See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/228554838
Success and Failure Criteria for Knowledge Management Systems Article
CITATIONS
READS
2
877
3 authors: Mario Benassi
Paolo Bouquet
University of Milan
Università degli Studi di Trento
22 PUBLICATIONS 1,121 CITATIONS
122 PUBLICATIONS 2,594 CITATIONS
SEE PROFILE
SEE PROFILE
Roberta Cuel Università degli Studi di Trento 51 PUBLICATIONS 511 CITATIONS SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Okkam - Enabling the Web of Entities (FP7) View project
Academic spin-offs View project
All content following this page was uploaded by Paolo Bouquet on 26 May 2014. The user has requested enhancement of the downloaded file.
Success and Failure Criteria for Knowledge Management Systems
Mario Benassi1, Paolo Bouquet2, Roberta Cuel3 2 University of Milan,
Department of Information Sciences, via Comelico 39/41 – 20100 Milan, Italy
[email protected]
2 University of Trento,
Department of Information and Communication Technology, Via Sommarive, 14 -- 38050 Povo, Trento, Italy
[email protected]
3 University of Trento,
Department of Computer and Management Science, Via Inama, 5 -- 38100 Trento, Italy
[email protected]
Abstract
Knowledge management and more generally systems dealing with information creation, storage and retrieval are supposed to be a key factor for firms’ success. However, enthusiasm about their introduction is often paralleled
by disenchantment and delusion about their effectiveness and real use. Taking such failures seriously, we investigate reasons leading to success or failure of KM dissemination. Using a case study approach, we investigate the link between organizational profile and structure and information needs to be supported by KM systems, and offer a general framework to explain why –holding technology constant- the end result may differ widely. We conclude our investigation by suggesting that coherence and match are crucial ingredients to make KM work. Keywords. Knowledge Management, Distributed Knowledge Management, Knowledge Node, Technological architecture, Organizational model. Track: Strategy and management in knowledge-based organizations
1. Introduction In the last 20 years, knowledge has been considered as one of the most, if not the most, important organizational asset and KM has been considered as a discipline "realized in practice", which allows organizations to improve their capabilities capitalizing “organizational knowledge”. Organizational knowledge derives both from knowledge of single individuals and units that participate to the firm activities. Definitions about KM abound [Harris1], but despite existing differences KM can be considered as a set of practices allowing organizations to create, refine, store, and share knowledge, often through information communication technology (ICT). Increasing popularity of KM as a discipline is paralleled by the huge amount of investments made by firms and organizations on technological systems to transform capture, organize, store, and share workers’ knowledge [Zilich1].
Despite the claim, made by businessmen and consultants, that complex KM systems are a right answer to knowledge management users’ needs, these systems are often deserted by users. In other words, very often KM systems don’t reach the expected results, and are not effectively used by workers, who continue to produce and share knowledge in the way they did before the introduction of such systems. As explained in [Bonifacio2] the effectiveness of KM systems depends on different variables, which are both technological and organizational. It is not only the type of technology (for example obsolete databases, or communication protocols) of KM systems, but their technological architectures and a lot of other organizational variables – such as organizational model, social systems, individuals skills, organizational culture, and so on – that might influence their effectiveness. In particular we argue that KM systems partially fail because their representations of knowledge des not satisfy the needs and the interpretation schemas of users. Especially in complex organizations workers – specialized in different sectors, with different needs, different ways of thinking, and different interpretation schemas – cannot be forced to use a unique system of knowledge representation that they might consider it either as oppressive or irrelevant [Bowker1]. In our research we analyze two important variables which determine success or failure of KM systems: -
organizational model: it describes the way in which organization is structured in functions, and activities. From a KM point of view, it represents groups, teams, or communities that manage knowledge according to their local interpretation schema. In other words it reflects the system of interpretation schemas of individuals or groups of individuals (such as teams, communities, offices) that participate to the firm activities;
-
technological architecture: it reflects the way in which knowledge is represented within a KM system, namely how many representation schemas are allowed to be managed within the system.
In this paper, it is argued that success (or failure) of KM systems depends on the coherence (or incoherence) – in terms of centralization or distribution – between the technological architecture of KM systems and the organizational model of the firm. This claim is demonstrated by three case studies, three complex organizations in which (using the same technology - Lotus Notes) different KM systems have been implemented. Through these analysis we demonstrate that KM systems (with different technological architecture) are effective and successfully implemented when organizational models and technology architectures are coherent (both centralized, or both distributed); on the other hand KM systems fail, or partially fail, when technological architectures are centralized and organizational models are distributed, or vice versa, when technological architectures are distributed and organizational models are centralized.
2. Coherence
between
Organizational
Models
and
Technological
Architectures Different approaches are used to manage knowledge, and these have an influence on the technological architecture of KM systems implemented within firms. In the next paragraphs both organizational models and technological architecture are defined, focusing on centralization and distribution criteria.
2.1. Organizational Models In literature it is widely described that two dual approaches on KM have evolved and developed: the centralized approach (or northwestern approach) and the distributed approach (or subjectivist/Japanese approach) [Nonaka1]. According to the KM centralized approach, knowledge is considered reproducible within an organization only when it can be made objective, refining it by all individuals, environmental, and contextual conditions considered “raw” materials [Bonifacio2]. In other words, knowledge cannot be reused or shared in other organizational units, when it contains “details” which cannot be replicated. The “refining” processes of knowledge are done by individuals, who – wanting to create a “pure” and unique knowledge – eliminate some characteristics of knowledge in favor of others, supposed more useful. Then, a unique and apparently objective representation of knowledge is spread out within the firm, and all workers are forced to use it. From the KM point of view, we can say that organizational models are centralized when they allow the creation of a common and unique knowledge representation – called interpretation schema –. Centralized organizational models are very useful for organizational units (such as small firms, offices, teams, or communities) whose workers need to participate all together, sharing a common view and a unique way of thinking, namely a unique interpretation schema. Interpretation schema is continuously created, and modified through processes of participation and reification [Wenger1]. Often participation and reification are connected processes, and people – participating in a group – reify their interpretation schema adopting, using, creating and modifying artifacts (i.e. procedures, routines, workflows, archives, spatial disposition of the workplace, document categories, file managers, document managers, databases) that better suit their emerging needs.
Even if a unique and supposedly shared interpretation schema is developed within complex organizations, workers might not understand it, or may interpret it in different ways. A lot of authors from both cognitive (e.g. [Bouquet1, Ghidini1, Mccarthy1, Fauconnier1, Dinsmore1]) and social sciences (e.g. [Goffman1, Dougherty1]) studied the coexistence of different interpretational schemas, and the resulting orientation shows that interpretation schemas are only partially reducible to each others, and no unique, common, and shared schema can satisfy all users. Therefore it doesn’t exist a unique and pure version of knowledge, and in [Benerecetti1] it is called local knowledge: the different, partial, approximate, perspective interpretation of the world generated by individuals or groups of individuals (such as organizational units) which use their personal interpretation schema. In other words, knowledge cannot be made objective, unique, and understandable for more than one organizational unit, because it is strongly connected with individuals’ interpretation schemas. A distributed approach of KM is necessary (or the subjectivist/Japanese approach) in which each organizational unit can manage its knowledge according to its local interpretational schema, and can exchange knowledge with others without a unique and shared interpretation schema. Organizational models which sustain these processes and allow the creation of autonomous interpretation schemas are called distributed organizational models. These, are very useful, for example, in complex knowledge based organizations composed by different organizational units, which are specialized in different activities, with different needs, and therefore with different interpretation schemas [Bonifacio3]. In conclusion we can say that there are two different organizational models which sustain the two KM approaches:
-
centralized organizational model: knowledge is continuously negotiated and created within an organizational unit (perspective making [Boland1], single loop learning [Argyris1]),
-
the distributed organizational model: knowledge is continuously managed within organizational units, and it is continuously negotiated by people who try to understand how other units look like from different interpretation schemas (perspective taking [Boland1], double loop learning [Argyris1]).
2.2. Technological architectures Companies have invested huge amounts of money in order to manage knowledge through ICT. Even though KM systems use different technologies, tools, and methodologies, they seem to share some typical features. These, described by [Davenport1] aim at sustain informal communities, connect individuals through corporate-wide Intranet, allow mutual understanding using and designing corporate languages, and finally create an Enterprise Knowledge Portal (EKP), a simple and
EKP
KB
CONTRIBUTION THROUGH COLLABORATIVE TOOLS
Figure 1. Centralized architecture of KM systems
unique interface that allows individuals to create, and share corporate knowledge. Focusing the attention on semantic features and looking at traditional KM systems, we can figure out that technological architectures are usually composed (as it is shown in Figure1) by: -
collaborative environments (such as communities tools, communication protocols) which sustain informal communities (virtual or not) where "raw" knowledge is produced through spontaneous and emerging social interactions;
-
contribution workflows to refine, organize, and codify “raw” knowledge into a unique interpretation schema;
-
Knowledge Bases (KBs) to collect contents organized according to a common and unique corporate conceptual schema which represent a unique and shared interpretation schema;
-
Enterprise Knowledge Portal (EKP) to provide a single point of access for the members of different organizational units (different personalization processes can be developed on the same KB, namely each user can have different profiles to obtain customized information from the same KB).
The system described in Figure 1. is based on a centralized technological architecture. It tries to force, de-facto, a privileged knowledge representation onto people who may not share it. Then it is necessary to develop a new kind of system which supports the processes of the distributed organizational model: -
the autonomous management of local knowledge within organizational units;
-
the processes of mutual understanding among organizational units, namely how the world looks like from different interpretation schemas (Figure 2.).
Figure 2. Distributed architecture of KM systems
2.3. Organizational Models and Technological Architectures Coherences Despite Orlikowski in [Orlikowski1] and [Orlikowski2] says that workers have to adapt on technology systems, and technology systems must be changed to better suit the users’ needs, Bowker and Star say that a system has to respect the workers way of thinking [Bowker1]. It is evident that, looking at KM and KM systems, there is a strong correlation between organizational models of the firms and technological architecture of KM systems. In other words, technological architectures of KM systems must sustain the organizational models, allowing workers to manage their local interpretation schemas. We argue that KM systems will satisfy users when KM is allowed according to workers’ interpretation schemas, and will fail when technological architecture of KM systems is not able to manage the interpretation schemas of the workers. According to Figure 3. KM systems will be effective (the introduction of KM system will be successful) in these two cases:
Distributed
FAILURE KM System is considered oppressive
SUCCESS
SUCCESS
FAILURE KM System is considered irrelevant
Organizational Model
Centralized Centralized uted
Technological Architecture
Distrib-
Figure 3. KM systems’ architectures and Organizational models
-
both the technological architecture and the organizational model are centralized: the KM system manages a unique representation of knowledge which corresponds to the privileged interpretation schema, shared by workers, developed within the firm;
-
both the technological architecture and the organizational model are distributed: the technological architecture of KM system allows workers to manage more than one interpretation schema developed within autonomous organizational units.
Otherwise KM systems will fail when workers don’t use it, because it is considered either irrelevant or oppressive. These situation are determined by a gap between technological architecture of KM systems, and organizational model of the firm: -
the technological architecture is centralized and the organizational model is distributed: the centralized technological architecture force people, who work in autonomous organizational units managing personal interpretation schemas, to manage their knowledge using a privileged and unique interpretation schema;
-
the technological architecture is distributed and the organizational model is centralized: a huge number of interpretation schemas can be managed within the KM systems, but workers might use and understand only one of these: the privileged and unique interpretation schema of the firm.
In the next paragraph three case studies will demonstrate the failure and the success of KM systems according to the coherence or incoherence between organizational models, and technological architectures.
2. Case studies In this paragraph three case studies are described: three complex organizations which have developed KM systems using the same technology, Lotus Notes. Analyzing different KM systems developed within the firms, we focus the attention on problems of coherency between organizational models of the firms and technological architecture of the systems.
3.1. Andersen Consulting case study Andersen Consulting was one of the most important global consultant firm. Its organizational model was distributed: composed by autonomous units (teams and communities) which operated contemporary in three dimensions: 1. divisions: Assurance & Business Advisory, Tax, Legal & Business Advisory, Business Consulting, and Global Corporate Finance; 2. geographic communities: Pacific area, Europe, Asia, USA, Italy, France, Canada, and so on;
3. industries communities: Banking, Tourism, Industrial production, and so on. As described in [Bonifacio1, Bonifacio5] Andersen Consulting was a knowledge based organization in which knowledge was considered one of the most important organizational assets. Andersen Consulting tried to develop a lot of KM systems, and Lotus Notes was one of the first technology used. The first system introduced was able to organize local knowledge bases with personalized and autonomous interpretation schema. Despite workers couldn’t share documents and information with other groups – because of a common and shared system of classification (or interpretation schema) was missed –, the system had a great success. For each organizational unit, the technological architecture and the organizational model were centralized, therefore a unique representation of knowledge within the system corresponded to the interpretation schema of the local unit. In the ‘90s another KM system based on Lotus Notes was developed: Andersen On-line. It was a peer to peer communication system used by knowledge workers to communicate, and share knowledge with other organizational units. Workers were able to manage their local knowledge in the way they preferred – using personal interpretation schema –, and to exchange knowledge with other units. In short time the Andersen On-line system had a great success and then became crawly of participants. Then some problems figured out, it became difficult to find other workers with similar experiences and backgrounds, therefore for workers it was not easy explain – to people without similar background – their needs and point of views. To understand each others, workers started to share very general information, and no more useful information were exchanged. They were not able to solve their specific problems and to improve their personal knowledge. There-
fore a lot of individuals decided to not use Andersen On-line because they perceived it irrelevant, too much general to obtain useful information. At the beginning Andersen On-line had a great success because both the organizational model and the technological architecture were distributed. Then, it had a partial failure, because despite the organizational model was distributed, the technological architecture didn’t allow differentiation among different interpretation schemas.
3.2. Gesto case study GESTO is an Italian service providing company started in the late 1960’s [Benassi1, Benassi2, Greve1]. GESTO has a network structure composed by a steering group that evaluates projects, and a variable number of self-organizing groups which use artefacts, and internal KM systems which better suit their needs. Its organizational model is distributed, each group manages knowledge in an autonomous way, according to its needs, and its locally defined interpretation schema. Moreover the steering group has the function of knowledge coordinator across the entire organization, it makes knowledge sharable through the matching of different interpretation schemas. KM systems in GESTO are based on a common Lotus Notes database which contain detailed information about projects (for example title and main objectives, a brief description, time and length, whether the project was completed or suspended, whether it had an external client or was done for internal purposes, the names of the initiator, team leader, and participants) and a list of skills and skill levels of each worker. GESTO employees use Lotus Notes to organize projects, they access the database of ongoing projects with their participants to ?nd who is available at what time to be included in a project. Therefore KM system’s architec-
ture is distributed: each self-organizing group can create and manage its own personal and local interpretation schema. Knowledge sharing among these autonomous groups is guaranteed by knowledge coordination processes of the steering group, and by general rules that make life easier for everyone in the organization.
3.3. Impresa Pizzarotti & C. S.p.A. case study Pizzarotti & C. S.p.A. is one of the most important Italian building company. Its organizational model is distributed, it is composed by a centre – with different service offices such as administrative, EDP, security, quality, and human resources offices –, and a constellation of temporary building firms which operate in a geographically distributed environment. These firms are temporary organized, and workers join a new organizational unit when a new temporary firm is created, or a new building area is started. Each unit, in particular temporary firm, is completely autonomous, it creates and manages its interpretation schema in the way that better suits workers’ needs. KM system in Pizzarotti & C. S.p.A is based on Lotus Notes and is composed by a huge number of databases, which satisfy all the needs of the service offices. People of different offices can autonomously manage information, that is customized through profiling systems. All the organizational KBs are organized, stored, and managed according to a common and unique interpretation schema, therefore the technological architecture of Pizzarotti & C. S.p.A’s KM system is centralized. The firm decided to develop centralized architecture of KM systems because data are considered crucial for strategic issue, and any duplication, asynchronous, or not updated information is perceived as a problem to overcome.
Within Pizzarotti & C. S.p.A. KM systems are perceived by workers in different ways. Service offices consider KM systems as successful instruments, mainly because their interpretation schemas are represented by particular and complex views of the KM systems. Otherwise, autonomous temporary firms consider Lotus Notes applications as an oppressive system which workers are forced to use to exchange knowledge with the centre.
3. The DKM approach To solve the gap between organizational models and technological architecture we are developing a new approach on KM: the Distributed Knowledge Management (DKM) approach. As described in [Bonifacio1] the DKM approach takes into account two very general principles: the Principle of Autonomy – each organizational unit should be granted a high degree of autonomy to manage its local knowledge and its interpretation schema –, and the Principle of Coordination – each organizational unit must be enabled to exchange knowledge with others not through the adoption of a single, common interpretation schema –. Using this approach we can look at organizations and analyze the type of organizational model they have, to define the technological architecture of KM systems. The organizational model is seen as a constellation of local organizational units with their local interpretation schema. These organizational units are reified by the concept of Knowledge Node (KN) [Bonifacio4], which considers: -
the knowledge owner: one or more individuals (the organizational unit) who manages the local interpretation schema;
-
the system of artifacts: all the procedure, databases, and other tools which allow individuals to manage knowledge through participation and reification;
-
the context: its interpretation schema.
A methodology to define organizational units, KNs, and therefore organizational model (from the KM point of view) is under development as part of EDAMOK, a joint project of the Institute for Scientific and Technological Research (IRST, Trento) and of the University of Trento. Moreover a technological system of KM which adopts the DKM approach (described in [Bonifacio6]) is developed in order to sustain distributed organizational models. It allows to manage more than one interpretation schema, in particular a number of the interpretation schemas that correspond to the number of KNs.
4. Conclusions Three case studies are not enough to draw robust conclusions, but, paraphrasing a famous Simon’s motto, they are better than nothing. Andersen Consulting, Gesto and Pizzarotti & C. S.p.A. are different organizations dealing with different competitive industries, a different history, a different organizational profile, a different strategy, and so on. Although differences are huge, these three companies can be treated as equivalent, for they have dealt with issues normally addressed by KM systems. Rather than exhibiting a common, homogeneous behaviour, each show how technological systems may be interpreted and used in a peculiar way. Technology and KM systems are therefore an “open book”, to be filled, arranged and used by organizations. Such flexibility is a double-edge sword: it offers various degrees of
freedom, but at the same time it may be illusive, for KM can turn to an information trap. This is especially true if organizations do not pay attention to the interinterdependence between organizational profile and KM architecture. In this paper, we have offered several examples of how this issue may be dealt with, and offered a general framework to explain why the same application tool can, alternatively, succeed or fail. By using as main variables the nature of technological architecture and the organizational model, we offer a simple matrix that can be of some help for planning and managing the introduction of KM applications. We do believe that, although information and knowledge are by nature distributed, several alternative solutions are possible. This is a bad and a good news at the same time. It is a bad news, for we cannot infer from a KM characteristics how information and knowledge will be created and the organization really work. It is a good news, for it forces researchers and businessmen to dig more, hoping to find a “correct” formula to leverage both technology and organizations’ strengths.
References [Argyris1] C. Argyris. Teaching smart people how to learn. Harvard business review, 1999. [Benassi1] M. Benassi, A. Greve GESTO —A Network Company? Journal of Market-Focused Management.Vol.1(4):301-323,1996. [Benassi2] M. Benassi, A. Greve, J. Harkola Looking for a network organization. Journal of Market Focused Management, Vol.4(3):205-229,1999.
[Benerecetti1] M. Benerecetti, P. Bouquet, and C. Ghidini. Contextual Reasoning Distilled. Journal of Theoretical and Experimental Artificial Intelligence, 12(3):279–305, July 2000. [Boland1] R. J. Boland, and R. V. Tenkasi Perspective Making and Perspective Taking in Communities of Knowing, Organization Science, 6, 4 (July-August), 1995. [Bonifacio1] M. Bonifacio, P. Bouquet, and A. Manzardo. A distributed intelligence paradigm for knowledge management. In AAAI Spring Symposium Series 2000 on Bringing Knowledge to Business Processes. AAAI, 2000. [Bonifacio2] M. Bonifacio, P. Bouquet, and P. Traverso. Enabling distributed knowledge management. Managerial and technological implications. Informatik - Informatique, 1/2002. [Bonifacio3] M. Bonifacio, P. Bouquet, and R. Cuel. The Role of Classification(s) in Distributed Knowledge Management. Proceedings of 6th International Conference on Knowledge-Based Intelligent Information Engineering Systems & Allied Technologies (KES'2002), Special Session on Classification. 2002 Amsterdam, IOS Press. [Bonifacio4] M. Bonifacio, P. Bouquet, and R. Cuel. KNs: the building blocks of a distributed approach to KM. Journal for Universal Computer Science, vol.6, 2002. [Bonifacio5] M.Bonifacio, P.Bouquet, P.F.Camussone Knowledge Management: Teoria e prassi a confonto. Il caso Andersen. AIDEA 2002, Novara 2002. [Bonifacio6] M.Bonifacio, R. Cuel, G. Mameli, M. Nori. A Peer-to-Peer Architecture for Distributed Knowledge Management, Proceedings of 3rd International Symposium on Multi-Agent Systems, Large Complex Systems, and EBusinesses (MALCEB'2002)", Erfurt/Thuringia, Germany.
[Bouquet1] P. Bouquet. Contesti e ragionamento contestuale. Il ruolo del contesto in una teoria della rappresentazione della conoscenza. Pantograph, Genova (Italy), 1998. [Bowker1] G. C. Bowker and S. L. Star. Sorting things out: classification and its consequences. MIT Press., 1999. [Davenport1] T. H. Davenport, D. W. De Long, and M. C. Beers. Successful knowledge management projects. Sloan Management Review, 39(2), Winter 1998. [Dinsmore1] J. Dinsmore. Partitioned Representations. Kluwer Academic Publishers, 1991. [Doughuerty1] D. Dougherty. Interpretative barriers to successful product innovation in large firms. Organization Science, 3(2), 1992. [Fauconnier1] G. Fauconnier. Conceptual Integration Networks. Cognitive Science, 22(2):133-187, 1998. [Ghidini1] C. Ghidini and F. Giunchiglia. Local Models Semantics, or Contextual Reasoning = Locality + Compatibility. Artificial Intelligence, 127(2):221-259, April 2001. [Goffman1] I. Goffamn. Frame Analysis. Harper & Row, New York, 1974. [Greve1] A. Greve, M. Benassi, J. Harkola “Combining Expertise Raises Productivity: The Contributions Of Human And Social Capital To Performance” in M. Benassi (2002) Longitudinally Exploring Organizations, Cedam [Harris1] K. Harris and M. Fleming and R. Hunter and B. Rosser and A. Cushman, The Knowledge Management Scenario: Trends and Directions for 1998-2003, Gardner Group 1998.
[McCarthy1] J. McCarthy. Notes on Formalizing Context. Proc. of the 13th International Joint Conference on Artificial Intelligence, pages 555-560, Chambery, France, 1993. [Nonaka1] I. Nonaka, H. Takeuchi “The knowledge-creating company” Oxford University Press, Inc.,1995 [Orlikowski1] W.J. Orlikowski and D.C. Gash. Technological Frames: Making Sense of Information Technology in Organization. ACM Transactions on Information Systems, 1994 Vol.12/2, April, pages 174-207. [Orlikowski2] W. J. Orlikowsy, and D. Robey Information Technology and the Structuring of Organizations. Information Systems Research, 2, 2 (June), 1991. [Wenger1] E. Wenger. Communities of Practice. Learning, Meaning and Identity, Cambridge University Press, 1998. [Zilich1] R. Zilich Organic Knowledge Management e Dynamic Workplace per la gestione della conoscenza in IBM, Sistemi & Impresa, n. 6 luglio/agosto 2002.
View publication stats