what's your strategy for managing …s your strategy for managing knowledge? ... boston...

Some companies automate knowledge management-others rely on their people to share knowledgethrough more traditional means. Emphasizing thewrong approach-or trying to pursue both at thesame time-can quickly undermine your business.

WHAT'S YOURSTRATEGY FOR

MANAGINGKNOWLEDGE?by Morten T. Hansen, NItin Nohria,and Thomas Tierney

KNOWLEDGE MANAGEMENT IS NOTHING NEW.For hundreds of years, owners of familybusinesses have passed their commercial

wisdom on to their children, master craftsmen havepainstakingly taught their trades to apprentices, andworkers have exchanged ideas and know-how on thejob. But it wasn't until the 1990s that chief executivesstarted talking about knowledge management. As thefoundation of industrialized economies has shifted fromnatural resources to intellectual assets, executives havebeen compelled to examine the knowledge underlying theirbusinesses and how that knowledge is used. At the sametime, the rise of networked computers has made it possibleto codify, store, and share certain kinds of knowledge moreeasily and cheaply than ever before.

Since knowledge management as a conscious practice is soyoung, executives have lacked successful models that they coulduse as guides. To help fill that gap, we have recently studied theknowledge management practices of companies in several indus-tries. We started by looking at management consulting firms. Be-cause knowledge is the core asset of consultancies, they were among

Morten T. Hansen is an assistant professor and Nitin Nobria is a professorof organizational behavior at Harvard Business School in Boston. Massa-chusetts. Thomas Tierney is the worldwide managing director of Bain &>Company in Boston.

7b discuss knowledge management strategies, join the authors in the HBRForum at http://www.hbr.org/forum.

106 ARTWORK BY THEO RUDNAK

the first businesses to pay attention to-andmake heavy investments in-the managementof knowledge. They were also among the first toaggressively explore the use of information tech-nology to capture and disseminate knowledge.Their experience, which is relevant to any com-pany that depends on smart people and the fiowof ideas, provides a window onto what worksand what doesn't.

Consultants, we found, do not take a uniformapproach to managing knowledge. The consultingbusiness employs two very different knowledgemanagement strategies. In some companies, thestrategy centers on the computer. Knowledge iscarefully codified and stored in databases, whereit can he accessed and used easily by anyone in thecompany. We call this the codification strategy. Inother companies, knowledge is closely tied to theperson who developed it and is shared mainlythrough direct person-to-person contacts. The chiefpurpose of computers at such companies is to helppeople communicate knowledge, not to store it. Wecall this the personalization strategy. A company'schoice of strategy is far from arbitrary-it depends onthe way the company serves its clients, the econom-ics of its business, and the people it hires. Emphasiz-ing the wrong strategy or trying to pursue both at thesame time can, as some consulting firms have found,quickly undermine a business.

The two strategies are not unique to consulting.When we looked beyond that business and analyzedcomputer companies and health care providers, wefound the same two strategies at work. In fact, we be-lieve that the choice between codification and person-alization is the central one facing virtually all compa-nies in the area of knowledge management. By betterunderstanding the two strategies and their strengthsand weaknesses, chief executives will be able to makemore surefooted decisions about knowledge manage-ment and their investments in it.

Codification or Personalization?Some large consulting companies, such as Andersen Con-sulting and Ernst & Young, have pursued a codificationstrategy. Over the last five years, they have developed

HARVARD BUSINESS REVIEW March-April 1999 107

KNOWLEDGE MANAGEMENT STRATEGY

elaborate ways to codify, store, and reuse knowl-edge. Knowledge is codified using a "people-to-doc-uments" approach: it is extracted from the personwho developed it, made independent of that person,and reused for various purposes. Ralph Poole, direc-tor of Ernst & Young's Center for Business Knowl-edge, describes it like this: "After removing client-sensitive information, we develop 'knowledgeobjects' by pulling key pieces of knowledge such as

interview guides, work

codification I^^ ' ^i^ 'lT'T*

the possibilityin111

thp

"Ki 1 ci -npcc

mentation analyses outof documents and storingthem in the electronic re-posnoryforpeoplctouse."This approach allows manypeople to search for and re-"ieve codified knowledgewithout having to contact

-i-nrl tVii ic r\f the person who originally de-d l l U U i U b U i veloped it. That opens up the

PO^ i i i y °^ achieving scaleij^ knowledge reuse and thus°^growing the business.

r^^^^ tj g example of RandallLove, a partner in the Los Ange-

les office of Ernst &. Young. Love was preparing animportant bid for a large industrial manufacturerthat needed help installing an enterprise resourceplanning system. He had already directed projectsfor implementing information systems for severalmanufacturers in other industries, but he hadn't yetworked on a manufacturing project in this one. Heknew other Ernst & Young teams had, however, sohe searched the electronic knowledge managementrepository for relevant knowledge. For help with thesales process, he found and used several presenta-tions on the industry-documents containing previ-ously developed solutions-as well as value proposi-tions that helped him estimate how much moneythe client would save by implementing the system.

Because Love reused this material, Ernst & Youngwon the project and closed the sale in two monthsinstead of the typical four to six. In addition, histeam found programming documents, technicalspecifications, training materials, and change man-agement documentation in the repository. Becausethese documents were availahle. Love and his teamdid not have to spend any time tracking down andtalking with the people who had first developedthem. The codification of such knowledge saved theteam and the elient one full year of work.

Ernst & Young executives have invested a lot tomake sure that the codification process works effi-

ciently. The 250 people at the Center for BusinessKnowledge manage the electronic repository andhelp consultants find and use information. Special-ists write reports and analyses that many teams canuse. And each of Ernst &. Young's more than 40practice areas has a staff member who helps codifyand store documents. The resulting area databasesare linked through a network.

Naturally, people-to-documents is not the onlyway consultants in firms like Ernst &. Young andAndersen Consulting share knowledge - they talkwith one another, of course. What is striking, how-ever, is the degree of emphasis they place on the cod-ification strategy.

By contrast, strategy consulting firms such asBain, Boston Consulting Group, and McKinsey em-phasize a personalization strategy. They focus on di-alogue between individuals, not knowledge objectsin a database. Knowledge that has not heen codified-and probably couldn't be - i s transferred in brain-storming sessions and one-on-one conversations.Consultants collectively arrive at deeper insights bygoing back and forth on problems they need to solve.

Marcia Blenko, for example, a partner in Bain'sLondon office, had to consider a difficult strategyproblem for a large British financial institution.The client wanted Bain to help it expand by offeringnew products and services. The assignment re-quired geographic and product-line expertise, abroad understanding of the industry, and a largedose of creative thinking. Blenko, who had beenwith Bain for 12 years, knew several partners withexpertise relevant to this particular prohlem. Sheleft voice mail messages with them and checkedBain's "people finder" database for more contacts.Eventually she connected with nine partners andseveral managers who had developed growth strate-gies for financial services institutions. She metwith a group of them in Europe, had videoconfer-ences with others from Singapore and Sydney, andmade a quick trip to Boston to attend a meeting ofthe financial services practice. A few of these col-leagues became ongoing advisers to the project, andone of the Asian managers was assigned full timeto the case team. During the next four months,Blenko and her team consulted with expert part-ners regularly in meetings and through phone callsand e-mail. In the process of developing a uniquegrowth strategy, the team tapped into a worldwidenetwork of colleagues' experience.

To make their personalization strategies work,firms like Bain invest heavily in building networksof people. Knowledge is shared not only face-to-facebut also over the telephone, hy e-mail, and via video-conferences. McKinsey fosters networks in many

108 HARVARD BUSINESS REVIEW March-April 1999


ways: by transferring people between offices,- bysupporting a culture in which consultants arc ex-pected to return phone calls from colleaguespromptly; by creating directories of experts; and byusing "consulting directors" within the firm to as-sist project teams.

These firms have also developed electronic docu-ment systems, but the purpose of the systems is notto provide knowledge objects. Instead, consultantsscan documents to get up to speed in a particulararea and to find out who has done work on a topic.They then approach those people directly.

When we initially looked at how consulting com-panies manage knowledge, we found that they all

used both the codification and the personalizationapproaches. When we dug deeper, however, wefound that effective firms excelled by focusing onone of the strategies and using the other in a sup-porting role. They did not try to use both approach-es to an equal degree.

Different Strategies, Different DriversA company's knowledge management strategyshould refiect its competitive strategy: how it cre-ates value for customers, how that value supports aneconomic model, and how the company's people de-hver on the value and the economics.

How Consulting Firms Manage Their Knowledge

CODIFICATION

Provide high-quality, reliable, and fastinformation-systems implementationby reusing codified knowledge.

CompetitiveStrategy

REUSE ECONOMICS:

Invest once in a knovifledge asset;reuse it many times.

Use large teams with a high ratio ofassociates to partners.

Focus on generating large overall revenues.

PEOPLE-TO-DOCUMENTS:

Oevelop an electronic document systemthat codifies, stores, disseminates, andallows reuse of knowledge.

Invest heavily in IT; the goal is to connectpeople with reusable codified knowledge.

Hire new college graduates who are well5uited to the reuse of knowledge and theimplementation of solutions.

Train people in groups and throughcomputer-based distance learning.

Reward people for using and contributingto document databases.

Andersen Consulting, Ernst & Young

EconomicModel

KnowledgeManagement

Strategy

InformationTechnology

HumanResources

PERSONALIZATION

Provide creative, analytically rigorousadvice on high-level strategic problemsby channeling individual expertise.

Examples

EXPERT ECONOMICS:

Charge high fees for highly customizedsolutions to unique problems.

Use small teams with a tow ratio ofassociates to partners.

Focus on maintaining high profit margins.

PERSON-TO-PERSON:

Develop networks for linking people so thattacit knowledge can be shared.

Invest moderately in IT; the goal is tofacilitate conversations and the exchangeof tacit knowledge.

Hire M.B.A.s who like problem solving andcan tolerate ambiguity.

Train people through one-on-onementoring.

Reward people for directly sharingknowledge with others.

McKinsey & Company, Sain & Company



A company'sstrategy forknowledge

managementshould reflect

its competitivestrategy

Creating Value for Customers. Randall Love'sapproach to implementing the information systemis typical of consulting companies where the effi-cient reuse of codified knowledge is essential be-cause they are dealing with similar prohlems over

and over. In suchfirm^s, the serviceoffering is very clear:the customer hene-fits hecause the con-sultants can huild areliahle, high-qualityinformation systemfaster and at a hetterprice than others hy us-ing work plans, soft-ware code, and solutionsthat have heen fine-tunedand proven successful.That's not to say that theprocess operates on auto-matic pilot. It's like huild-ing with Lego hlocks: con-

sultants reuse existing bricks while applying theirskills to construct something new.

Strategy consulting firms offer customers a verydifferent kind of value. Consultants like MarciaBlenko tackle problems that don't have clear solu-tions at the outset. They seek advice from colleaguesto deepen their understanding of the issues, hut inthe end they must create a highly customized solu-tion to a unique problem. Because their clients'problems are difficult and one of a kind, the consul-tants can charge high fees for their services.

Turning a Profit. Companies that follow a codifi-cation strategy rely on the "economics of reuse."Once a knowledge asset-software code or a man-ual, for example -is developed and paid for, it can beused many times over at very low cost, provided itdoes not have to be substantially modified eachtime it is used. Because the knowledge is containedin electronic repositories, it can be employed inmany jobs by many consultants. Many consultantscan be assigned to a project; big projects will have ahigh ratio of consultants to partners. For example,there are more than 30 consultants for each partnerat Andersen Consulting.

The reuse of knowledge saves work, reduces com-munications costs, and allows a company to take onmore projects. As a consequence, firms such as An-dersen Consulting and Ernst & Young have beenable to grow at rates of 20% or more in recent years.Ernst &. Young's worldwide consulting revenues, forexample, increased from $1.5 billion in 1995 to $2.7billion in 1997.

By contrast, tbe personalization strategy relies ontbe logic of "expert economics." Strategy consultingfirms offer tbeir clients advice that is rich in tacitknowledge. The process of sharing deep knowledgeis time consuming, expensive, and slow. It can't tru-ly be systematized, so it can't be made efficient.That means, first, that the ratio of consultants topartners in these firms is relatively low-there areapproximately seven consultants for each partner atMcKinsey and Bain. And second, it means that it'sdifficult to hire many new consultants in a short pe-riod because every new person needs so much one-on-one training. For tbose two reasons, strategy con-sulting firms find it difficult to grow rapidly withoutsacrificing the customized approach.

Nevertheless, their highly customized offeringsallow them to charge much higher prices thanfirms offering more standardized services can. In1997, for example, daily fees for a McKinsey consul-tant were on average more than $2,000; at AndersenConsulting, the figure was slightly more than $600.

Managing People. Not surprisingly, the two kindsof firms hire different kinds of people and train andreward them differently. Ernst & Young and Ander-sen Consulting hire undergraduates from top uni-versities and train them to develop and implementchange programs and information systems. Ander-sen's recruits are trained at the firm's Center forProfessional Education, a 150-acrc campus in St.Charles, Illinois. Using the knowledge manage-ment repository, the consultants work through sce-narios designed to improve business processes.They arc implemcnters, not inventors; the "not in-vented here" attitude has no place in a reuse firm.

McKinsey, BCG, and Bain hire top-tier M.B.A.graduates to be inventors - that is, to use their ana-lytic and creative skills on unique business prob-lems. These firms also want people who will beable to use the person-to-person knowledge-sharingapproach effectively. To be sure of obtaining peoplewith that mix of skills, they recruit with extraordi-nary care. Partners and senior consultants inter-view a candidate six to eight times before making ajob offer. At Bain, i out of 60 applicants gets an of-fer. Once on board, their most important trainingcomes from working with experienced consultantswho act as mentors.

From Health Care to High TechThe strategies of codification and personalization donot apply only to the world of consulting. We foundthat providers of health care and manufacturers ofcomputers also need to choose a knowledge man-agement approach that fits their needs and goals.



Access Health, a call-in medical center, exploitsa reuse model. When someone calls the center, aregistered nurse uses the company's "clinical deci-sion architecture" to assess the caller's symptoms,rule out possible conditions, and recommend ahome remedy, doctor's visit, or emergency roomtrip. The knowledge repository contains algorithmsof the symptoms of more than 500 illnesses. CEOJoseph Tallman describes the company's strategy:"We are not inventing a new way to cure disease.We are taking available knowledge and inventingprocesses to put it to better use."

Access Health provides a prime example of thebenefits that come from reusing codified knowl-edge-in this instance, software algorithms. Thecompany spent a lot to develop those algorithms,but it has been repaid handsomely for its invest-ment. The first 300 algorithms that Access Healthdeveloped have each been used an average of 8,000times per year. That level of reuse allows it to chargelow prices per call. In turn, the company's payingcustomers-insurance companies and providergroups-save money because many callers wouldhave made expensive trips to the emergency roomor doctor's office when they could have been diag-nosed over the phone.

Contrast Access Health's re-use strategy with the highly de-veloped personalization modelused at Memorial Sloan-Ketter-ing Cancer Center in New YorkCity. The center provides thebest, most customized adviceand treatment to cancer patients.A variety of experts consults oneach patient's case, and manag-ing the experts' collaborationis, in essence, managing thecenter's knowledge. Dr. JamesDougherty, its deputy physicianin chief, describes this collabora-tion as follows: "We coordinateintensive face-to-face communi-cation in order to ensure thatknowledge is transferred be-tween researchers and cliniciansand betw^een different types ofclinicians." Employees work to-gether in 17 disease-specificteams. The breast cancer team,for example, has 40 specialists-medical oncologists, surgeons,radiation therapists, psycholo-gists, and others- as well as acore of basic scientists.

To make person-to-person communication easy,a team's memhers are all located in the same areaof the hospital. Each team has several face-to-facemeetings per week that everyone attends. The meet-ings cover basic science initiatives, clinical findings,patient care, and ongoing research.

The center's human resource policy is alignedwith its knowledge management strategy. Top can-cer clinicians are attracted by Memorial Sioan-Ket-tering's state-of-the-art technology and excellentreputation. These clinicians are highly paid-mostreceive salaries that place them in the ninety-fifthpercentile or above relative to their counterparts atother academic institutions. The center hires clini-cians from two pools of candidates. Junior peopleare hired from top university residency programsand trained as fellows. The best fellows are movedinto an "up or out" pyramid system. The centeralso hires senior, nationally recognized clinicianswho often bring teams of people with them.

It is hard to imagine two business models in thesame industry as different as those of Access Healthand Memorial Sloan-Kettering. Yet both assess pa-tients' symptoms and make recommendations fortheir care, and both are highly successful. By pro-

Any company that depends on smart people and the flow of ideasmust choose a knowledge management strategy.



viding reliable service at low cost. Access Healthhas captured 50% of the call-center market and isgrowing at 40% a year. One insurer using its ser-vices saw its emergency-room admissions drop by15% and its physician office visits by 11%. For itspart. Memorial Sloan-Kettering is consistentlyranked as the top cancer research and treatment in-stitution in the country.

Medicine, like management consulting and otherservices, is built on unique knowledge. But thetwo knowledge management models also apply in

the indus-trial sector.

The strategy consulting consider thevery different

w e s tudied all approaches tak-, - , en by two com-

to gr ie i w i t h puter companies,, , , Dell and Hewlett-

document-dr iven PackardDell's competitive

strategy is to assemble in-expensive PCs that are

made to order and sell themdirectly to customers. A sophisticated knowledgemanagement system lies behind that businessmodel. Dell has invested heavily in an electronicrepository that contains a list of available compo-nents. The system drives the operation: customerschoose configurations from a menu, suppliers pro-vide components hased on their orders, and manu-facturing retrieves orders from the system andschedules assembly. Dell does not deliver highlycustomized orders, and it raises its prices consider-ably for orders with special components.

Dell has to invest a good deal up front to determineand specify configurations, but its investment paysoff because of the knowledge's reuse. In 1997, Dellsbipped 11 million PCs. Those systems were put to-gether from 40,000 possible configurations (competi-tors typically offer only about 100 configurations),which means that each configuration was used onaverage 275 times. That level of reuse allows Dell tolower its costs and charge less than the competition.Propelled in part hy its knowledge reuse model.Dell's net income for 1997 was $944 million on salesof $12.3 billion; the company's revenues have grown8 3 % annually over the last four years.

Hewlett-Packard, by contrast, uses a personaliza-tion approach to support its business strategy,which is to develop innovative products. For thatstrategy to succeed, technical knowledge must gettransferred to product development teams in atimely way. The company channels such knowl-edge through persoii-to-person exchanges.

For example, engineers routinely use one of thecompany's planes to visit other divisions and shareideas about possible new products. Rather thanlimiting travel budgets, executives encourage suchtravel. Every employee has access to the corporateairplanes, which travel daily between HP offices.Remarkably, the company manages effective per-son-to-person knowledge sharing despite its size-with 120,000 employees, HP dwarfs the largest con-sulting company, Andersen Consulting, which hasabout 60,000 people.

Consider this example. An HP team recently de-veloped a very successful electronic oscilloscopewith a Windows operating system and interface.Executives wanted to be sure that other divisionsunderstood and applied the interface. To keep thecosts of knowledge transfer low, they consideredtrying to codify the acquired know-how. They real-ized, however, that the knowledge they wanted tocapture was too rich and subtle to incorporate in awritten report. And they understood that writinganswers to the many questions that would comefrom HP's divisions would take an extraordinaryamount of time. So they took the person-to-personapproach and sent engineers from product develop-ment teams to meetings at divisions around theworld and to a companywide conference.

The executives' decision didn't come cheap: byone estimate, the company spent Si million on com-munication costs alone on this process. But the in-vestment paid off as the interface gained wide-spread acceptance throughout the company.

In all the companies and institutions we exam-ined, managers had chosen a distinct knowledgemanagement strategy. Although their approachesdiffered slightly, there was a common patternamong them. Those that pursued an assemblc-to-order product or service strategy emphasized thecodification and reuse of knowledge. Those thatpursued highly customized service offerings, or aproduct innovation strategy, invested mainly inperson-to-person knowledge sharing.

Do Not StraddleAs we've said, companies that use knowledge effec-tively pursue one strategy predominantly and usethe second strategy to support the first. We think ofthis as an 80-20 split: 80% of their knowledge shar-ing follows one strategy, 20% the other. Executiveswho try to excel at both strategics risk failing at both.Management consulting firms bave run into serioustrouble when they failed to stick with one approach.

The strategy consulting firms we studied all cameto grief with document-driven systems. Consultants



were tempted to use the systems to deUver standard-ized solutions, but their customers were paying forhighly customized services. When the systems weremisused, customers became dissatisfied.

As the CEO of a major U.S. company told us, "Ihave been using a particular consulting companyfor over a decade now. One of the main reasons Ihave used them so regularly is because tbey baveintimate knowledge of my company and our indus-try. The firm's partners who have worked with mealso know my style and my strengths and weak-nesses. The advice I have gotten from them has beensensitive to our unique needs. Recently, tbough, Ihave found that they are trying to push cookie-cutter solutions. It's almost as if they are simplychanging the names on the same set of presenta-tions. While some of their advice is useful, I am notsure if that's enough. Frankly I expect more-andthey sure as hell have not reduced their rates."

Another consulting firm. Bain, learned a hardlesson about relying on documents. In the 1980s,before electronic document systems became fash-ionable, managers at Bain developed a large paper-based document center at its Boston headquarters;it stored slide books containing disguised presenta-tions, analyses, and information on various indus-tries. The library's purpose was to help consultantslearn from work done in the pastwithout having to contact theteams that did the work. But asone partner commented, "Thecenter offered a picture of a cakewithout giving out the recipe."The documents could not conveythe richness of the knowledge orthe logic that had been applied toreach solutions-that understand-ing had to be communicated fromone person to another. Bain's man-agement eventually developed anentirely new system, but the failedapproach wasted time and money.

Other strategy consulting com-panies report different problemswitb electronic document sys-tems. For example, after subject ex-perts at one firm contributed docu-ments to electronic libraries, tbeywere fiooded with callers askingvery basic questions. Two compa-nies that we studied have scrappedtheir investment in electronicknowledge databases; their exist-ing databases are used simply toconnect people.

HARVARD BUSINESS REVIEW March-April 1999

Similarly, firms that rely on codification have runinto trouble by ovorinvcsting in person-to-personsystems. When they overinvest in this way, theyundermine their value proposition- reliable sys-tems at reasonable prices-as well as the economicsof reuse. That's because their people may feel en-couraged to develop a novel solution to a problemeven when a perfectly good solution already existsin the electronic repository. Unnecessary innova-tions are expensive: programming and then debug-ging new software, for instance, eats a lot of re-sources. And person-to-person knowledge sharinginvolves expensive travel and meeting time; thosecosts dilute the advantage that is created when cod-ified knowledge is reused.

Companies that straddle the two strategics mayalso find themselves with an unwieldy mix of peo-ple. Having both inventors and implementers rub-bing elbows can be deadly. The downfall of CSCIndex, the consulting company tbat invented thereengineering concept in the earlyi99os, under-scores how serious this problem can be.

The founders of what was originally known sim-ply as Index had strong backgrounds in IT systems.Success with reengineering, however, catapultedthe company into the general management arena. Itthen tried to leverage its newfound access to the

m GEGETTING THE INCENTIVES RIGHT

People need incentives to participate in the knowledge sharingprocess. The two knowledge manaKcment strategies call for dif-ferent incentive systems. In the codification model, managersneed to develop a system that encourages peopk to write downwhat they know and to get those documents int<i the electronicrepository. And real incentives-not small enticements-are re-quired to get people to take those steps. In fact, the level andquality of employees' contributions to the document datahaseshoutii be part of their annual performance reviews. Ernst &Young, for example, does just that. At performance reviews, con-sultants are evaluated along five dimensions, one of which istheir "contribution to and utilization of the knowledge asset ofthe firm."

Incentives to stimulate knowledge sharing should be very dif-ferent at companies that are following the personalization ap-proach. Managers need to reward people for sharing knowlctlgedirectly with other people, At Bain, the partners are evaluatedeach year on a variety of dimensions, including how much directhelp they have given colleagues. The degree of high-quality per-son-to-person dialogue a partner has had with others can accountfor as much as one-quarter of his or her annual compensation.

113


I HOW MUCH INFORMATION TECHNOLOGYI DO YOU NEED?

I ^The level of IT support a company needs depends on its choice of knowledgemanagement strategy. For the codification model, heavy IT support is criti-

; for the personalization model, it is much less important. Managers whoare implementing the former should be prepared to spend a lot on large, so-phisticated electronic repository systems. Andersen Consulting, for exam-ple, has developed proprietary search engines. Ernst &. Young has installed ahierarchy of databases, At the top arc "elite" databa.scs that are restricted insize and contain the best knowledge on a particular topic. Next come largerdatabases containing specific "knowledge objects"; finally there are themuch larger "holding tanks" for all kinds of other materials.

Over the past few years, Andersen Consulting and Ernst &. Young have eachspent more than $500 million on IT and people to support their knowledgemanagement strategics. On a much smaller seale. Access Health initially in-vested $16 million in its knowledge management system when its revenueswere a modest $20 million; later it spent another $40 million on the systemin order to have sufficient scale to generate $100 million in revenues.

The two knowledge management strategies require different IT infra-structures as well as different levels of support. In the codification model,managers need to implement a system that is much like a traditionallibrary-it must eontain a large cache of documents and include search en-gines that allow people to find and use the documents they need. In the per-sonalization model, it's most important to have a system that allows peopleto find other people.

CEO by aggressively hiring senior consultants fromestablished strategy consulting firms. It also startedto recruit M.B.A.s from leading business schools.Soon the firm had two populations: an old guardthat focused on IT systems and had strong imple-mentation skills, and a new guard that focused oncorporate strategy and had strong conceptual skills.

As reengineering became a commodity businesslater in the decade, some of the old guard recog-nized the need to standardize tbeir metbods andcreate more reusable knowledge. But members oftbe new guard bad little interest in working oncommodity-like reengineering projects. Tbey hadjoined tbe firm because tbey wanted to work oncutting-edge strategy problems.

As a result of this clasb, CSC Index was unable tokeep up witb competitors like Andersen Consult-ing and Ernst &. Young, wbicb leveraged a reusestrategy to deliver reengineering projects more reli-ably and at a lower price. Nor did tbe firm baveenougb deptb in strategy consulting to competewith tbe likes of MeKinsey, BCG, and Bain. In amarket tbat grew 10% annually from 1994 to 1996,

CSC Index's annual rev-enues slipped from $200million to an estimated$150 million. Tbe firmwas subsequently foldedinto its parent company.

Although it is impor-tant to avoid straddling,an exclusive focus on onestrategy is also unwise.Companies pursuing tbepersonalization modelsbould bave a modestelectronic documentsystem tbat supportspeople in two ways: byproviding backgroundmaterials on a topic andby pointing tbem to ex-perts wbo can providefurtber advice. As MarkHorwitcb, a partner atBain, explains, "Infor-mation in firms pursu-ing the person-to-personapproacb is leveraged asan input to tbe analyti-cal process ratber tbanas an output."

Companies tbat pri-marily adhere to tbereuse model will want

about 20% of tbeir knowledge sharing to be person-to-person. Tbus tbey will bave to pay to bring somepeople witbin tbe company togetber at meetings.Tbey sbould encourage tbe beavy use of e-mail andelectronic discussion forums. Sucb person-to-per-son communication is needed to make sure thatdocuments are not blindly applied to situations forwhicb tbey are ill suited.

Choosing the Right StrategyCompetitive strategy must drive knowledge man-agement strategy. Executives must be able to artic-ulate wby customers buy a company's products orservices ratber than those of its competitors. Wbatvalue do customers expect from tbe company? Howdoes knowledge that resides in the company addvalue for customers? If a company does not haveclear answers to tbose questions, it sbould not at-tempt to cboose a knowledge management strategybecause it could easily make a bad choice.

Assuming tbe competitive strategy is clear, man-agers will want to consider tbree furtber questions



that can belp tbem cboose a primary knowledgemanagement strategy. Altbough the implicationsof the answers may seem obvious, it is importantfor managers to make tbe explicit connection be-tween tbeir company's competitive strategy andhow they use knowledge to support it.

Do you offer standardized or customized prod-ucts? Companies that follow a standardized productstrategy sell products that do not vary much, if atall. Even Dell, whose assemble-to-order computersvary more tban mass-marketed products, sells prod-ucts tbat can be considered standardized. A knowl-edge management strategy based on reuse fits com-panies tbat are creating standardized products.

A company sells customized produets and ser-vices if most of its work goes toward meeting par-ticular eustomers' unique needs. Because tboseneeds will vary dramatically, codified knowledgeis of limited value. Companies tbat follow a cus-tomized product approach sbould consider the per-sonalization model.

Do you have a mature or innovative product? Abusiness strategy based on mature products typical-ly benefits most from a reuse model. Tbe processesfor developing and selling sueb products involvewell-understood tasks and knowledge that can becodified. A strategy based on product irmovation,on the other band, is best supported by a personal-ization strategy. People in companies seeking inno-vation need to sbare information that would getlost in document form.

Do your people rely on explicit or tacit knowl-edge to solve problems? Explicit knowledge isknowledge tbat can be codified, sucb as simple soft-ware code and market data. Wben a company's em-ployees rely on explicit knowledge to do their work,the people-to-documents approacb makes tbe mostsense. Tacit knowledge, by contrast, is difficult toarticulate in writing and is acquired througb per-sonal experience. It includes scientific expertise,operational know-bow, insights about an industry,business judgment, and technological expertise.Wben people use tacit knowledge most often tosolve problems, the person-to-person approachworks best.

Managers sometimes try to turn inherently tacitknowledge into explicit knowledge. Tbat can leadto serious problems. Xerox, for example, once at-tempted to embed tbe know-bow of its service andrepair tecbnicians into an expert system that wasinstalled in tbe copiers. Tbey hoped tbat techni-cians responding to a call could be guided by tbesystem and complete repairs from a distance. But itturned out that technicians could not solve prob-lems using the system by itself. Wben the copier

designers looked into tbe matter more closely, theydiscovered that technicians learned from one an-other by sbaring stories about bow tbey bad fixedthe machines. Tbe expert system could not repli-cate tbe nuance and detail tbat were exchanged infaee-to-faee conversations.

Your answers to tbe three questions above willoften suggest which knowledge management strat-egy to emphasize. But tbe issue is sometimes com-plicated by two additional concerns: the existenceof multiple business units and tbe commoditiza-tion of knowledge over time.

It is tempting to tbink tbat tbe two knowledgemanagement models can coexist in different busi-ness units witbin one corporation. Indeed, they eancoexist-but only in corporations where businessunits operate like stand-alone companies. In a com-pany like GeneralMotors, wbere the - . -icar divisions have L o m p a n i e s t h a tlittle to do with thecredit and finance

management mwork in each business •fnrir'f i r»nalunit. Companies with ^ H a ltightly integrated busi- d e p a r t m e n t sness units, however, rshould eitber focus on lilcp "F-TR o r ITonly one of tbe strategiesor spin off units tbat don't r i^V l o Q i n o ifQfit the mold. °

Some knowledge-inten- b e n e f i t s .sive produets and serviees-like reengineering con-sulting, for example-mature over time and becomecommodities. At first, the process of reengineeringrequired unique solutions, but it wasn't long beforea step-by-step approacb was needed. CSC Index be-gan witb tbe rigbt match-a personalization modelsupporting a customized offering-but tbat becamea mismatch as the concept of reengineeringchanged. Tbe firm bad a cboice: change its knowl-edge management strategy or get out of tbe reengi-neering business. By not choosing eitber, it fell ondifficult times.

In effective companies, tbe knowledge manage-ment model stays the same even as new productsand serviees mature. For eonsulting companies fo-cused on highly customized solutions, tbe trick isto get out of areas Hke reengineering before they be-come commodities. At firms that reuse knowledgeand solutions, tbe opposite is true: such firms ex-ploit an approacb as it matures. Peter Novins, apartner at Ernst & Young, puts it like tbis: "We try



to commoditize the expertise in one area as fast aspossible and move it to scale and reuse, which ben-efits both the chent and the company."

Don't Isolate Knowledge ManagementSome CEOs have put knowledge management atthe top of their agendas. Others have not given it thesame attention as they have given cost cutting, re-structuring, or international expansion. In compa-nies where that is the case, knowledge managementtakes plaee-if at all-in functional departmentssuch as HR or IT. But companies that isolate knowl-edge management risk losing its benefits, which arehighest when it is coordinated with HR, IT, andcompetitive strategy.

That coordination requires the leadership of thegeneral manager. When CEOs and general man-

agers actively choose a knowledge managementapproach-one that supports a clear competitivestrategy-both the company and its customers ben-efit. When top people fail to make such a choice,both suffer. Customers may end up paying for a cus-tomized solution when a standard solution wouldhave worked perfectly well. Or they may get paint-by-the-numbers advice when they really need helpwith a unique problem. Within the organization,employees will be confused about priorities. The is-sue will quickly become politicized, and peoplewill battle for resources without seeing the wholepicture. Only strong leadership can provide the di-rection a company needs to choose, implement,and overcome resistance to a new knowledge man-agement strategy. ^

Reprint 99206 Tu order reprints, see the last page of this issue.

'When I say 'now,' I mean this now, not the now we were talking about two minutes ago."


Harvard Business Review Notice of Use Restrictions, May 2009

Harvard Business Review and Harvard Business Publishing Newsletter content on EBSCOhost is licensed for

the private individual use of authorized EBSCOhost users. It is not intended for use as assigned course material

in academic institutions nor as corporate learning or training materials in businesses. Academic licensees may

not use this content in electronic reserves, electronic course packs, persistent linking from syllabi or by any

other means of incorporating the content into course resources. Business licensees may not host this content on

learning management systems or use persistent linking or other means to incorporate the content into learning

management systems. Harvard Business Publishing will be pleased to grant permission to make this content

available through such means. For rates and permission, contact [email protected].

what's your strategy for managing …s your strategy for managing knowledge? ... boston...

Documents