Wise executives tailor their approach to fit the complexity of theDocument Transcript
Wise executives tailor their approach to fit the complexity of thecircumstances they face.In January 1993, a gunman murdered sevenpeople in a fast-food restaurant in Palatine, asuburb of Chicago. In his dual roles as an administrativeexecutive and spokesperson forthe police department, Deputy Chief WalterGasior suddenly had to cope with severaldifferent situations at once. He had to dealwith the grieving families and a frightenedcommunity, help direct the operations of anextremely busy police department, and takequestions from the media, which inundatedthe town with reporters and film crews.“There would literally be four people comingat me with logistics and media issues all atonce,” he recalls. “And in the midst of all this,we still had a department that had to keeprunning on a routine basis.”Though Gasior was ultimately successful injuggling multiple demands, not all leadersachieve the desired results when they facesituations that require a variety of decisionsand responses. All too often, managers relyon common leadership approaches that workwell in one set of circumstances but fall shortin others. Why do these approaches fail evenwhen logic indicates they should prevail?The answer lies in a fundamental assumptionof organizational theory and practice: that acertain level of predictability and order existsin the world. This assumption, groundedin the Newtonian science that underlies scientificmanagement, encourages simplificationsthat are useful in ordered circumstances.Circumstances change, however, and as theybecome more complex, the simplificationscan fail. Good leadership is not a one-sizefits-all proposition.We believe the time has come to broadenthe traditional approach to leadership anddecision making and form a new perspectivebased on complexity science. (For more onthis, see the sidebar “Understanding Complexity.”)Over the past ten years, we have appliedthe principles of that science to governmentsand a broad range of industries. Working withother contributors, we developed the Cynefinframework, which allows executives to seethings from new viewpoints, assimilate complexThis article made available with compliments of COGNITIVE EDGE PTE LTD. Further posting, copying ordistributing is copyright infringement. To order more copies go to www.hbr.org or call 800-988-0886.A Leader’s Framework for Decision Makingharvard business review • november 2007 page 2concepts, and address real-world problemsand opportunities. (Cynefin, pronouncedkunev-in, is a Welsh word that signifies themultiple factors in our environment and ourexperience that influence us in ways we cannever understand.) Using this approach,leaders learn to define the framework withexamples from their own organization’s historyand scenarios of its possible future. This
enhances communication and helps executivesrapidly understand the context in whichthey are operating.The U.S. Defense Advanced ResearchProjects Agency has applied the frameworkto counterterrorism, and it is currently a keycomponent of Singapore’s Risk Assessmentand Horizon Scanning program. Over time,the framework has evolved through hundredsof applications, from helping a pharmaceuticalcompany develop a new productstrategy to assisting a Canadian provincialgovernment in its efforts to engage employeesin policy making.The framework sorts the issues facingleaders into five contexts defined by the natureof the relationship between cause and effect.Four of these—simple, complicated, complex,and chaotic—require leaders to diagnosesituations and to act in contextually appropriateways. The fifth—disorder—applies whenit is unclear which of the other four contextsis predominant.Using the Cynefin framework can help executivessense which context they are in sothat they can not only make better decisionsbut also avoid the problems that arise whentheir preferred management style causesthem to make mistakes. In this article, wefocus on the first four contexts, offering examplesand suggestions about how to lead andmake appropriate decisions in each of them.Since the complex domain is much moreprevalent in the business world than mostleaders realize—and requires different, oftencounterintuitive, responses—we concentrateparticularly on that context. Leaders whounderstand that the world is often irrationaland unpredictable will find the Cynefinframework particularly useful.Simple Contexts: The Domain ofBest PracticeSimple contexts are characterized by stabilityand clear cause-and-effect relationships thatare easily discernible by everyone. Often, theright answer is self-evident and undisputed. Inthis realm of “known knowns,” decisions areunquestioned because all parties share an understanding.Areas that are little subject tochange, such as problems with order processingand fulfillment, usually belong here.Simple contexts, properly assessed, requirestraightforward management and monitoring.Here, leaders sense, categorize, and respond.That is, they assess the facts of the situation,categorize them, and then base their responseon established practice. Heavily processorientedsituations, such as loan paymentprocessing, are often simple contexts. If somethinggoes awry, an employee can usuallyidentify the problem (when, say, a borrowerpays less than is required), categorize it(review the loan documents to see how partialpayments must be processed), and respondappropriately (either not accept the paymentor apply the funds according to the termsof the note). Since both managers and employees
have access to the information necessaryfor dealing with the situation in this domain,a command-and-control style for settingparameters works best. Directives are straightforward,decisions can be easily delegated,and functions are automated. Adhering tobest practices or process reengineering makessense. Exhaustive communication amongmanagers and employees is not usually requiredbecause disagreement about whatneeds to be done is rare.Nevertheless, problems can arise in simplecontexts. First, issues may be incorrectly classifiedwithin this domain because they havebeen oversimplified. Leaders who constantlyask for condensed information, regardless ofthe complexity of the situation, particularlyrun this risk.Second, leaders are susceptible to entrainedthinking, a conditioned response that occurswhen people are blinded to new ways of thinkingby the perspectives they acquired throughpast experience, training, and success.Third, when things appear to be goingsmoothly, leaders often become complacent.If the context changes at that point, a leaderis likely to miss what is happening and reacttoo late. In the exhibit “The Cynefin Framework,”the simple domain lies adjacent tothe chaotic—and for good reason. The mostfrequent collapses into chaos occur becauseThis article made available with compliments of COGNITIVE EDGE PTE LTD. Further posting, copying ordistributing is copyright infringement. To order more copies go to www.hbr.org or call 800-988-0886.A Leader’s Framework for Decision Makingharvard business review • november 2007 page 3success has bred complacency. This shift canbring about catastrophic failure—think ofthe many previously dominant technologiesthat were suddenly disrupted by moredynamic alternatives.Leaders need to avoid micromanaging andstay connected to what is happening in orderto spot a change in context. By and large,line workers in a simple situation are morethan capable of independently handlingany issues that may arise. Indeed, thosewith years of experience also have deep insightinto how the work should be done. Leadersshould create a communication channel—ananonymous one, if necessary—that allowsdissenters to provide early warnings aboutcomplacency.Finally, it’s important to remember thatbest practice is, by definition, past practice.Using best practices is common, and often
appropriate, in simple contexts. Difficultiesarise, however, if staff members are discouragedfrom bucking the process even whenit’s not working anymore. Since hindsight nolonger leads to foresight after a shift in context,a corresponding change in managementstyle may be called for.Complicated Contexts: The Domainof ExpertsComplicated contexts, unlike simple ones,may contain multiple right answers, andthough there is a clear relationship betweencause and effect, not everyone can see it. Thisis the realm of “known unknowns.” Whileleaders in a simple context must sense, categorize,and respond to a situation, those in acomplicated context must sense, analyze, andrespond. This approach is not easy and oftenrequires expertise: A motorist may know thatsomething is wrong with his car because theengine is knocking, but he has to take it to amechanic to diagnose the problem.Because the complicated context calls forinvestigating several options—many of whichmay be excellent—good practice, as opposed tobest practice, is more appropriate. For example,the customary approach to engineering aUnderstanding ComplexityComplexity is more a way of thinking aboutthe world than a new way of working withmathematical models. Over a century ago,Frederick Winslow Taylor, the father of scientificmanagement, revolutionized leadership.Today, advances in complexity science, combinedwith knowledge from the cognitivesciences, are transforming the field onceagain. Complexity is poised to help currentand future leaders make sense of advancedtechnology, globalization, intricate markets,cultural change, and much more. In short,the science of complexity can help all of usaddress the challenges and opportunities weface in a new epoch of human history.A complex system has the following characteristics:•It involves large numbers of interactingelements.•The interactions are nonlinear, andminor changes can produce disproportionatelymajor consequences.•The system is dynamic, the whole isgreater than the sum of its parts, andsolutions can’t be imposed; rather, theyarise from the circumstances. This isfrequently referred to asemergence.•The system has a history, and the pastis integrated with the present; the elementsevolve with one another andwith the environment; and evolutionis irreversible.•Though a complex system may, in retrospect,appear to be ordered and predictable,
hindsight does not lead to foresightbecause the external conditions andsystems constantly change.•Unlike in ordered systems (where thesystem constrains the agents), or chaoticsystems (where there are no constraints),in a complex system the agents and thesystem constrain one another, especiallyover time. This means that we cannotforecast or predict what will happen.One of the early theories of complexity isthat complex phenomena arise from simplerules. Consider the rules for the flockingbehavior of birds: Fly to the center of theflock, match speed, and avoid collision. Thissimple-rule theory was applied to industrialmodeling and production early on, and itpromised much; but it did not deliver inisolation. More recently, some thinkers andpractitioners have started to argue thathuman complex systems are very differentfrom those in nature and cannot be modeledin the same ways because of human unpredictabilityand intellect. Consider the followingways in which humans are distinct fromother animals:•They have multiple identities and canfluidly switch between them withoutconscious thought. (For example, a personcan be a respected member of thecommunity as well as a terrorist.)•They make decisions based on pastpatterns of success and failure, ratherthan on logical, definable rules.•They can, in certain circumstances, purposefullychange the systems in whichthey operate to equilibrium states (thinkof a Six Sigma project) in order to createpredictable outcomes.Leaders who want to apply the principlesof complexity science to their organizationswill need to think and act differently thanthey have in the past. This may not be easy,but it is essential in complex contexts.This article made available with compliments of COGNITIVE EDGE PTE LTD. Further posting, copying ordistributing is copyright infringement. To order more copies go to www.hbr.org or call 800-988-0886.A Leader’s Framework for Decision Makingharvard business review • november 2007 page 4new cell phone might emphasize feature A overfeature B, but an alternative plan—emphasizingfeature C—might be equally valuable.Another example is the search for oil ormineral deposits. The effort usually requires ateam of experts, more than one place will potentiallyproduce results, and the location ofthe right spots for drilling or mining involvescomplicated analysis and understanding ofconsequences at multiple levels.Entrained thinking is a danger in complicatedcontexts, too, but it is the experts(rather than the leaders) who are prone toit, and they tend to dominate the domain.When this problem occurs, innovative suggestionsby nonexperts may be overlooked ordismissed, resulting in lost opportunities. Theexperts have, after all, invested in building
their knowledge, and they are unlikely totolerate controversial ideas. If the contexthas shifted, however, the leader may needaccess to those maverick concepts. To getaround this issue, a leader must listen to theexperts while simultaneously welcomingnovel thoughts and solutions from others.Executives at one shoe manufacturer didthis by opening up the brainstorming processfor new shoe styles to the entire company.As a result, a security guard submitteda design for a shoe that became one of theirbest sellers.Another potential obstacle is “analysisparalysis,” where a group of experts hits astalemate, unable to agree on any answersbecause of each individual’s entrainedthinking—or ego.Working in unfamiliar environments canhelp leaders and experts approach decisionmaking more creatively. For instance, we putretail marketing professionals in several militaryresearch environments for two weeks.The settings were unfamiliar and challenging,but they shared a primary similarity with theretail environment: In both cases, the marketershad to work with large volumes of datafrom which it was critical to identify smalltrends or weak signals. They discovered thatthere was little difference between, say,handling outgoing disaffected customers andanticipating incoming ballistic missiles. Theexercise helped the marketing group learnhow to detect a potential loss of loyaltyand take action before a valued customerswitched to a competitor. By improving theirstrategy, the marketers were able to retainfar more high-volume business.Games, too, can encourage novel thinking.We created a game played on a fictionalplanet that was based on the culture of a realclient organization. When the executives“landed” on the alien planet, they were askedto address problems and opportunities facingthe inhabitants. The issues they encounteredwere disguised but designed to mirror realsituations, many of which were controversialor sensitive. Because the environment seemedso foreign and remote, however, the playersfound it much easier to come up with freshideas than they otherwise might have done.Playing a metaphorical game increases managers’willingness to experiment, allows themto resolve issues or problems more easilyThe Cynefin FrameworkThe Cynefin framework helps leadersdetermine the prevailing operative contextso that they can make appropriatechoices. Each domain requires differentactions.Simpleandcomplicatedcontextsassume an ordered universe, wherecause-and-effect relationships are perceptible,and right answers can be determined
based on the facts.Complexandchaoticcontexts are unordered—there isno immediately apparent relationshipbetween cause and effect, and the wayforward is determined based on emergingpatterns. The ordered world is theworld of fact-based management; theunordered world represents patternbasedmanagement.The very nature of the fifth context—disorder—makes it particularly difficult torecognize when one is in it. Here, multipleperspectives jostle for prominence,factional leaders argue with one another,and cacophony rules. The way out of thisrealm is to break down the situation intoconstituent parts and assign each to oneof the other four realms. Leaders canthen make decisions and intervene incontextually appropriate ways.This article made available with compliments of COGNITIVE EDGE PTE LTD. Further posting, copying ordistributing is copyright infringement. To order more copies go to www.hbr.org or call 800-988-0886.A Leader’s Framework for Decision Makingharvard business review • november 2007 page 5and creatively, and broadens the range ofoptions in their decision-making processes.The goal of such games is to get as manyperspectives as possible to promote unfetteredanalysis.Reaching decisions in the complicated domaincan often take a lot of time, and there isalways a trade-off between finding the rightanswer and simply making a decision. Whenthe right answer is elusive, however, and youmust base your decision on incomplete data,your situation is probably complex ratherthan complicated.Complex Contexts: The Domain ofEmergenceIn a complicated context, at least one rightanswer exists. In a complex context, however,right answers can’t be ferreted out. It’s like thedifference between, say, a Ferrari and theBrazilian rainforest. Ferraris are complicatedmachines, but an expert mechanic can takeone apart and reassemble it without changinga thing. The car is static, and the whole is thesum of its parts. The rainforest, on the otherhand, is in constant flux—a species becomesextinct, weather patterns change, an agriculturalproject reroutes a water source—and thewhole is far more than the sum of its parts.This is the realm of “unknown unknowns,” andit is the domain to which much of contemporarybusiness has shifted.Most situations and decisions in organizationsare complex because some majorchange—a bad quarter, a shift in management,a merger or acquisition—introduces unpredictabilityand flux. In this domain, we can understandwhy things happen only in retrospect.Instructive patterns, however, can emerge ifthe leader conducts experiments that are safeto fail. That is why, instead of attempting
to impose a course of action, leaders mustpatiently allow the path forward to revealitself. They need to probe first, then sense,and then respond.There is a scene in the film Apollo 13 whenthe astronauts encounter a crisis (“Houston,we have a problem”) that moves the situationinto a complex domain. A group of experts isput in a room with a mishmash of materials—bits of plastic and odds and ends that mirrorthe resources available to the astronauts inflight. Leaders tell the team: This is whatyou have; find a solution or the astronautswill die. None of those experts knew a prioriwhat would work. Instead, they had to let asolution emerge from the materials at hand.And they succeeded. (Conditions of scarcityoften produce more creative results thanconditions of abundance.)Another example comes from YouTube.The founders could not possibly have predictedall the applications for streamingvideo technology that now exist. Once peoplestarted using YouTube creatively, however,the company could support and augment theemerging patterns of use. YouTube has becomea popular platform for expressing politicalviews, for example. The company built on thispattern by sponsoring a debate for presidentialhopefuls with video feeds from the site.As in the other contexts, leaders face severalchallenges in the complex domain. Ofprimary concern is the temptation to fallback into traditional command-and-controlmanagement styles—to demand fail-safebusiness plans with defined outcomes. Leaderswho don’t recognize that a complex domainrequires a more experimental mode of managementmay become impatient when theydon’t seem to be achieving the results theywere aiming for. They may also find it difficultto tolerate failure, which is an essential aspectof experimental understanding. If they tryto overcontrol the organization, they willpreempt the opportunity for informativepatterns to emerge. Leaders who try to imposeorder in a complex context will fail, but thosewho set the stage, step back a bit, allowpatterns to emerge, and determine whichones are desirable will succeed. (See thesidebar “Tools for Managing in a ComplexContext.”) They will discern many opportunitiesfor innovation, creativity, and newbusiness models.Chaotic Contexts: The Domain ofRapid ResponseIn a chaotic context, searching for right answerswould be pointless: The relationshipsbetween cause and effect are impossible to determinebecause they shift constantly and nomanageable patterns exist—only turbulence.This is the realm of unknowables. The eventsof September 11, 2001, fall into this category.In the chaotic domain, a leader’s immediatejob is not to discover patterns but to stanch thebleeding. A leader must first act to establishThis article made available with compliments of COGNITIVE EDGE PTE LTD. Further posting, copying or
distributing is copyright infringement. To order more copies go to www.hbr.org or call 800-988-0886.A Leader’s Framework for Decision Makingharvard business review • november 2007 page 6order, then sense where stability is presentand from where it is absent, and then respondby working to transform the situation fromchaos to complexity, where the identificationof emerging patterns can both help preventfuture crises and discern new opportunities.Communication of the most direct top-downor broadcast kind is imperative; there’s simplyno time to ask for input.Unfortunately, most leadership “recipes”arise from examples of good crisis management.This is a mistake, and not only becausechaotic situations are mercifully rare. Thoughthe events of September 11 were not immediatelycomprehensible, the crisis demandeddecisive action. New York’s mayor at the time,Rudy Giuliani, demonstrated exceptionaleffectiveness under chaotic conditions byissuing directives and taking action to reestablishorder. However, in his role asmayor—certainly one of the most complexjobs in the world—he was widely criticizedfor the same top-down leadership style thatproved so enormously effective during thecatastrophe. He was also criticized afterwardfor suggesting that elections be postponed sohe could maintain order and stability. Indeed,a specific danger for leaders following a crisisis that some of them become less successfulwhen the context shifts because they are notable to switch styles to match it.Moreover, leaders who are highly successfulin chaotic contexts can develop an overinflatedself-image, becoming legends in theirown minds. When they generate cultlike adoration,leading actually becomes harder forthem because a circle of admiring supporterscuts them off from accurate information.Tools for Managing in a Complex ContextGiven the ambiguities of the complex domain,how can leaders lead effectively?• Open up the discussion.Complexcontexts require more interactivecommunication than any of the otherdomains. Large group methods (LGMs),for instance, are efficient approachesto initiating democratic, interactive,multidirectional discussion sessions.Here, people generate innovative ideasthat help leaders with development andexecution of complex decisions and strategies.For example, “positive deviance” isa type of LGM that allows people todiscuss solutions that are already workingwithin the organization itself, ratherthan looking to outside best practicesfor clues about how to proceed. ThePlexus Institute used this approach toaddress the complex problem of hospitalacquiredinfections, resulting in behaviorchange that lowered the incidence byas much as 50%.• Set barriers.Barriers limit or delineate
behavior. Once the barriers are set, thesystem can self-regulate within thoseboundaries. The founders of eBay, forexample, created barriers by establishinga simple set of rules. Among them arepay on time, deliver merchandisequickly, and provide full disclosure onthe condition of the merchandise.Participants police themselves byrating one another on the quality oftheir behavior.• Stimulate attractors.Attractors arephenomena that arise when small stimuliand probes (whether from leaders orothers) resonate with people. As attractorsgain momentum, they providestructure and coherence. EBay againprovides an illustrative example. In1995, founder Pierre Omidyar launchedan offering called Auction Web on hispersonal website. His probe, the firstitem for sale, quickly morphed intoeBay, a remarkable attractor for peoplewho want to buy and sell things. Today,sellers on eBay continue to provide experimentalprobes that create attractorsof various types. One such probe, sellinga car on the site, resonated with buyers,and soon automobile sales became apopular attractor.• Encourage dissent and diversity.Dissent and formal debate are valuablecommunication assets in complexcontexts because they encourage theemergence of well-forged patterns andideas. A “ritual dissent” approach, forinstance, puts parallel teams to work onthe same problem in a large groupmeeting environment. Each team appointsa spokesperson who movesfrom that team’s table to anotherteam’s table. The spokesperson presentsthe first group’s conclusions whilethe second group listens in silence. Thespokesperson then turns around to faceaway from the second team, which ripsinto the presentation, no holds barred,while the spokesperson listens quietly.Each team’s spokesperson visits othertables in turn; by the end of the session,all the ideas have been well dissectedand honed. Taking turns listening in silencehelps everyone understand thevalue of listening carefully, speakingopenly, and not taking criticism personally.• Manage starting conditions andmonitor for emergence.Because outcomesare unpredictable in a complexcontext, leaders need to focus on creatingan environment from which goodthings can emerge, rather than tryingto bring about predetermined resultsand possibly missing opportunitiesthat arise unexpectedly. Many yearsago, for instance, 3M instituted a rule allowingits researchers to spend 15% oftheir time on any project that interestedthem. One result was a runaway success:the Post-it Note.
This article made available with compliments of COGNITIVE EDGE PTE LTD. Further posting, copying ordistributing is copyright infringement. To order more copies go to www.hbr.org or call 800-988-0886.A Leader’s Framework for Decision Makingharvard business review • november 2007 page 7Decisions in Multiple Contexts: A Leader’s GuideEffective leaders learn to shift their decision-making styles to match changing business environments. Simple,complicated,complex, and chaotic contexts each call for different managerial responses. By correctly identifying the governingcontext,staying aware of danger signals, and avoiding inappropriate reactions, managers can lead effectively in a variety ofsituations.
This article made available with compliments of COGNITIVE EDGE PTE LTD. Further posting, copying ordistributing is copyright infringement. To order more copies go to www.hbr.org or call 800-988-0886.A Leader’s Framework for Decision Makingharvard business review • november 2007 page 8Yet the chaotic domain is nearly alwaysthe best place for leaders to impel innovation.People are more open to novelty anddirective leadership in these situations thanthey would be in other contexts. One excellenttechnique is to manage chaos and innovationin parallel: The minute you encounter a crisis,
appoint a reliable manager or crisis managementteam to resolve the issue. At the sametime, pick out a separate team and focus itsmembers on the opportunities for doing thingsdifferently. If you wait until the crisis is over,the chance will be gone.Leadership Across ContextsGood leadership requires openness to changeon an individual level. Truly adept leaders willknow not only how to identify the contextthey’re working in at any given time but alsohow to change their behavior and their decisionsto match that context. They also preparetheir organization to understand the differentcontexts and the conditions for transition betweenthem. Many leaders lead effectively—though usually in only one or two domains(not in all of them) and few, if any, preparetheir organizations for diverse contexts.During the Palatine murders of 1993, DeputyChief Gasior faced four contexts at once. Hehad to take immediate action via the mediato stem the tide of initial panic by keepingthe community informed (chaotic); he had tohelp keep the department running routinelyand according to established procedure(simple); he had to call in experts (complicated);and he had to continue to calm thecommunity in the days and weeks followingthe crime (complex). That last situationproved the most challenging. Parents wereafraid to let their children go to school, andemployees were concerned about safety intheir workplaces. Had Gasior misread thecontext as simple, he might just have said,“Carry on,” which would have done nothingto reassure the community. Had he misread itas complicated, he might have called in expertsto say it was safe—risking a loss of credibilityand trust. Instead, Gasior set up a forumfor business owners, high school students,teachers, and parents to share concerns andhear the facts. It was the right approach for acomplex context: He allowed solutions toemerge from the community itself ratherthan trying to impose them.•••Business schools and organizations equipleaders to operate in ordered domains (simpleand complicated), but most leaders usuallymust rely on their natural capabilities whenoperating in unordered contexts (complex andchaotic). In the face of greater complexitytoday, however, intuition, intellect, and charismaare no longer enough. Leaders needtools and approaches to guide their firmsthrough less familiar waters.In the complex environment of the currentbusiness world, leaders often will be calledupon to act against their instincts. They willneed to know when to share power and whento wield it alone, when to look to the wisdomof the group and when to take their owncounsel. A deep understanding of context, theability to embrace complexity and paradox,and a willingness to flexibly change leadershipstyle will be required for leaders who
want to make things happen in a time ofincreasing uncertainty.Reprint R0711CTo order, see the next pageor call 800-988-0886 or 617-783-7500or go to www.hbrreprints.orgThis article made available with compliments of COGNITIVE EDGE PTE LTD. Further posting, copying ordistributing is copyright infringement. To order more copies go to www.hbr.org or call 800-988-0886.Further ReadingTo Orderpage 9TheHarvard Business ReviewPaperback SeriesHere are the landmark ideas—bothcontemporary and classic—that haveestablished Harvard Business Review as requiredreading for businesspeople around the globe.Each paperback includes eight of the leadingarticles on a particular business topic. Theseries includes over thirty titles, including thefollowing best-sellers:Harvard Business Review on