On a "Common Language"for a "Best Practices" Daniel E. O’Leary University of Southern California 3660 Trousdale Parkway Los Angeles, CA90089-1421 Oleary@rcf.usc.edu From: AAAI Technical Report WS-99-13. Compilation copyright © 1999, AAAI (www.aaai.org). All rights reserved. Abstract firms, this is probably not a problemunless the dictator changes (e.g., through executive turnover) or there is attempt to implement the commonlanguages in some other setting (e.g., sell the best practices knowledge base). Knowledgemanagementsystems are becomingembedded in knowledge work. As part of those knowledge management systems, increasingly, firms are developing best practices knowledgebases that summarizea wide rangeof enterprise processes.Central to those particular knowledgebases are common languagesused to facilitate access and navigation through the knowledgebase. This paper summarizes someof the evidenceas to the necessity of common languagesin best practices databases. Further, somebarriers standing in the wayof development and use of these common languagesare summarized.In addition, this paper developsa modelfinding it is "impossible"to rationally choosea common languagethat meets needs of all individualsandthe firm, unlessdictatorshipis allowed. Scope Consulting firms, such as the "big six" have developed best practices knowledgebases for their owninternal direct use. Arthur Andersen and Price Waterhouse apparently were amongthe first such developers. Each of these firms have publicly available materials regarding their best practices knowledge bases (Arthur Andersen 1997, APQC 1997, International Benchmarking Clearinghouse 1997, Price Waterhouse 1995 and 1997). Asa result, the descriptive scopeof this paper is primarily limited to information available from those sources. 1. Introduction Increasingly, firms are developing"best practices" knowledgebases as part of their knowledgemanagement systems. Best practices (or leading practices) knowledge bases provide access to enterprise processes that appear to define the best waysof doing things. At the base of these best practices knowledgebases is what the developers (e.g., Price Waterhouse,1997) call a "common language" or (International BenchmarkingClearinghouse 1997) "commonvocabulary." 2. Best Practices KnowledgeBases Best practices knowledgebases capture information and knowledge about the best way to do things. Best practices knowledgebases have found use in a wide range of enterprises. For example, as noted by Davenport (1997), General Motors- HughesElectronics capture best process reengineering practices in a database. In addition, major consulting firms, including Arthur Andersen and Price Waterhouse, have developed best practices knowledgebases. The purpose of this paper is to summarize the use of commonlanguages in best practice knowledgebases and present an analytic model of the choice of common language concepts that are included in or excluded from the commonlanguage. This paper finds that based on a small set of assumptionsit is "impossible" to rationally choose a commonlanguage that will be optimal for individual membersof a group and the group itself, unless "dictatorship" is allowed. Thus, the resulting common languages are likely to meet the needs of the dictator or the needs that the dictator sees are important for the organization. Since the organizations that have developed best practices knowledgebases typically are for-profit Best Practice Firm Knowledge Bases as Models of the Best practices knowledge bases typically are based on process models of the firm, with emphasis on particular processes and howthose processes relate to each other. The best practice models are used to organize those processes for Arthur Andersenand Price Waterhousebest practices knowledge bases are in exhibits 1 and 2. 60 Exhibit1 -- ArthurAndersen ¯ Understandl I 2. Develop I I 3. Design I I 4. Market Markets& [---t Customers ] [ Vision Strategy& [-’t ] ] Products ServicesM& &Sell 5. Produce&l Deliverfor Manufaeturin i Organization I 7. Invoice&I Service Customers Deliverfor Service Organization 8. Develop and Manage HumanResources I 9. ManageInformation I 10. ManageFinancial and Physical Resources ] I I 1. Execute Environmental ManagementProgram I l 12. ManageExtrernal Relationships I [ 13. ManageImprovement and Change Global Best Practice I Classification Scheme (Arthur Andersen) Exhibit2 -- Price Waterhouse \ Perform Logistics and/,~ Customer Service Distribution/ Poo\lo Marketing and/,~ Products and/~ Products a Sales / Services / Services Perform Business Improvement ManaEeEnvironmental Concerns ManageExternal Re!at!onships Manage Corporate Services and Fa~ L ManaE~Financials Mana~ Human Resources Provide Lesal Services Perform Planning and Man*~ement Perform Procurement Develnn. ~¢zMaintain gwt~q18, andTot,h~,^l^~,_ ..... KnowledgeVlew Multi-Industry Process Technology (Price Waterhouse) Whatis in Best Practice KnowledgeBases? maybe referenceto articles or otherdescriptionsof the processes. Process measurements are also summarized providinga basis for benchmarking. Somebest practices knowledgebases include warstories, andinformation relating processesandtechnologyenablerinformation. Finally, the knowledgebase mayhave reference to Best practice knowledge bases include a range of materials. Typically they include text and or graphic representationof best practice processes. Best processes maybe generic or designedfor specific industries. There 61 particular experts on the processes. with other Knowledgebase Languages? What is the Relationship Between Common Languages and Knowledge ManagementOntologies? The information describing the process provides a basis for organization of the best processes. As a result, best practices are organized by process, performance measure benchmarks, industry-based process information and technology enabler. Are Best Practices or External Use? To what extent does a Best Practices KnowledgeBase Require a CommonLanguage? There is evidence that developers have found the need for a common language in the best practices databases to be a critical issue. For example, as noted by Price Waterhouse (1997) Knowledge Bases for Internal A CommonLanguage It is almost impossible to make intelligent comparisons without a commonset of reference point to describe the key processes and core capabilities of a business. Even within the same company,different divisions cannot compareprocesses whenthey lack a commonlanguage to describe what they do. The challenge becomes even greater when executives attempt to compare separate companies in the same industry or across different industries. Best practices must be accompanied by a commonlanguage that breaks business processes into activities that all companiesrecognize, understand and share. Initially best practice knowledgebases were designed for internal usage. Using best practice knowledge bases, consultants and auditors could have access to the best practices in order to help or understand their client’s processes. Clients benefited indirectly through having more knowledgeableauditors and consultants. However, recently firms apparently have become increasingly interested in direct access in order to facilitate "bench marking" with other firms and improving their work processes. As a result, some consulting firms have madetheir best practice knowledge bases directly available to users. For example, Arthur Anderson’s KnowledgeSpace (http://www.knowledgespace.corn) was made available as a service over the internet to subscribers in 1998. (In addition to access to best practices information, subscribers can access news, and other resources.) As a result, clients can nowdirectly access best practice information. How are CommonLanguages Developed? There is limited information available regarding how these different commonlanguages have been developed and how choice was made between alternative language representations. It is assumedthat different groups that will be using the commonlanguage are represented by an individual who represents their interests. Within that group choices are made regarding which concepts to include and exclude, what to call particular concepts, etc. The International Benehmarking Clearinghouse apparently madeheavy use of a single source of expertise in order to develop their common language. Best Practice Knowledge Bases are Part of a Portfolio of Knowledge Bases Typically, best practice knowledgebases are treated as a standalone knowledge base. However, best practice knowledge bases are only a part of the portfolio of knowledge management system knowledge bases. In consulting firms, those knowledge bases also include proposal knowledgebases, engagement knowledgebases, news, information and expertise databases and others. The Center and Arthur Andersen & Co. have collaborated closely to bring the Process Classification Frameworkto life and enhanceit over the past three years. The center would like to acknowledgethe staff of Arthur Andersen for their research and numerousinsights during this effort. (International BenchmarkingClearinghouse 1997) 3. Common Languagesin Best Practices KnowledgeBases Best practice knowledgebases typically are organized around a commonlanguage / taxonomy. The existence of a commonlanguage for best practices knowledge bases raises a numberof questions. ¯ To what extent does a Best Practices Knowledge Base Require a CommonLanguage? ¯ Howare CommonLanguages Developed? ¯ How many Best Practice CommonLanguages are there? ¯ Howdo Best Practice CommonLanguages Interface How many Best Practice there? CommonLanguages are It is unclear howmanydifferent best practice common languages have been developed. However, to-date there have been reports of a few large companies, such as General Motors (e.g., Davenport 1997), and large consultants (Arthur Andersen 1997 and Price Waterhouse 62 organization,knowledge navigation, facilitation of cross industry comparisons, need to eliminate "insider terminology,"and the broad base of constituencies, some of whichare discussedin moredetail here, whileall are discussedin detail in a longerversionof this paper. 1995and 1997)each developingbest practices databases and their owncorrespondingcommon languages. Which CommonLanguage is Optimal? Withthe existence of all of these common languages, whichis best? Thefact that different common languages are emergingprobably is evidencethat developers have different needsand as a result developdifferent common languagesto meet those needs. However,there are some apparent similarities. For example,two of the common languages that have gotten the most publicity (Price Waterhouse 1995 and International Benchmarking Clearinghouse/ArthurAndersen1997) appear to be based on a common overall model,Porter’s value chain model (Porter, 1980). Knowledge Navigation A commonlanguage facilitates knowledgenavigation througha best practices database(ArthurAndersen1997). Ourexperiencetaught us that the common organizing framework wasvery valuable -- it providesus with a common and understandablewayto navigate through the knowledge. Cross Industry Usage How do Best Practice Common Languages Interface with other Knowledgebase Languages? Oneof the benefitsof best practicesis to be able to take a processin one industry andadapt it to anotherindustry. Apparently, a common language can facilitate cross industry comparisons (Price Waterhouse 1997). As a stand alone knowledgebase, best practice common languagesgenerallydo not needto directly interface with other knowledgebase languages.However,this is likely to change as knowledgemanagementsystems become increasinglyintegrated. The challenge becomesevengreater whenexecutives attempt to compareseparate companiesin the same industryor acrossdifferent industries. Best practices must be accompaniedby a commonlanguage that breaks business processes into activities that all companiesrecognize, understandand share. What is the Relationship Between Common Languages and Knowledge Management Ontologies? Issues relating to common languagesin best practices knowledge basesare part of a larger set of issues referred to as a knowledgemanagement ontologies (KMOs). Eliminates Needfor "Insider Terminology" The International BenchmarkingClearinghouse (1997) argued that a commonlanguage allows them to break awayfrom specialized language. In particular, they indicatedthat they At the broadestlevel, an ontologyhas beendefinedas an explicit specificationof a conceptualization (e.g., Gruber, 1993). Within artificial intelligence, ontologies are necessary for multiple independentcomputingagents to communicate withoutambiguity.In addition, in artificial intelligence, ontologiesare the center of muchresearchon reusability of knowledgebases. A KMO is a knowledgebased specification that typically describes a taxonomy that defines the knowledge. Within the context of knowledge management systems, ontologies are specifications of discourse in the form of a shared vocabularyfor humanactors. Ontologiescan differ by developerandindustry, dependingon their humanusers. ... were convincedthat a common vocabulary, not tied to any specific industry was necessary to classify information by process and to help companiestranscend the limitations of insider terminology. Broad Rangeof Constituencies In addition, best practices knowledge bases are designed for a widerange of users for a widerange of uses. For example,Price Waterhouse (1995, p. 12) notes that "More than 30,000... professionals worldwidehave access to this tool for the purposeof collecting, refining, and sharingtheir knowledge with clients to help themenhance organizational competitiveness." As noted by Price Waterhouse, (1995, p. 9) 4. Why is it Necessary to have a Common Language? Witheachof the reports of best practice knowledge bases, there is also discussion of the uniquecommon language used to access and organize the knowledgebase. The need for a commonlanguage derives from a numberof factors including, knowledge reuse, knowledge Witha common language to describe the processes 63 and activities of all companies,any companycan compareitself to another. Price Waterhouse,The Knowledge Viewtaxonomy earl serve as the basis for best practice comparisons across industries, languages, and time zones. In fact the taxonomyis already being used on five continents and has been translated into several worldlanguages. TheFramework does not list all processeswithin any specific organization. Likewise, not every process listed in the Frameworkis present in every organization. Howdo developers decide what should be included or excluded? Howdo they choose betweendifferent terms to be includedin their best practices knowledge bases? 5. Barriers to a CommonLanguage in a Best Practices Knowledge Base Individual Differences Eachof these barriers suggestthat different individuals wouldhavedifferent preferences regarding the various terms and concepts included in or excluded from the common language. For example, someusers mayprefer informationcapturing industry-basedinformation,while others wouldprefer a "generic"view. Asa result, evenif there is a commonlanguage, there are likely to be divergentneeds. Theseindividual differences, and varyinglevels of expertise suggest that firms wouldemploya broad range of individuals in order to generate these common languages. Accordingly,a modelof howthat group can rationally makethe choices required for a common languagewouldbe a helpful andimportantcontribution. Althoughcommon languagesare seen as necessary, there are a number of barriers to their implementation. Common Languages are Costly to Develop Best practices knowledgebases are particularly complex anddifficult to develop ... weunderestimatedthe sheer effort necessaryto translate ... knowledgeabout best practice into useful explicit knowledge.The central team could not, on its ownextract ... knowledgeof the consultantsand the professionalsin the field .... After a significant effort, the teamhadproduceda CD-ROM with the classification scheme,but only 10 of the 170 processespopulated,andwith limited information. The initial offering almost died an early death--it seemedmucheffort for little payoff.(ArthurAndersen,1997,p. 4) 6. A Model of Choosing a Common Language This section discusses a model of group choice of a commonlanguage and vocabulary (Keeny and Raiffa 1993). Five rationality assumptionsare elicited and methodis sought to satisfy those five assumptionsin order to generate a rational approach to generate a commonlanguage. Generalvs. Specific Vocabulary A commonlanguage forces all users into the same vocabulary,so that users are unableto take advantageof vocabularies that meet specific needs. This is unfortunate, since specific technical vocabularies are typically developed because general language is insufficient (e.g., Kuhn1970). Notation Let S = (a,b,c, ...) be the set of alternativesavailablefor some concept in the framework. S is the set of alternatives fromwhichthe choicewill be made.Let I.I indicate eardinality. Let aRbindicate that onevocabulary term a is preferredto or indifferent than someother term b. R is used to stand for the preference relationships betweenall a and b. For example,R earl be specified througha utility function UwhereaRbif and only if U (a) > U (b). Suppose that there are n individuals responsible for deciding amongeach of the set of alternatives. Eachindividualwill be treated as oneof the participants in the set G, the groupmakingthe decision. Eachmemberin G maybe representatives fromdifferent subgroupswithin a company,as with Price Waterhouse (1995, 1997) or from multiple companies, as in the frameworkdevelopedby the International Benehmarking Clearinghouse(1997). Each participant is assumed Incorrect Usage of the Common Language Simplyhaving a common languagedoes not ensure that users knowhowto use that language. For example,as noted by Kuhn(1970p. 204) "Totranslate a theory into one’s ownlanguageis not to makeit one’s own." As a result, there is potential to misusethe common language resulting in comparisons that are not sensible. Whatis included and what is excluded? Commonlanguages make specific inclusion and exclusion of particular concepts. For example,as noted by the International Benchmarking Clearinghouse(1997) 64 have preferences between the choices madefor the common language. For example,consulting and auditing wouldlikely havedifferent preferences. Let P~ represent the preferencerelationshipof participant j andFj be the utility functionof participantj. Let aRabindicatethat the groupas a wholeprefers a over b. 1970).Further, only dictatorships satisfy assumptionsAD. Discussion The only group choice method that meets Arrow’s conditionsis "dictatorial choice." This wasa concernof Arrowbecause of equity issues. As noted by Arrow (1970,p. 86) Assumption A (Complete Domain) Assume that n > 2, [Sl ~ 3, andaRabexists for all possible aRjb. Thatis, there are at least twomembers in the group, they choosebetweenat least three alternative concepts and a group order can be specified for all possible individualorderings. ... the very act of establishinga dictator or elite to decideon the social goodmaylead to a distortion of the pragmatic from the moral imperative. "Power always corrupts; and absolute power corrupts absolutely"(LordAction). AssumptionB (Positive Association of Social and Individual Orderings) For decisions that influence intra firm common languagesfor best practices, there maybe no "equity" issue. Ideally, common languages wouldbe chosen to maximizethe overall company utility. That is not to say that those whoestablish common languagesalways will always makeoptimal choices. In addition, within companies,common languagechoices are likely to leave somedivisions or subgroupsas beneficiaries over other divisions and subgroups. However, as soon as the language is used outside of the organization of the dictator, beyondthe boundsof the dictator, there can be concern about the decisions that have been made. Choicesmadethat met the needs of the dictator, maynot meetthe needsof users in other organizations. If aRobfor a set of individualratings, andif (1) cRib, and c ~ a are not changed,(2) aRjband bRjaV b are not changedor are modifiedin b’s favor, then aRab. This says that if the groupprefers a to b andthe only changes that occur amongindividuals is with someadditional groupmembers preferringa, then a is preferredover b. Assumption C (Independence Alternatives) of Irrelevant Anytwoset of the rankingsmusthaveidentical corporate choicesin S or subsetsof S. In particular, a choicerule has independence of irrelevant alternatives if andonly if aR6b(for a,b ~ S’, VS’ c_ S) impliesaRab(for a,b ~ If one or morealternatives are removedfrom the set of choicesof concepts, then the choiceamongthe remaining alternativesis identical to the original orderingfor those alternatives. 7. Summary This paper has investigated commonlanguages and vocabulariesused in best practices knowledge bases, also referred to here as knowledgemanagement ontologies. The paper briefly discussed best practices knowledge bases and the importance of commonlanguages. Advantages and disadvantages of using common languages were investigated in order to illustrate divergent user needs. A model was developed to understand the group decision process associated with developinga common language. AssumptionD (Individual’s Sovereignty) For eacha,b, 3 someset of individual orderingsaRjbsuch that aRGb.Thatis, for eachpair of alternatives a andb, there is someset of individual orderings such that the groupprefers a to b. References AssumptionE (Nondietatorship) Arrow,K. 1970. Social Choice and Individual Values, YaleUniversityPress. There does not exist somek ~ G, such that aRobif and only if aRkb,Va,b~ S. Adictatorship wouldexist if one person (or division or companyrepresented by that person) determinedthe group’s common language. Arthur Andersen. 1997. "Knowledge, The Global Currencyof the 21st Century." Arrow’s Impossibility Theorem APQC - AmericanProductivity & Quality Center, 1997. "Arthur Andersen" There is no solution that meets assumptionsA-E. This result is knownas Arrow’simpossibility theorem(Arrow 65 Bernstein, L. 1974. Financial Statement Analysis, Irwin, Homewood, Illinois. Davenport, T.,"Some Principles of Knowledge Management,http://knowman.bus.utexas.edu/kmprin.htm Ernst & Young, 1997. Center for Business Knowledge, "Leading Practices KnowledgeBase. Gruber, T. 1993. "A Translational Approachto Portable Ontologies," KnowledgeAcq, 5, 2, pp.199-220. International Benchmarking Clearinghouse 1997. "Process Classification Framework." Keeny, R. and Raiffa, H. 1993. Decisions with Multiple Objectives, CambridgeUniversity Press Kuhn,T. 1970. The Structure of Scientific Revolution, SecondEdition, University of ChicagoPress. Porter, M. 1980. Competitive Strategy, Free Press, New York. Price Waterhouse. 1995. "Welcome to Knowledge View." Price Waterhouse. 1997. "International Business Language," formerly http://www.pw.com/kv/ibltxt.htm 66