hbr need to solve a problem - take a break from collaborating
DESCRIPTION
Harvard Business Review Article--public domainTRANSCRIPT
04 May 2015 Research & Ideas
Need to Solve a Problem? Take a Break FromCollaborating
Organizations spend a lot ofmoney enabling employees tosolve problems collectively. Butinducing more collaboration mayactually hinder the mostimportant part ofproblem-solving: actually solvingthe problem. Research by JesseShore, Ethan Bernstein, andDavid Lazer.
by Carmen Nobel
Got a problem? Throw somecollaboration software at it.
It's a common strategy amongtoday's managers: Organizationsspend a lot of money ontechnology that enables employeesto tackle problems collectively.Hence, the market is rife withconnectivity tools and servicessuch as Salesforce.com's Chatter,Microsoft's Yammer, and JiveSoftware's suite of namesakeproducts. The global enterprisesocial software market is forecastto grow to $8.14 billion in 2019,according to MarketWatch, from$4.77 billion in 2014.
"When it comes to solvingproblems, connectedness isa double-edged sword"
Unfortunately, it turns out thatinducing more collaboration mayhinder the most important part ofproblem-solving: actually solvingthe problem. While connectingemployees does increase the abilityto gather facts during the earlystages of tackling a problem, it alsoinhibits the ability to analyze thosefacts and find a solution, a team ofBoston researchers reports.
"When it comes to solvingproblems, connectedness is adouble-edged sword," saysHarvard Business School AssistantProfessor Ethan Bernstein,co-author of Facts and Figuring:An Experimental Investigation ofNetwork Structure andPerformance in Information andSolution Spaces, with Jesse Shore,an assistant professor at BostonUniversity's Questrom School ofBusiness, and David Lazer, apolitical science professor atNortheastern University. The paperappears in the journal OrganizationScience.
A HIGH-STAKES
WHODUNIT GAME
Previous academic researchgenerally has focused separately ontwo aspects ofproblem-solving—either gatheringfacts or figuring out solutions. Andthe findings often havecontradicted each other.
"Substantial recent researchimplies that clustering—the degreeto which people with whom aperson is connected are themselvesconnected to each other—canimprove problem-solvingperformance by increasingcoordination," the authors write inthe "Facts and Figuring" paper."By contrast, equally powerfulresearch suggests that clusteringcan undermine performance byfostering an unproductiveimbalance between exploration andexploitation, even for simpletasks."
Thus, Bernstein, Shore, and Lazerset out to figure out howcollaborating could both help andhurt the problem-solving process.To pursue the efficacy ofcollaboration the research teamdeveloped a straightforwardexperiment that mirroredreal-world problem-solving work.
COPYRIGHT 2013 PRESIDENT AND FELLOWS OF HARVARD COLLEGE 1
Rather than start from scratch, theycustomized a platform calledELICIT (Experimental Laboratoryfor Investigating Collaboration,Information-sharing, and Trust),developed by the US Departmentof Defense's Command andControl Research Program. In theDoD's high-stakes "whodunit"game, players try to solve severalaspects of an imaginary pendingterrorist attack: the identity of theterrorists, the target of the attack,and where and when the attackwould happen.
For the study, the researchers hired417 participants to play the game.Players received two clues at thestart of each round and wereallowed to search for more cluesonce per minute; they had 25minutes to solve the problem. (Theexperiment took place at HarvardBusiness School's Computer Labfor Experimental Research.)
Participants were randomlyassigned to one of 70 16-personnetworks, some of which weremore interconnected—or"clustered," in academicparlance—in terms of who couldshare information with whomduring the game. "In themost-clustered conditions, peoplewere connected in a clear teamstructure," Shore explains. "In theleast clustered, nobody's partnerswere also partners with eachother."
Participants received 60¢ for eachminute (equivalent to $36 per hour)that they had the correct answerregistered before the clock ranout—a considerable incentive,considering the majority of theparticipants were undergraduateswho needed pocket money.
"Generally speaking, these werestudents from northeasternuniversities, all of whom arerelatively smart, so we're nottalking about the averagepopulation," Bernstein says. "Butactually, that's probably good forus, because the average populationis not who would be hired to solveintelligence problems for theDepartment of Defense."
In terms of seeking out uniquefacts, the results showed that themost-clustered groups gathered 5percent more information than theleast-clustered groups, becauseclustering prevented networkmembers from unknowinglyconducting duplicative searches."By being in a cluster, individualstended to contribute more to thecollective exploration throughinformation space—not from moresearch but rather by being morecoordinated in their search," theresearchers write.
However, clustering also seemed toinhibit the breadth and number ofanswers that the players proposed.The least-connected networkscame up with 17.5 percent moretheories and solutions than did themost-connected networks. Lessclustering also increased thelikelihood of correct solutions inthat those in very clusteredpositions were more likely to copyan incorrect theory from a neighborthan their less-clusteredcounterparts.
"We realized that the networkstructure seemed to have oppositeeffects for searching forinformation and searching forsolutions," Shore says. "That wassort of the 'aha' moment."
IMPLICATIONS FORTHE WORKPLACE
The team's findings highlight thepros and cons of ever-presentconnectivity. On the one hand,tools like collaboration softwarecan facilitate coordinatedinformation gathering at both alocal and a global scale. On theother hand, too much connectivitycan lead to a premature consensus,to the detriment of organizationsthat aim to transform informationinto knowledge. Fostering effectiveproblem-solving requiresflexibility—flexible technology,flexible employees, and a flexibleworkspace.
"The optimal communication andcollaboration structures changeover time," Shore says. "Tocomplicate matters, in longer-termprojects you're going to be iteratingbetween information gathering andsolution generating. So theorganizational structure needs toswitch back and forth to facilitatethat."
Bernstein notes that some firms dothat already. "These results seem tomake a lot of sense toorganizations like The BostonConsulting Group or McKinsey &Company, which deliberately usevery different structures fordifferent functional stages of aproject," he says. "Problem-solvinglooks different from other stages ofa project in which the team isdefining the problem, gatheringdata, or synthesizing results."
In terms of how to design corporateoffices, the findings in "Facts andFiguring" provide rationale for thehot architectural trend of "agile"
HARVARD BUSINESS SCHOOL | WORKING KNOWLEDGE | HBSWK.HBS.EDU
COPYRIGHT 2013 PRESIDENT AND FELLOWS OF HARVARD COLLEGE 2
workspaces, including a mix ofopen floors, private spaces, andmovable surfaces. "Architects havedesigned our workspaces flexiblyso that we can change them anduse them in different ways,"Bernstein says. "But nobody hastrained us on how to use them. Weneed to think about how to use ourspaces wisely."
For knowledge-intensivecompanies, the findings highlightthe need for enterprise softwarethat addresses the disparate aspectsof problem-solving.
"Another organizational responsewould be to designcommunications infrastructuresthat could somehow separate factsfrom figuring and adoptdifferently-structuredcommunication networks for eachcategory," the authors write. "Inother words, rather than allow themarch of technology to dictateorganizational performance, it ispossible to imagine technologybeing harnessed to achievedifferent performance goals. Evenwithout the separation of facts andfiguring, the results of this studyare likely to be especially relevant
for computer-mediatedproblem-solving because of theease of manipulating the structurein which participantscommunicate."
For managers, the findingshighlight the need to determinefrom the get-go whether aproblem-solving task requires asearch for facts or a search foranswers, and then, if possible,tackle the problemaccordingly—enforcingcollaboration only where it makessense. "That also means that theircollaboration tools, and the policiesthat others set around them, need tohave an off switch—to be usedvery selectively, but at thediscretion of those at the front lineof problem-solving," Bernsteinsays.
And for academics, the researchprovides a good jumping-off pointfor future studies about physicaland virtual networking. In theirnext stage of research, the teamplans to conduct several fieldstudies in real-world officeenvironments as well as furtherexperimental studies on digitalcollaboration. In the meantime,
"Facts and Figuring" helps toexplain how knowledge works,both in the workplace and in theclassroom.
"Students sometimes want you justto tell them the answer—like 'justgive me the answer to leadership!'"says Bernstein, who co-teaches therequired Leadership andOrganizational Behavior course tofirst-year MBA students at HBS."We have never done it that way.We want to give them the facts,and we do, but we want them toexplore their own individualsolutions based on those facts.Because ultimately, that's going toget the field of study furtheralong—and get them furtheralong—over the course of time. Sothat's how we teach. But it's justnot the way we've designedorganizations, historically."
About the author
Carmen Nobel is senior editor ofHarvard Business School WorkingKnowledge.
HARVARD BUSINESS SCHOOL | WORKING KNOWLEDGE | HBSWK.HBS.EDU
COPYRIGHT 2013 PRESIDENT AND FELLOWS OF HARVARD COLLEGE 3