Implications forknowledge sharing in organizations. Inform. Organ. 22(1):37–59.

Leonardi PM, Huysman M, Steinfield C (2013) Enterprise socialmedia: Definition, history, and prospects for the study ofsocial technologies in organizations. J. Comput.-Mediated Comm.19(1):1–19.

Liang D, Moreland R, Argote L (1995) Group versus individualtraining and group performance: The mediating role of trans-active memory. Personality Soc. Psych. Bull. 21(4):384–393.

Liebeskind JP (1996) Knowledge, strategy and the theory of thefirm. Strategic Management J. 17(2):93–107.

Majchrzak A, Cooper LP, Neece OE (2004) Knowledge reuse forinnovation. Management Sci. 50(2):174–188.

March JG (1991) Exploration and exploitation in organizationallearning. Organ. Sci. 2(1):71–87.

Miles MB, Huberman AM (1994) Qualitative Data Analysis: AnExpanded Sourcebook, 2nd ed. (Sage Publications, ThousandOaks, CA).

Nardi BA, Engeström Y (1999) A web on the wind: The structureof invisible work. Comput. Supported Cooperative Work 8(1):1–8.

Newell S, Robertson M, Scarbrough H, Swan J (2009) Manag-ing Knowledge Work and Innovation (Palgrave Macmillan, Bas-ingstoke, UK).

Palazzolo ET, Serb DA, She YC, Su CK, Contractor NS (2006)Coevolution of communication and knowledge networks intransactive memory systems: Using computational models fortheoretical development. Comm. Theory 16(2):223–250.

Peltokorpi V (2008) Transactive memory systems. Rev. GeneralPsych. 12(4):378–394.

Reagans R, McEvily B (2003) Network structure and knowledgetransfer: The effects of cohesion and range. Admin. Sci. Quart.48(2):240–267.

Reagans R, Argote L, Brooks D (2005) Individual experience andexperience working together: Predicting learning rates fromknowing who knows what and knowing how to work together.Management Sci. 51(6):869–881.

Ren Y, Argote L (2011) Transactive memory systems 1985–2010:An integrative framework of key dimensions, antecedents, andconsequences. Acad. Management Ann. 5(1):189–229.

Ren Y, Carley KM, Argote L (2006) The contingent effects of trans-active memory: When is it more beneficial to know what othersknow? Management Sci. 52(5):671–682.

Strauss A, Corbin J (1998) Basics of Qualitative Research: Techniquesand Procedures for Developing Grounded Theory, 2nd ed. (SagePublications, Thousand Oaks, CA).

Suchman L (2007) Human–Machine Reconfigurations: Plans and Situ-ated Actions (Cambridge University Press, New York).

Szulanski G, Jensen RJ (2006) Presumptive adaptation and theeffectiveness of knowledge transfer. Strategic Management J.27(10):937–957.

Tortoriello M, Reagans R, McEvily B (2012) Bridging the knowl-edge gap: The influence of strong ties, network cohesion, andnetwork range on the transfer of knowledge between organi-zational units. Organ. Sci. 23(4):1024–1039.

Treem JW, Leonardi PM (2012) Social media use in organizations:Exploring the affordances of visibility, editability, persistence,and association. Comm. Yearbook 36:143–189.

Von Hippel E, Tyre MJ (1995) How learning by doing is done:Problem identification in novel process equipment. Res. Policy24(1):1–12.

Walsham G (2002) What can knowledge management systemsdeliver? Management Comm. Quart. 16(2):267–273.

INFORMS

holdsco

pyrightto

this

artic

lean

ddistrib

uted

this

copy

asaco

urtesy

totheau

thor(s).

Add

ition

alinform

ation,

includ

ingrig

htsan

dpe

rmission

policies,

isav

ailableat

https://p

ubso

nline.inform

s.org/.


Wasko MM, Faraj S (2005) Why should I share? Examining socialcapital and knowledge contribution in electronic networks ofpractice. MIS Quart. 29(1):35–57.

Wegner DM (1987) Transactive memory: A contemporary analysisof group mind. Mullen B, Goethals G, eds. Theories of GroupBehavior (Springer-Verlag, New York), 185–208.

Wegner DM (1995) A computer network model of human transac-tive memory. Soc. Cognition 13(3):319–339.

Weick KE (1995) Sensemaking in Organizations (Sage Publications,Thousand Oaks, CA).

Yuan Y, Fulk J, Shumate M, Monge PR, Bryant JA, Matsaganis M(2005) Individual participation in organizational informa-tion commons. The impact of team level social influenceand technology-specific competence. Human Comm. Res. 31(2):212–240.

Yuan YC, Fulk J, Monge PR (2007) Access to information in connec-tive and communal transactive memory systems. Comm. Res.34(2):131–155.

Zuboff S (1988) In the Age of the Smart Machine: The Future of Workand Power (Basic Books, New York).

Social Media, Knowledge Sharing, and Innovation: Toward a ... · shape knowledge sharing in...

Documents

Transcript of Social Media, Knowledge Sharing, and Innovation: Toward a ... · shape knowledge sharing in...