- Email: [email protected]
DEMO: Towards a discipline of organisation engineering
European Journal of Operational Research 128 (2001) 351±363
www.elsevier.com/locate/dsw
DEMO: Towards a discipline of organisation engineering Jan L.G. Dietz
*
Delft University of Technology, Faculty of Mathematics and Informatics, P.O. Box 356, 2600 AJ Delft, The Netherlands
Abstract Dynamic Essential Modelling of Organisation (DEMO) is a theory about the ÔconstructionÕ and the ÔoperationÕ of organisations, that is rooted in the Communicative Action Paradigm regarding human communication and action. In this theory, the Ôworking principleÕ of an organisation consists of the entering into and the complying with commitments between human beings, where authority, responsibility and competence play an important role. The ÔconstructionÕ of an organisation consists of a coherent whole of transactions, i.e., recurrent patterns of communication and action in which the commitments are entered into and complied with. DEMO is also a methodology for modelling, (re)designing and (re)engineering organisations based on that theory. Up to now some thirty projects have been carried out with DEMO. These projects are of very dierent kinds and they have been carried out in various types of organisations. The success factor has been the same for all these projects, namely the practical relevance of the concepts of DEMO, as well as their clear and precise de®nitions. Both managers and employees appear to be able to participate readily in identifying and understanding problems, and in generating appropriate and feasible solutions. Ó 2001 Elsevier Science B.V. All rights reserved. Keywords: Modelling; Organisation; Communication; Business process; Information system
``Communication is the thread of which Organisation is woven''
1. Introduction We live in a time in which more and more organisations need to become innovative, competitive, and ¯exible enterprises, in order to survive.
*
Tel.: +15-278-4475. E-mail address: [email protected] (J.L.G. Dietz).
This does not only hold for commercial companies, but also for organisations with a public function. A prerequisite for being an innovative, competitive and ¯exible organisation, is that the business processes are ecient, easy to manage and easy to change, and that they are eectively supported by information systems. To enable this, all hope is placed in technology, particularly modern information and communication technology (ICT). However, this hope turns out to be vain if one does not dispose of the right expertise. Those who are involved in business processes and information systems, can roughly be divided into O-professionals and I-professionals.
0377-2217/01/$ - see front matter Ó 2001 Elsevier Science B.V. All rights reserved. PII: S 0 3 7 7 - 2 2 1 7 ( 0 0 ) 0 0 0 7 7 - 1
352
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363
O-professionals emanate from the organisational sciences, including the major part of the ®eld of information systems. They are primarily oriented towards function and behaviour and for that reason perfectly well able to specify the requirements business processes have to ful®ll in order to realise the business objectives. I-professionals emanate from the information processing sciences, including computer science and software engineering. They are primarily oriented towards construction and operation and for that reason perfectly well able to build information systems. There is a gap between O-professionals and I-professionals that cannot be bridged easily, as the history of the application of ICT in organisations has learnt. To bridge the gap one needs experts with a construction/operation oriented knowledge of business processes. Only those experts are able to design and engineer business processes in a well-founded and knowledgeable way. These experts could very appropriately be called Ôorganisation engineersÕ. There is worldwide a large and increasing shortage of them. This paper is about the construction and operation of organisations, as envisioned by a group of researchers at Delft University of Technology, participating in the research program Dynamic Essential Modelling of Organisations (DEMO). DEMO incorporates a way of thinking about organisation and technology that has originated from a deep dissatisfaction with current ways of thinking about information systems and business processes. These current ways of thinking fail to explain coherently and precisely how organisation and ICT are interrelated. They fail to provide assistance in articulating what is essential and invariant about the business processes and what are more or less incidental ways of doing. This is what seems to be needed: separating ÔessenceÕ from ÔtechnologyÕ. Fig. 1 shows the four core concepts of DEMO for understanding business processes and information systems. DEMO ®ts in a fairly new and promising perspective on business processes and information systems, called the Language/Action Perspective, or L/A Perspective for short. The theoretical foundation of this new perspective is constituted by Speech Acts Theory (Austin, 1962; Searle,
Fig. 1. The core concepts.
1969), and the Theory of Communicative Action (Habermas, 1981). The pioneer of the L/A Perspective is undoubtedly Fernando Flores (Flores and Ludlow, 1980; Winograd and Flores, 1986). Contrary to the prevailing notion that communication is exchanging sentences, expressing some proposition with regard to the world, the L/A Perspective assumes that communication is a kind of action in that it creates commitments between the communicating parties. To communicate then is to perform language acts (Searle, 1969) or communicative acts (Habermas, 1981), like requesting or promising. Other approaches in the same L/A Perspective can be found in e.g., Auramaki et al. (1988), Medina-Mora et al. (1992), Taylor (1993). Three international workshops have been held up to now focussing on the L/A Perspective (Dignum et al., 1996; Dignum and Dietz, 1997; Goldkuhl et al., 1998). These proceedings contain several papers concerning DEMO. Some other relevant papers are Dietz (1994), Van der Rijst and Van Reijswoud (1995), Dietz and Mulder (1996). The outline of the paper is as follows: Sections 2± 5 deal with the theory of DEMO. In Section 2, the concept of communication is discussed. This constitutes the ground on which the three other core concepts, information, action and organisation are founded. They are discussed in Sections 3±5, respectively. Section 6 deals with the DEMO methodology, and Section 7 provides some conclusions.
2. Communication In DEMO, communication is de®ned to be the sharing of mental states or thoughts between social
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363
individuals or subjects. The unit of communication consists of the sharing of one thought between two subjects, and is called the communicative act. Fig. 2 illustrates the process of performing a communicative act. The subject that is going to share one of her thoughts is called the locutor (L) of the act, and the subject with whom the thought is shared is called the addressee (A). The thought to be shared is formulated in some language that is common to L and A. The produced language expression is called (an elementary piece of) information. This information must somehow be made perceivable to A such that it can be interpreted by A. The eect of interpretation is the creation of a thought of A (of which both L and A usually hope that it is very similar to the thought of L). A thought is de®ned as a triple áI, F, Tñ, where I is the illocutionary kind, F is a fact, i.e., an elementary state of aairs, in some world, and T denotes the time period in which F is the case (or should be, depending on the kind of I). T has several default values, depending on I. As an example of a communicative act, let us assume that someone, the guest (G) for short, addresses a re-
Fig. 2. The communicative act.
353
ception employee (E) of a hotel and utters the next sentence: ``Do you have suites?'' The illocutionary kind I of the formulated thought is the question, the fact F is Ôthe hotel does have suitesÕ, and the time period T is a not clearly speci®ed (default) period, most probably Ônow and in the near futureÕ. In order to denote complete communicative acts, DEMO uses the so-called OER-notation áL:I:A:F:Tñ. The OER-notation of the example communicative act above is: áG: question: E: the hotel does have suites: now and in the near futureñ The reply by the hotel employee to this question could be: ``Yes, we do.'' is:
The OER-notation of this communicative act
áE: assertion: G: the hotel does have suites: now and in the near futureñ DEMO distinguishes six illocutionary kinds: question, assertion, request, promise, statement and acceptance. Fig. 3 exhibits the illocutionary kinds of DEMO, as well as how they are related to the categories as distinguished by Searle and Habermas. They fall into two illocutionary categories of Habermas, viz., constativa and regulativa. The classi®cation of Habermas diers from that of Searle because of the dierent philosophical stances they take regarding the interrelation-
Fig. 3. The illocutionary categories of DEMO.
354
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363
ships between communicating subjects, and between subjects and the world to which the facts belong. These stances are extensively discussed in Dietz and Widdershoven (1991) (Note: Contrary to the ®gure as presented there, HabermasÕ category of imperatives is left out in Fig. 3. This is in accordance with his later reconsideration that the illocutions in that category were already contained in the category of regulativa (Honneth and Joas, 1991).) It thus appears that the illocutionary kinds of DEMO can be explained fully using HabermasÕ Theory of Communicative Action, while SearleÕs Speech Act Theory fails to make the (for DEMO important) distinctions between questions and requests on the one hand, and between assertions and statements on the other hand. The ®gure also shows that DEMO focusses on business communication, to the exclusion of expressions of psychological or emotional states (expressiva and expressives). This exclusion does not mean that we neglect or underestimate the role of these acts in the functioning of organisations. Instead we believe that they constitute the indispensible Ôlubricant in the organisational machineryÕ. The only reason for not taking them into account, is that they are not directly related to the business at hand. The distinction between constativa and regulativa should be understood as follows: in every communicative act all three validity claims (truth, justice and sincerity) are present, however the dominant claim in constativa is the claim to truth, and the dominant claim in regulativa is the claim to justice. ÔJustÕ here means what is socially correct, i.e., what is valid given the actual social norms and values. As will be explained in Section 4, the six kinds include related illocutions like e.g., the denial as an illocution related to the assertion. A sequence of to and fro communicative acts between two subjects is called a conversation. DEMO distinguishes between informative and performative conversations. Informative conversations are conversations in which only questions and assertions occur. An example of an informative conversation between G and E is the combination of the two utterances, mentioned earlier: G: Do you have suites?
E: Yes, we do Performative conversations are conversations in which only requests, promises, statements and acceptances occur. Two subtypes are distinguished: actagenic and factagenic. An actagenic conversation is a conversation in which the request and the promise are the chief illocutionary kinds. An example in the hotel situation is: G: IÕd like to have a suite for 3 nights starting January the 3rd E: Let me see . . . yes, I can arrange that for you The OER-notation of this conversation (in which ÔasapÕ means Ôas soon as possibleÕ) is: áG: request: E: a suite is reserved for G from January 3 till January 6: asapñ áE: promise: G: a suite is reserved for G from January 3 till January 6: asapñ The result of this actagenic conversation is that E has committed him/her-self to make the agreed upon reservation. An example of a corresponding factagenic conversation is: E: Madame, I have reserved a suite for you for 3 nights starting January the 3rd G: Thank you very much The OER-notation of this conversation is: áE: statement: G: a suite is reserved for G from January 3 till January 6: asapñ áG: acceptance: E: a suite is reserved for G from January 3 till January 6: asapñ The example conversations show that the real meaning of a sentence can often not be deduced from a grammatical analysis of the sentence, because it depends heavily on the context (of other sentences) in which it is uttered. Furthermore, the examples show clearly that every sentence possesses the components I (illocution), F (fact) and T (time period). 3. Information In DEMO, information is understood to be form given thought, as was already mentioned in the previous section. The notion of information is
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363
355
closely related to the notion of communication. To be more speci®c: there is no information without communication, since information is produced only for the purpose of communicating (cf. Fig. 1). Information serves to bridge distances in both (physical) space and time. It allows therefore also to communicate with ÔoneselfÕ at dierent points in time. There are three aspects of information to be distinguished, called the forma, the in-forma and the per-forma, as shown in Fig. 4. Every piece of information has a forma, meaning that it has some perceivable structure carried in some physical substance. This structure must be recognisable as being an expression in some ÔlanguageÕ, which is the case if the forma conforms to the syntactical rules of that language. The notion of language has to be taken broadly. Every phenomenon that counts as a forma according to the institutional rules of a society, is by de®nition a forma (cf. Searle, 1995). To every forma belongs at least one in-forma (if there are more in-formaÕs, these are called homonyms). The in-forma is the meaning of the forma, the reference to some common world or Universe of Discourse, as de®ned by the semantics of the language. The aspect in-forma also includes the pragmatic rules of the language, like the choice of the right or best forma to express some in-forma in speci®c circumstances. The per-forma of a piece of information is the eect on the relationship between the communicating subjects, caused by communicating the thought. It is determined by both the illocution and the proposition of the communicative act, and is further dependent on the current norms and values in the shared culture of the communicating subjects. For example, if one asks someone the
time, one is supposed to need the answer (it is not considered just to ask such a question for fun), and the one who answers is expected to answer to the best of his knowledge. However, this ÔnormÕ may be overwritten, e.g., if the one has just robbed the other one. In such a situation it is acceptable to be dishonest. The distinction between the three aspects performa, in-forma, and forma, gives rise to the distinction of three corresponding levels at which information in an organisation has to be managed. These levels are called the essential level, the informational level, and the documental level, respectively (cf. Fig. 5). Likewise, these levels can be understood as ÔglassesÕ through which one can look at an organisation. Looking through the essential glasses, one observes socially acting elements, i.e., elements that execute per-formative actions like requesting or promising objective actions, and like stating or accepting results of objective actions. (Note: The notion of objective action will be explained in Section 4.) Looking through the informational glasses, one observes rationally acting elements, i.e., elements that execute in-formative actions like collecting, providing, recalling and computing knowledge about the common world or Universe of Discourse. Lastly, looking through the documental glasses, one observes formally acting elements, i.e., elements that execute formative actions like gathering, distributing, storing, copying, and destroying documents containing the aforementioned knowledge about per-formative actions and their results. Put in other terms, one observes through the three distinct glasses, systems of three distinct categories, which we call, in accordance with
Fig. 4. The aspects of information.
Fig. 5. The information management levels.
356
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363
Fig. 6. The system categories.
current practice, business processes, information systems and ICT-infrastructure components, respectively. These three system categories and their relationships are exhibited in Fig. 6. The double arrows from Business Process to Information System, and from the latter to ICT Infrastructure indicate one-to-many relationships, which should be understood as follows. Given a particular business process, one may conceive of a number of (collections of) information system(s) that support, and in doing so make operational, that business process. However at any point of time, there is only one (collection of) information system(s). Likewise, given a particular information system, one can conceive of a number of ICT infrastructure con®gurations that are able to realise, and in doing so make operational, that information system. At any point in time, however, there is one such con®guration. To conclude this section, we like to emphasize that the core or essence of a business process is, that the participants enter into and comply with commitments, and thus ®gure as social individuals. In order to do this in a responsible way, they may need to perform a lot of rational work like inquiring, calculating and evaluating alternative decisions (and probably also some formal work). This work however is only supportive to the business process, it is not part of it.
4. Action Action is a core notion in studying any dynamic system from an engineering point of view. According to the distinctions as proposed in the
previous section, DEMO distinguishes between essential actions, informational actions and documental actions. The main interest however is in the essential actions, because these are the Ôauthentic, genuineÕ business actions, all other actions only serve to support them. The class of essential actions is further divided into objective actions and social or intersubjective actions. By executing objective actions, the members of an organisation ful®ll the mission of the organisation. The nature of an objective action can be material or immaterial. Examples of material actions are all manufacturing actions in the production of goods as well as all storage and transportation actions. Examples of immaterial actions are the judgement by a court to condemn someone, the decision to grant an insurance claim, and appointing someone to be president. By executing intersubjective actions, subjects enter into and comply with commitments. In doing so, they initiate and coordinate the execution of objective actions. All intersubjective actions at the essential level of abstraction, fall into the category of regulativa (cf. Fig. 3). In order to abstract from the particular subject that performs an action and to concentrate on the functional or organisational role of the subject in performing that action, the notion of actor is introduced (e.g., sales and purchasing). An actor role can be ful®lled by a number of subjects (concurrently as well as collectively), and a subject may ful®ll concurrently a number of actor roles. In correspondance with the distinction between objective and intersubjective actions, DEMO distinguishes between two worlds in which each of these kinds of actions have eect: the object world and the intersubject world, respectively. The eect of every action is a state transition. These transitions are considered to take place instantaneously. A particular transition (e.g., requesting a beer by someone from the barkeeper) at a particular point in time (e.g., the requesting of a beer by a particular person at a particular moment) is called an event. Objective actions and their related intersubjective actions appear to occur in a particular pattern, called the (business) transaction, as illustrated by Fig. 7. It consists of three phases: the order phase
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363
357
Fig. 7. The business transaction.
or O-phase, the execution phase or E-phase, and the result phase or R-phase. (Note: The three letters O, E and R constitute the Dutch word OER which means primal, original, essential.) A transaction is carried through by two actors, who alternately perform actions. The one who starts the transaction and eventually completes it, is called the initiator (A1 in Fig. 7), the other one, who actually performs the objective action, is called the executor (A2 in Fig. 7). The order phase is an actagenic conversation, and the result phase is a factagenic conversation. Both conversations consist of communicative ( intersubjective) actions, having as eect a transition in the intersubject world (ISW). These actions are executed alternately by the initiator and the executor of the transaction. In between the two conversations, the objective action is executed, by the executor of the transaction. The eect of this action is a transition in the object world (OW). Because events in the object world are principally not knowable to the initiator (and to other actors) as long as they are not stated by the executor, transaction status 3 is coloured grey. This principal position is important. On the one hand, it stresses the supremacy of events in the intersubject world. On the other hand, it allows material and immaterial objective actions (and resulting facts) to be dealt with in the same manner. For immaterial facts it is obvious that they cannot be said to exist unless they are stated and subsequently accepted, and thus that they come into existence at the moment of acceptance, i.e., when reaching transaction status 5. Although at ®rst sight, and intuitively, material facts seem to come into existence in status 3, this appears not to be the case on
closer observation. In every organisation with material objective actions (like maufacturing or transporting ®rms), there appears always to be someone who is held responsible for a particular material fact being the case. Only he or she has the authority to declare that something is the case, which corresponds to an event in the intersubject world. Fig. 7 only shows the so-called success path of a transaction. The complete, and theoretically closed transaction process model is exhibited in Fig. 8. It has been demonstrated that this model is universal, i.e., all transactions within and between all organisations follow this pattern (Van Reijswoud, 1996). Transactions only dier in the kind of objective action they are dealing with, and in the way the intersubjective actions take place. Buying a loaf at the bakerÕs shop is an example of a very simple transaction type, where there are usually very few deviations from the success path, and where the intersubjective actions are simple utter-
Fig. 8. The transaction process model.
358
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363
ences or non-verbal acts. Moreover, they are often implicit according to the silent agreement Ôno news is good newsÕ. Buying a house is quite something else. Now, every transition in the intersubject world is performed explicitly and as unambiguously as possible. Moreover, it is documented carefully and extensively. The normal case is that a transaction follows the success path, as presented in Fig. 7. This path corresponds with the success layer in Fig. 8. The transaction then successfully ends in the terminal state 5. The discussion layer contains the possible Ôside pathesÕ a transaction may follow. It basically represents all kinds of negotiations and discussions the two actors may conduct for the sake of reaching consensus. Sometimes this consensus is not reached, meaning that the transaction ends unsuccessfully (terminal state 8). For example, instead of promising to establish the agreed upon proposition 1 ! 2, the executor can reject it 1 ! 6 for one or more of three reasons, re¯ected by the three validity claims (cf. Fig. 3). A rejection based on the claim to truth e.g., means that he is not able to establish the proposition (but perhaps a slightly dierent one). A rejection based on the claim to justice means that the executor challenges the initiatorÕs right to request him what was requested. A rejection based on the claim to sincerity means that the executor casts doubt on the sincerity of the initiator. In either case, the initiator has the option to respond in defense 6 ! 1 or to act otherwise, e.g., to stop the transaction 6 ! 8. All illocutions used in the discussion layer are regulativa, they are variants of the four kinds as presented in Fig. 3. The discourse layer represents the discussions in HabermasÕ theory about the social background, the norms and values, from which each of the actors act. The states 11 and 12 must be understood as compound states, i.e., as processes themselves. They include e.g., the recurrent policy discussions in an organisation like discussions about the acceptance rules to be applied to customer orders. The notions of action and event as presented above dier from those applied in most common approaches to information systems development and/or business process re-engineering. These approaches seem to adopt the mechanical notions of
cause and eect. Mechanical energy is always directly transferred between masses (cf. the collissions among snooker balls). Regarding mechanical processes, it is therefore perfectly alright to speak of trigger and eect. This is however not true for business processes. It is not at all harmless to say that (business) events trigger business processes. Apart from the imprecise use of the notion of event, it suggests a correspondence with mechanical cause and eect sequences. Actors however are not mechanical operators. For every action they take, they have a reason to take that particular action instead of an other one. These reasons or considerations are the basis for being held responsible for what actors do or refrain from. Even if at some time a particular rule is agreed upon, the actor applying it stays responsible for doing so, and is accountable for any eect that results from apparently irresponsible behaviour. 5. Organisation There exists a variety of de®nitions for the word ÔorganisationÕ. These de®nitions represent as many dierent notions, apparently depending on the point of view on takes towards the phenomenon organisation. As indicated in the introduction, DEMO takes an engineering point of view. This is rather new and alien for the O-professionals, since their involvement is dominated by notions of organisations of which the common denominator is that they take the black-box model of a system for granted. The black-box model is a very powerful mental tool for understanding organisations, however it is not the only one and, more importantly, it is inappropriate for understanding and changing the internal operation of a system. Fig. 9 exhibits the common graphical representation of the black-box model. Taking the black-box model when studying organisations (or any variant of it like the control model) one tries to ®nd and understand the relevant input variables and output variables, and the relationship between them, called the transfer function. Knowing the transfer function means knowing how the system responds to variations in the values of the input variables. By manipulating
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363
359
Fig. 9. The black-box model.
the input variables, one is able to change the behaviour of the system. If the transfer function is too complicated to understand, the technique of functional decomposition can be applied through which the system is replaced by a structure of subsystems of which the transfer functions are more readily understandable. One has to bear in mind however that the knowledge acquired about the system under investigation is still functional or behavioural knowledge, that it is insucient and inappropriate for changing the system. A functional decomposition of an organisation, whether it be performed in an exact way or more loosely, is very often taken to be a picture of how the organisation is ÔconstructedÕ and ÔworksÕ. However, a functional or black-box model of an organisation is only a particular conception of the organisation, a means to understand its (functional) behaviour. Like every other system (e.g., an alarm clock or a racing car), the functional behaviour of an organisation is brought about by the collective working of the constructional components. The construction and the working of a system is most near to what a system really is, to its ontological description. A very exact and very general, ontological de®nition of a system is provided by Bunge (1979). Based on this de®nition, DEMO uses the next de®nition of an organisation: Something is an organisation if and only if it ful®lls the next properties: · It has composition, i.e., it is composed of actors, where an actor is de®ned as one or more subjects in a particular role (cf. Section 4). These actors act on the basis of assigned authority and with corresponding responsibility. · It has structure, i.e., the actors in¯uence each other. Two kinds of mutual in¯uencing are distinguished. Interaction consists of executing transactions. Interstriction consists of taking into account the results or the status of other
transactions when carrying through a transaction. · It has boundary. The composition (i.e., the set of constituting actors) is divided into two subsets, called the kernel and the environment, such that every actor in the environment in¯uences, either through interaction or through interstriction, one or more actors in the kernel, and such that there are no ÔisolatedÕ parts in the kernel. The closed line that separates the kernel from the environment is called the boundary. On the basis of this de®nition, a white-box model of a system can be made, as exhibited in Fig. 10. The system elements of an organisation are the actors. On the one hand, they perform objective actions, thus changing the state of the OW. On the other hand, they perform intersubjective actions, thus changing the state of the intersubjective world (the system world in Fig. 10). When being active the actors take into account the state of the object world as well as the state of the ISW. Analoguous to the technique of functional (de)composition, there is a technique for composing and decomposing white-box models of a system. It is called constructional (de)composition. A central concept in the technique is the concept of subsystem. This concept is in DEMO de®ned as follows: S2 is a subsystem of S1 if and only if:
Fig. 10. The white-box model.
360
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363
· S1 and S2 are systems according to the de®nition above. · The kernel of S2 is a subset of the kernel of S1. · The structure of S2 is a subset of the structure of S1. · The environment of S2 is a subset of the composition of S1. In conclusion, functional (de)composition and constructional (de)composition are similar techniques, but applied to very dierent system notions. An important consequence is, that a functional component does not necessarily have a constructional counterpart. Attempting to relate the one kind of component to the other kind just does not make sense, their natures are inherently incompatible. To illustrate this, we take the decomposition of a car as an example. A functional decomposition might result into the distinction of the driving subsystem, the brake subsystem, the steering subsystem, the electrical subsystem etc. A constructional decomposition on the other hand would result into the chassis, the radiator, the wheels, the battery etc. It is obvious that each of the functional components does not correspond to some constructional component. More importantly, such correspondance is principally impossible. 6. The DEMO methodology The DEMO methodology was originally intended for carrying out the requirements engineering phase of information systems development, and it is being applied very successfully for that purpose indeed. Next to that however, it appeared to be very eective for a variety of other purposes, like Business Process Reengineering, Work Flow Management, Electronic Commerce and Virtual Organisations. In short, the methodology has proven to be a coherent, integral, and comprehensible approach to the modelling, the (re)design and (re)engineering, as well as the managing, of business processes and information systems. For several years now it is being applied in practice by a growing number of consultancy enterprises in all kinds of organisations: commercial, industrial and governmental.
The main contribution of the methodology in an organisationÕs change process, is the provision of the essential model of the organisation. This is a description by means of a number of diagramming techniques of its business processes (as de®ned by the DEMO theory), thereby abstracting fully from the informational/documental as well as from the organisational (structural) realisation. In each of the about thirty projects undertaken up to now, it has turned out that this essential model provides an ideal starting point for rethinking ®rstly the business of the organisation as such, and secondly the way in which it is realised now. Invariably, employees and managers are taken with the clarity, the conciseness, the relevance, and the eectiveness of the methodology and the produced results. Not only do DEMO projects cost a fraction of the costs of a ÔtraditionalÕ project, it also succeeds where the ÔtraditionalÕ methodologies keep failing. An extensive discussion of one of the smaller projects is provided in (Van Reijswoud et al., 1997). As an illustration of the DEMO methodology, let us look at the well-known Ford Case (Hammer, 1990). Hammer provides the next description of the business activities concerned: ``When Ford's purchasing department wrote a purchase order, it sent a copy to accounts payable. Later, when material control received the goods, it sent a copy of the receiving document to accounts payable. Meanwhile, the vendor sent an invoice to accounts payable. It was up to accounts payable, then, to match the purchase order against the receiving document and the invoice. If they matched, the department issued payment. The department spent most of its time on mismatches, instances where the purchase order, receiving document, and invoice disagreed...... One way to improve things might have been to help the accounts payable clerk investigate more eciently, but a better choice was to prevent the mismatch in the ®rst place. To this end, Ford instituted ``invoiceless processing''. Now when the purchasing department initiates an order, it enters the
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363
information into an on-line database. It doesn't send a copy of the purchase order to anyone. When the goods arrive at the receiving dock, the receiving clerk checks the database to see if they correspond to an outstanding purchase order. If so, he or she accepts them and enters the transaction into the computer system. (If receiving cannot ®nd a database entry for the received goods, it simply returns the order.)'' According to Hammer, Ford opted by the chosen solution for radical change, and achieved dramatic improvement. To illustrate this, it was mentioned that initially there were 500 people working at the accounts payable department, and that a 75% reduction of this ®gure was achieved after the solution had been implemented. Let us analyse this case brie¯y form the point of view of DEMO. To start with, Fig. 11 exhibits the relevant part of the essential model of the business activities concerned. On top of Fig. 11, a so-called transaction table is shown. It lists the distinct transaction types, including the speci®cation of the resulting fact types, as well as the speci®cation of the initiator and the executor of each transaction type. In the Ford Case, there are two transaction types: order delivery, and order payment. In the lower right corner of the ®gure the Interaction Model is shown, which can be derived directly from the transaction table. The lower left corner contains the Process Model. In this diagram each of the two transaction types is divided into its three constituent phases, the O-phase, the E-phase, and the R-phase. The solid arrows represent initiation
361
relationships. Their meaning is that the process or activity at the point side is started from the process or activity at the shaft side. For every transaction it holds that the E-phase is started from the Ophase, and that the R-phase is started form the Ephase. This is shown in Fig. 11 for the transaction types T1 and T2. Also shown is that the O-phase of transaction type T2 is started from the E-phase of transaction type T1, and that the O-phase of transaction type T1 is started externally (this is indicated by the small circle). The dotted arrow from T1/R to T2/E represents a conditional relationship. It means that T2/E can only be completed after T1/R has been completed. The Process Model expresses that the request for payment is issued in parallel with the statement that the order is delivered, and that the payment is actually performed after the order delivery has been completed. The interesting point in the Ford case is, that the essential model is the same before and after the radical(!) change. In the terminology of DEMO, this means that there has only been reengineering, not redesign. Redesign would imply that (part of) the essential model, e.g., the Process Model, would have been changed. In this respect, it sounds misleading to say that Ford has become invoiceless. This is only true at the documental level, meaning that there are no paper invoices sent anymore. The essential meaning of an invoice however is that it is a request for payment, and this request is still being performed. The dierence between the new situation and the old one, is that in the new situation the delivery of goods (also) counts as requesting for payment, while in the old situation the paper invoice conveyed this request.
Fig. 11. The essential model of the Ford Case.
362
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363
The essential model as presented in Fig. 11 (and including the transaction process models of each of the transaction types as exhibited in Fig. 8), ®rstly oers the right redesign freedom for the informational and the documental level, because it only shows the what instead of the how. It so triggers the mind for creating new solutions for the how, like the proposed change. Secondly, it oers strong support to the proposer of a change for assuring himself and others that the change is allowable, since it leaves the essential level unaffected. The sceptic reader would object, of course, that the change could also be found by applying common sense (as was done in the Ford Case). Common sense is however not always reliable, and is unevenly distributed among people. Moreover, it cannot be taught whereas a systematic approach like DEMO can. 7. Conclusions To bridge the gap between the O-people and the I-people requires more than the willingness to listen to each other. A new, engineering, perspective on business processes appears to be needed. The DEMO theory provides an appropriate framework for understanding business processes in this perspective. The essence of an organisation lies in the entering into and the complying with commitments by authorised and responsible subjects. This constitutes the working principle of any organisation. The DEMO transaction is the elementary building block of every business process, irrespective of the nature of the business, i.e., of the kind of the objective actions (material or immaterial). At the same time it becomes clear that a business process diers fundamentally from a production or a logistic process, and that so-called information intensive organisations (banks, insurance companies etc.) do have business processes like all other organisations, they only do not have production and logistic processes (at least at the essential level, they may have documental ones). Modelling business processes is a prerequisite for (re)designing and (re)engineering them. Most approaches to modelling business processes how-
ever either rely on black-box models or on ¯ow models, and therefore do not embody an appropriate understanding of the notion of business process; consequently they do not provide eective support. It has been illustrated that the DEMO methodology can be an eective help in various activities concerning the analysis and optimisation of business processes, thereby applying in an optimal way modern information and communication technology. Experiences with the methodology up to now sustain unequivocally the phrase that nothing is more practical than a good theory. In providing this theory, DEMO takes a step in the direction of the development of the badly needed discipline of Organisation Engineering.
References Auram aki, Lehtinen, E.E., Lyytinen, K., 1988. A speech act based oce modelling approach. ACM TOIS 6 (2). Austin, J.L., 1962. How to Do Things with Words. Harvard University Press, Cambridge, MA. Bunge, M.A., 1979. Treatise on Basic Philosophy, vol. 4. Reidel, Dordrecht, Netherlands. Dietz, J.L.G., 1994. Modelling business processes for the purpose of redesign. In: Proceedings of the IFIP TC8 Open Conference on BPR. North-Holland, Amsterdam. Dietz, J.L.G., Mulder, J.B.F., 1996. Realising strategic reengineering objectives with DEMO. In: Proceedings of the International Symposium on Business Process Modelling. Springer, Berlin. Dietz, J.L.G., Widdershoven, G.A.M., 1991. Speech acts or communicative action? In: Proceedings of the Second European Conference on CSCW. Kluwer Academic Publishers, Boston, MA. Dignum, F., Dietz, J.L.G., Verharen, E., Weigand, H., 1996. Communication Modeling-the Language/Action Perspective. Electronic Workshops in Computing. Springer, Berlin (available at http://www.springer.co.nk/ewic/workshops/ CM96/). Dignum, F., Dietz, J.L.G., 1997. In: Proceedings of the Second International Workshop on Communication Modelling. Computing Science Reports. Eindhoven University of Technology. Flores, F.M., Ludlow, J.J., 1980. Doing and speaking in the oce. In: Decision Support Systems: Issues and Challenges. Pergamon, Oxford. Goldkuhl, G., Lind, M., Seigerroth, U., 1998. The language action perspective on communication modelling. In: Proceedings of the Third International Workshop, Department of Informatics, J onk oping International Business School, Sweden.
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363 Habermas, J., 1981. Theorie des Kommunikatives Handelns. Erster Band. Suhrkamp, Frankfurt am Main. Hammer, M., 1990. Reengineering Work: DonÕt Automate, Obliterate. Harvard Business Review. Honneth, A., Joas, H. (Eds.), 1991. Communicative Action ± Essays on J urgen HabermasÕ ÔThe theory of Communicative ActionÕ. Basil Blackwell, Oxford. Medina-Mora, R., Winograd, T., Flores, R., Flores, F., 1992. The action work¯ow approach to work¯ow management technology. In: Proceedings of the Fourth International Conference on CSCW, ACM, New York. Searle, J.R., 1969. Speech Acts, an Essay in the Philosophy of Language. Cambridge University Press, Cambridge, MA. Searle, J.R., 1995. The Construction of Social Reality. Allen Lane, Penguin, London. Taylor, J.R., 1993. Rethinking the Theory of Organisational Communication: How to Read An Organisation. Ablex, Norwood, NJ.
363
Van Reijswoud, V.E., 1996. The Structure of Business Communication ± Theory, Model and Application. Doctoral Thesis. Delft University of Technology. Van der Rijst, N.B.J., Van Reijswoud, V.E., 1995. Comparing speech act based modeling approaches for the purpose of information systems development. In: Proceedings of the Third European Conference on Information Systems, Athens, Greece. Van Reijswoud, V.E., Mulder, J.B.F., Dietz, J.L.G., 1999. Communicative action-based business process and information systems modelling with DEMO. Information Systems Journal 9, 117±138. Winograd, T., Flores, F., 1986. Understanding Computers and Cognition: A New Foundation for Design. Ablex, Norwood, NJ.
www.elsevier.com/locate/dsw
DEMO: Towards a discipline of organisation engineering Jan L.G. Dietz
*
Delft University of Technology, Faculty of Mathematics and Informatics, P.O. Box 356, 2600 AJ Delft, The Netherlands
Abstract Dynamic Essential Modelling of Organisation (DEMO) is a theory about the ÔconstructionÕ and the ÔoperationÕ of organisations, that is rooted in the Communicative Action Paradigm regarding human communication and action. In this theory, the Ôworking principleÕ of an organisation consists of the entering into and the complying with commitments between human beings, where authority, responsibility and competence play an important role. The ÔconstructionÕ of an organisation consists of a coherent whole of transactions, i.e., recurrent patterns of communication and action in which the commitments are entered into and complied with. DEMO is also a methodology for modelling, (re)designing and (re)engineering organisations based on that theory. Up to now some thirty projects have been carried out with DEMO. These projects are of very dierent kinds and they have been carried out in various types of organisations. The success factor has been the same for all these projects, namely the practical relevance of the concepts of DEMO, as well as their clear and precise de®nitions. Both managers and employees appear to be able to participate readily in identifying and understanding problems, and in generating appropriate and feasible solutions. Ó 2001 Elsevier Science B.V. All rights reserved. Keywords: Modelling; Organisation; Communication; Business process; Information system
``Communication is the thread of which Organisation is woven''
1. Introduction We live in a time in which more and more organisations need to become innovative, competitive, and ¯exible enterprises, in order to survive.
*
Tel.: +15-278-4475. E-mail address: [email protected] (J.L.G. Dietz).
This does not only hold for commercial companies, but also for organisations with a public function. A prerequisite for being an innovative, competitive and ¯exible organisation, is that the business processes are ecient, easy to manage and easy to change, and that they are eectively supported by information systems. To enable this, all hope is placed in technology, particularly modern information and communication technology (ICT). However, this hope turns out to be vain if one does not dispose of the right expertise. Those who are involved in business processes and information systems, can roughly be divided into O-professionals and I-professionals.
0377-2217/01/$ - see front matter Ó 2001 Elsevier Science B.V. All rights reserved. PII: S 0 3 7 7 - 2 2 1 7 ( 0 0 ) 0 0 0 7 7 - 1
352
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363
O-professionals emanate from the organisational sciences, including the major part of the ®eld of information systems. They are primarily oriented towards function and behaviour and for that reason perfectly well able to specify the requirements business processes have to ful®ll in order to realise the business objectives. I-professionals emanate from the information processing sciences, including computer science and software engineering. They are primarily oriented towards construction and operation and for that reason perfectly well able to build information systems. There is a gap between O-professionals and I-professionals that cannot be bridged easily, as the history of the application of ICT in organisations has learnt. To bridge the gap one needs experts with a construction/operation oriented knowledge of business processes. Only those experts are able to design and engineer business processes in a well-founded and knowledgeable way. These experts could very appropriately be called Ôorganisation engineersÕ. There is worldwide a large and increasing shortage of them. This paper is about the construction and operation of organisations, as envisioned by a group of researchers at Delft University of Technology, participating in the research program Dynamic Essential Modelling of Organisations (DEMO). DEMO incorporates a way of thinking about organisation and technology that has originated from a deep dissatisfaction with current ways of thinking about information systems and business processes. These current ways of thinking fail to explain coherently and precisely how organisation and ICT are interrelated. They fail to provide assistance in articulating what is essential and invariant about the business processes and what are more or less incidental ways of doing. This is what seems to be needed: separating ÔessenceÕ from ÔtechnologyÕ. Fig. 1 shows the four core concepts of DEMO for understanding business processes and information systems. DEMO ®ts in a fairly new and promising perspective on business processes and information systems, called the Language/Action Perspective, or L/A Perspective for short. The theoretical foundation of this new perspective is constituted by Speech Acts Theory (Austin, 1962; Searle,
Fig. 1. The core concepts.
1969), and the Theory of Communicative Action (Habermas, 1981). The pioneer of the L/A Perspective is undoubtedly Fernando Flores (Flores and Ludlow, 1980; Winograd and Flores, 1986). Contrary to the prevailing notion that communication is exchanging sentences, expressing some proposition with regard to the world, the L/A Perspective assumes that communication is a kind of action in that it creates commitments between the communicating parties. To communicate then is to perform language acts (Searle, 1969) or communicative acts (Habermas, 1981), like requesting or promising. Other approaches in the same L/A Perspective can be found in e.g., Auramaki et al. (1988), Medina-Mora et al. (1992), Taylor (1993). Three international workshops have been held up to now focussing on the L/A Perspective (Dignum et al., 1996; Dignum and Dietz, 1997; Goldkuhl et al., 1998). These proceedings contain several papers concerning DEMO. Some other relevant papers are Dietz (1994), Van der Rijst and Van Reijswoud (1995), Dietz and Mulder (1996). The outline of the paper is as follows: Sections 2± 5 deal with the theory of DEMO. In Section 2, the concept of communication is discussed. This constitutes the ground on which the three other core concepts, information, action and organisation are founded. They are discussed in Sections 3±5, respectively. Section 6 deals with the DEMO methodology, and Section 7 provides some conclusions.
2. Communication In DEMO, communication is de®ned to be the sharing of mental states or thoughts between social
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363
individuals or subjects. The unit of communication consists of the sharing of one thought between two subjects, and is called the communicative act. Fig. 2 illustrates the process of performing a communicative act. The subject that is going to share one of her thoughts is called the locutor (L) of the act, and the subject with whom the thought is shared is called the addressee (A). The thought to be shared is formulated in some language that is common to L and A. The produced language expression is called (an elementary piece of) information. This information must somehow be made perceivable to A such that it can be interpreted by A. The eect of interpretation is the creation of a thought of A (of which both L and A usually hope that it is very similar to the thought of L). A thought is de®ned as a triple áI, F, Tñ, where I is the illocutionary kind, F is a fact, i.e., an elementary state of aairs, in some world, and T denotes the time period in which F is the case (or should be, depending on the kind of I). T has several default values, depending on I. As an example of a communicative act, let us assume that someone, the guest (G) for short, addresses a re-
Fig. 2. The communicative act.
353
ception employee (E) of a hotel and utters the next sentence: ``Do you have suites?'' The illocutionary kind I of the formulated thought is the question, the fact F is Ôthe hotel does have suitesÕ, and the time period T is a not clearly speci®ed (default) period, most probably Ônow and in the near futureÕ. In order to denote complete communicative acts, DEMO uses the so-called OER-notation áL:I:A:F:Tñ. The OER-notation of the example communicative act above is: áG: question: E: the hotel does have suites: now and in the near futureñ The reply by the hotel employee to this question could be: ``Yes, we do.'' is:
The OER-notation of this communicative act
áE: assertion: G: the hotel does have suites: now and in the near futureñ DEMO distinguishes six illocutionary kinds: question, assertion, request, promise, statement and acceptance. Fig. 3 exhibits the illocutionary kinds of DEMO, as well as how they are related to the categories as distinguished by Searle and Habermas. They fall into two illocutionary categories of Habermas, viz., constativa and regulativa. The classi®cation of Habermas diers from that of Searle because of the dierent philosophical stances they take regarding the interrelation-
Fig. 3. The illocutionary categories of DEMO.
354
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363
ships between communicating subjects, and between subjects and the world to which the facts belong. These stances are extensively discussed in Dietz and Widdershoven (1991) (Note: Contrary to the ®gure as presented there, HabermasÕ category of imperatives is left out in Fig. 3. This is in accordance with his later reconsideration that the illocutions in that category were already contained in the category of regulativa (Honneth and Joas, 1991).) It thus appears that the illocutionary kinds of DEMO can be explained fully using HabermasÕ Theory of Communicative Action, while SearleÕs Speech Act Theory fails to make the (for DEMO important) distinctions between questions and requests on the one hand, and between assertions and statements on the other hand. The ®gure also shows that DEMO focusses on business communication, to the exclusion of expressions of psychological or emotional states (expressiva and expressives). This exclusion does not mean that we neglect or underestimate the role of these acts in the functioning of organisations. Instead we believe that they constitute the indispensible Ôlubricant in the organisational machineryÕ. The only reason for not taking them into account, is that they are not directly related to the business at hand. The distinction between constativa and regulativa should be understood as follows: in every communicative act all three validity claims (truth, justice and sincerity) are present, however the dominant claim in constativa is the claim to truth, and the dominant claim in regulativa is the claim to justice. ÔJustÕ here means what is socially correct, i.e., what is valid given the actual social norms and values. As will be explained in Section 4, the six kinds include related illocutions like e.g., the denial as an illocution related to the assertion. A sequence of to and fro communicative acts between two subjects is called a conversation. DEMO distinguishes between informative and performative conversations. Informative conversations are conversations in which only questions and assertions occur. An example of an informative conversation between G and E is the combination of the two utterances, mentioned earlier: G: Do you have suites?
E: Yes, we do Performative conversations are conversations in which only requests, promises, statements and acceptances occur. Two subtypes are distinguished: actagenic and factagenic. An actagenic conversation is a conversation in which the request and the promise are the chief illocutionary kinds. An example in the hotel situation is: G: IÕd like to have a suite for 3 nights starting January the 3rd E: Let me see . . . yes, I can arrange that for you The OER-notation of this conversation (in which ÔasapÕ means Ôas soon as possibleÕ) is: áG: request: E: a suite is reserved for G from January 3 till January 6: asapñ áE: promise: G: a suite is reserved for G from January 3 till January 6: asapñ The result of this actagenic conversation is that E has committed him/her-self to make the agreed upon reservation. An example of a corresponding factagenic conversation is: E: Madame, I have reserved a suite for you for 3 nights starting January the 3rd G: Thank you very much The OER-notation of this conversation is: áE: statement: G: a suite is reserved for G from January 3 till January 6: asapñ áG: acceptance: E: a suite is reserved for G from January 3 till January 6: asapñ The example conversations show that the real meaning of a sentence can often not be deduced from a grammatical analysis of the sentence, because it depends heavily on the context (of other sentences) in which it is uttered. Furthermore, the examples show clearly that every sentence possesses the components I (illocution), F (fact) and T (time period). 3. Information In DEMO, information is understood to be form given thought, as was already mentioned in the previous section. The notion of information is
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363
355
closely related to the notion of communication. To be more speci®c: there is no information without communication, since information is produced only for the purpose of communicating (cf. Fig. 1). Information serves to bridge distances in both (physical) space and time. It allows therefore also to communicate with ÔoneselfÕ at dierent points in time. There are three aspects of information to be distinguished, called the forma, the in-forma and the per-forma, as shown in Fig. 4. Every piece of information has a forma, meaning that it has some perceivable structure carried in some physical substance. This structure must be recognisable as being an expression in some ÔlanguageÕ, which is the case if the forma conforms to the syntactical rules of that language. The notion of language has to be taken broadly. Every phenomenon that counts as a forma according to the institutional rules of a society, is by de®nition a forma (cf. Searle, 1995). To every forma belongs at least one in-forma (if there are more in-formaÕs, these are called homonyms). The in-forma is the meaning of the forma, the reference to some common world or Universe of Discourse, as de®ned by the semantics of the language. The aspect in-forma also includes the pragmatic rules of the language, like the choice of the right or best forma to express some in-forma in speci®c circumstances. The per-forma of a piece of information is the eect on the relationship between the communicating subjects, caused by communicating the thought. It is determined by both the illocution and the proposition of the communicative act, and is further dependent on the current norms and values in the shared culture of the communicating subjects. For example, if one asks someone the
time, one is supposed to need the answer (it is not considered just to ask such a question for fun), and the one who answers is expected to answer to the best of his knowledge. However, this ÔnormÕ may be overwritten, e.g., if the one has just robbed the other one. In such a situation it is acceptable to be dishonest. The distinction between the three aspects performa, in-forma, and forma, gives rise to the distinction of three corresponding levels at which information in an organisation has to be managed. These levels are called the essential level, the informational level, and the documental level, respectively (cf. Fig. 5). Likewise, these levels can be understood as ÔglassesÕ through which one can look at an organisation. Looking through the essential glasses, one observes socially acting elements, i.e., elements that execute per-formative actions like requesting or promising objective actions, and like stating or accepting results of objective actions. (Note: The notion of objective action will be explained in Section 4.) Looking through the informational glasses, one observes rationally acting elements, i.e., elements that execute in-formative actions like collecting, providing, recalling and computing knowledge about the common world or Universe of Discourse. Lastly, looking through the documental glasses, one observes formally acting elements, i.e., elements that execute formative actions like gathering, distributing, storing, copying, and destroying documents containing the aforementioned knowledge about per-formative actions and their results. Put in other terms, one observes through the three distinct glasses, systems of three distinct categories, which we call, in accordance with
Fig. 4. The aspects of information.
Fig. 5. The information management levels.
356
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363
Fig. 6. The system categories.
current practice, business processes, information systems and ICT-infrastructure components, respectively. These three system categories and their relationships are exhibited in Fig. 6. The double arrows from Business Process to Information System, and from the latter to ICT Infrastructure indicate one-to-many relationships, which should be understood as follows. Given a particular business process, one may conceive of a number of (collections of) information system(s) that support, and in doing so make operational, that business process. However at any point of time, there is only one (collection of) information system(s). Likewise, given a particular information system, one can conceive of a number of ICT infrastructure con®gurations that are able to realise, and in doing so make operational, that information system. At any point in time, however, there is one such con®guration. To conclude this section, we like to emphasize that the core or essence of a business process is, that the participants enter into and comply with commitments, and thus ®gure as social individuals. In order to do this in a responsible way, they may need to perform a lot of rational work like inquiring, calculating and evaluating alternative decisions (and probably also some formal work). This work however is only supportive to the business process, it is not part of it.
4. Action Action is a core notion in studying any dynamic system from an engineering point of view. According to the distinctions as proposed in the
previous section, DEMO distinguishes between essential actions, informational actions and documental actions. The main interest however is in the essential actions, because these are the Ôauthentic, genuineÕ business actions, all other actions only serve to support them. The class of essential actions is further divided into objective actions and social or intersubjective actions. By executing objective actions, the members of an organisation ful®ll the mission of the organisation. The nature of an objective action can be material or immaterial. Examples of material actions are all manufacturing actions in the production of goods as well as all storage and transportation actions. Examples of immaterial actions are the judgement by a court to condemn someone, the decision to grant an insurance claim, and appointing someone to be president. By executing intersubjective actions, subjects enter into and comply with commitments. In doing so, they initiate and coordinate the execution of objective actions. All intersubjective actions at the essential level of abstraction, fall into the category of regulativa (cf. Fig. 3). In order to abstract from the particular subject that performs an action and to concentrate on the functional or organisational role of the subject in performing that action, the notion of actor is introduced (e.g., sales and purchasing). An actor role can be ful®lled by a number of subjects (concurrently as well as collectively), and a subject may ful®ll concurrently a number of actor roles. In correspondance with the distinction between objective and intersubjective actions, DEMO distinguishes between two worlds in which each of these kinds of actions have eect: the object world and the intersubject world, respectively. The eect of every action is a state transition. These transitions are considered to take place instantaneously. A particular transition (e.g., requesting a beer by someone from the barkeeper) at a particular point in time (e.g., the requesting of a beer by a particular person at a particular moment) is called an event. Objective actions and their related intersubjective actions appear to occur in a particular pattern, called the (business) transaction, as illustrated by Fig. 7. It consists of three phases: the order phase
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363
357
Fig. 7. The business transaction.
or O-phase, the execution phase or E-phase, and the result phase or R-phase. (Note: The three letters O, E and R constitute the Dutch word OER which means primal, original, essential.) A transaction is carried through by two actors, who alternately perform actions. The one who starts the transaction and eventually completes it, is called the initiator (A1 in Fig. 7), the other one, who actually performs the objective action, is called the executor (A2 in Fig. 7). The order phase is an actagenic conversation, and the result phase is a factagenic conversation. Both conversations consist of communicative ( intersubjective) actions, having as eect a transition in the intersubject world (ISW). These actions are executed alternately by the initiator and the executor of the transaction. In between the two conversations, the objective action is executed, by the executor of the transaction. The eect of this action is a transition in the object world (OW). Because events in the object world are principally not knowable to the initiator (and to other actors) as long as they are not stated by the executor, transaction status 3 is coloured grey. This principal position is important. On the one hand, it stresses the supremacy of events in the intersubject world. On the other hand, it allows material and immaterial objective actions (and resulting facts) to be dealt with in the same manner. For immaterial facts it is obvious that they cannot be said to exist unless they are stated and subsequently accepted, and thus that they come into existence at the moment of acceptance, i.e., when reaching transaction status 5. Although at ®rst sight, and intuitively, material facts seem to come into existence in status 3, this appears not to be the case on
closer observation. In every organisation with material objective actions (like maufacturing or transporting ®rms), there appears always to be someone who is held responsible for a particular material fact being the case. Only he or she has the authority to declare that something is the case, which corresponds to an event in the intersubject world. Fig. 7 only shows the so-called success path of a transaction. The complete, and theoretically closed transaction process model is exhibited in Fig. 8. It has been demonstrated that this model is universal, i.e., all transactions within and between all organisations follow this pattern (Van Reijswoud, 1996). Transactions only dier in the kind of objective action they are dealing with, and in the way the intersubjective actions take place. Buying a loaf at the bakerÕs shop is an example of a very simple transaction type, where there are usually very few deviations from the success path, and where the intersubjective actions are simple utter-
Fig. 8. The transaction process model.
358
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363
ences or non-verbal acts. Moreover, they are often implicit according to the silent agreement Ôno news is good newsÕ. Buying a house is quite something else. Now, every transition in the intersubject world is performed explicitly and as unambiguously as possible. Moreover, it is documented carefully and extensively. The normal case is that a transaction follows the success path, as presented in Fig. 7. This path corresponds with the success layer in Fig. 8. The transaction then successfully ends in the terminal state 5. The discussion layer contains the possible Ôside pathesÕ a transaction may follow. It basically represents all kinds of negotiations and discussions the two actors may conduct for the sake of reaching consensus. Sometimes this consensus is not reached, meaning that the transaction ends unsuccessfully (terminal state 8). For example, instead of promising to establish the agreed upon proposition 1 ! 2, the executor can reject it 1 ! 6 for one or more of three reasons, re¯ected by the three validity claims (cf. Fig. 3). A rejection based on the claim to truth e.g., means that he is not able to establish the proposition (but perhaps a slightly dierent one). A rejection based on the claim to justice means that the executor challenges the initiatorÕs right to request him what was requested. A rejection based on the claim to sincerity means that the executor casts doubt on the sincerity of the initiator. In either case, the initiator has the option to respond in defense 6 ! 1 or to act otherwise, e.g., to stop the transaction 6 ! 8. All illocutions used in the discussion layer are regulativa, they are variants of the four kinds as presented in Fig. 3. The discourse layer represents the discussions in HabermasÕ theory about the social background, the norms and values, from which each of the actors act. The states 11 and 12 must be understood as compound states, i.e., as processes themselves. They include e.g., the recurrent policy discussions in an organisation like discussions about the acceptance rules to be applied to customer orders. The notions of action and event as presented above dier from those applied in most common approaches to information systems development and/or business process re-engineering. These approaches seem to adopt the mechanical notions of
cause and eect. Mechanical energy is always directly transferred between masses (cf. the collissions among snooker balls). Regarding mechanical processes, it is therefore perfectly alright to speak of trigger and eect. This is however not true for business processes. It is not at all harmless to say that (business) events trigger business processes. Apart from the imprecise use of the notion of event, it suggests a correspondence with mechanical cause and eect sequences. Actors however are not mechanical operators. For every action they take, they have a reason to take that particular action instead of an other one. These reasons or considerations are the basis for being held responsible for what actors do or refrain from. Even if at some time a particular rule is agreed upon, the actor applying it stays responsible for doing so, and is accountable for any eect that results from apparently irresponsible behaviour. 5. Organisation There exists a variety of de®nitions for the word ÔorganisationÕ. These de®nitions represent as many dierent notions, apparently depending on the point of view on takes towards the phenomenon organisation. As indicated in the introduction, DEMO takes an engineering point of view. This is rather new and alien for the O-professionals, since their involvement is dominated by notions of organisations of which the common denominator is that they take the black-box model of a system for granted. The black-box model is a very powerful mental tool for understanding organisations, however it is not the only one and, more importantly, it is inappropriate for understanding and changing the internal operation of a system. Fig. 9 exhibits the common graphical representation of the black-box model. Taking the black-box model when studying organisations (or any variant of it like the control model) one tries to ®nd and understand the relevant input variables and output variables, and the relationship between them, called the transfer function. Knowing the transfer function means knowing how the system responds to variations in the values of the input variables. By manipulating
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363
359
Fig. 9. The black-box model.
the input variables, one is able to change the behaviour of the system. If the transfer function is too complicated to understand, the technique of functional decomposition can be applied through which the system is replaced by a structure of subsystems of which the transfer functions are more readily understandable. One has to bear in mind however that the knowledge acquired about the system under investigation is still functional or behavioural knowledge, that it is insucient and inappropriate for changing the system. A functional decomposition of an organisation, whether it be performed in an exact way or more loosely, is very often taken to be a picture of how the organisation is ÔconstructedÕ and ÔworksÕ. However, a functional or black-box model of an organisation is only a particular conception of the organisation, a means to understand its (functional) behaviour. Like every other system (e.g., an alarm clock or a racing car), the functional behaviour of an organisation is brought about by the collective working of the constructional components. The construction and the working of a system is most near to what a system really is, to its ontological description. A very exact and very general, ontological de®nition of a system is provided by Bunge (1979). Based on this de®nition, DEMO uses the next de®nition of an organisation: Something is an organisation if and only if it ful®lls the next properties: · It has composition, i.e., it is composed of actors, where an actor is de®ned as one or more subjects in a particular role (cf. Section 4). These actors act on the basis of assigned authority and with corresponding responsibility. · It has structure, i.e., the actors in¯uence each other. Two kinds of mutual in¯uencing are distinguished. Interaction consists of executing transactions. Interstriction consists of taking into account the results or the status of other
transactions when carrying through a transaction. · It has boundary. The composition (i.e., the set of constituting actors) is divided into two subsets, called the kernel and the environment, such that every actor in the environment in¯uences, either through interaction or through interstriction, one or more actors in the kernel, and such that there are no ÔisolatedÕ parts in the kernel. The closed line that separates the kernel from the environment is called the boundary. On the basis of this de®nition, a white-box model of a system can be made, as exhibited in Fig. 10. The system elements of an organisation are the actors. On the one hand, they perform objective actions, thus changing the state of the OW. On the other hand, they perform intersubjective actions, thus changing the state of the intersubjective world (the system world in Fig. 10). When being active the actors take into account the state of the object world as well as the state of the ISW. Analoguous to the technique of functional (de)composition, there is a technique for composing and decomposing white-box models of a system. It is called constructional (de)composition. A central concept in the technique is the concept of subsystem. This concept is in DEMO de®ned as follows: S2 is a subsystem of S1 if and only if:
Fig. 10. The white-box model.
360
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363
· S1 and S2 are systems according to the de®nition above. · The kernel of S2 is a subset of the kernel of S1. · The structure of S2 is a subset of the structure of S1. · The environment of S2 is a subset of the composition of S1. In conclusion, functional (de)composition and constructional (de)composition are similar techniques, but applied to very dierent system notions. An important consequence is, that a functional component does not necessarily have a constructional counterpart. Attempting to relate the one kind of component to the other kind just does not make sense, their natures are inherently incompatible. To illustrate this, we take the decomposition of a car as an example. A functional decomposition might result into the distinction of the driving subsystem, the brake subsystem, the steering subsystem, the electrical subsystem etc. A constructional decomposition on the other hand would result into the chassis, the radiator, the wheels, the battery etc. It is obvious that each of the functional components does not correspond to some constructional component. More importantly, such correspondance is principally impossible. 6. The DEMO methodology The DEMO methodology was originally intended for carrying out the requirements engineering phase of information systems development, and it is being applied very successfully for that purpose indeed. Next to that however, it appeared to be very eective for a variety of other purposes, like Business Process Reengineering, Work Flow Management, Electronic Commerce and Virtual Organisations. In short, the methodology has proven to be a coherent, integral, and comprehensible approach to the modelling, the (re)design and (re)engineering, as well as the managing, of business processes and information systems. For several years now it is being applied in practice by a growing number of consultancy enterprises in all kinds of organisations: commercial, industrial and governmental.
The main contribution of the methodology in an organisationÕs change process, is the provision of the essential model of the organisation. This is a description by means of a number of diagramming techniques of its business processes (as de®ned by the DEMO theory), thereby abstracting fully from the informational/documental as well as from the organisational (structural) realisation. In each of the about thirty projects undertaken up to now, it has turned out that this essential model provides an ideal starting point for rethinking ®rstly the business of the organisation as such, and secondly the way in which it is realised now. Invariably, employees and managers are taken with the clarity, the conciseness, the relevance, and the eectiveness of the methodology and the produced results. Not only do DEMO projects cost a fraction of the costs of a ÔtraditionalÕ project, it also succeeds where the ÔtraditionalÕ methodologies keep failing. An extensive discussion of one of the smaller projects is provided in (Van Reijswoud et al., 1997). As an illustration of the DEMO methodology, let us look at the well-known Ford Case (Hammer, 1990). Hammer provides the next description of the business activities concerned: ``When Ford's purchasing department wrote a purchase order, it sent a copy to accounts payable. Later, when material control received the goods, it sent a copy of the receiving document to accounts payable. Meanwhile, the vendor sent an invoice to accounts payable. It was up to accounts payable, then, to match the purchase order against the receiving document and the invoice. If they matched, the department issued payment. The department spent most of its time on mismatches, instances where the purchase order, receiving document, and invoice disagreed...... One way to improve things might have been to help the accounts payable clerk investigate more eciently, but a better choice was to prevent the mismatch in the ®rst place. To this end, Ford instituted ``invoiceless processing''. Now when the purchasing department initiates an order, it enters the
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363
information into an on-line database. It doesn't send a copy of the purchase order to anyone. When the goods arrive at the receiving dock, the receiving clerk checks the database to see if they correspond to an outstanding purchase order. If so, he or she accepts them and enters the transaction into the computer system. (If receiving cannot ®nd a database entry for the received goods, it simply returns the order.)'' According to Hammer, Ford opted by the chosen solution for radical change, and achieved dramatic improvement. To illustrate this, it was mentioned that initially there were 500 people working at the accounts payable department, and that a 75% reduction of this ®gure was achieved after the solution had been implemented. Let us analyse this case brie¯y form the point of view of DEMO. To start with, Fig. 11 exhibits the relevant part of the essential model of the business activities concerned. On top of Fig. 11, a so-called transaction table is shown. It lists the distinct transaction types, including the speci®cation of the resulting fact types, as well as the speci®cation of the initiator and the executor of each transaction type. In the Ford Case, there are two transaction types: order delivery, and order payment. In the lower right corner of the ®gure the Interaction Model is shown, which can be derived directly from the transaction table. The lower left corner contains the Process Model. In this diagram each of the two transaction types is divided into its three constituent phases, the O-phase, the E-phase, and the R-phase. The solid arrows represent initiation
361
relationships. Their meaning is that the process or activity at the point side is started from the process or activity at the shaft side. For every transaction it holds that the E-phase is started from the Ophase, and that the R-phase is started form the Ephase. This is shown in Fig. 11 for the transaction types T1 and T2. Also shown is that the O-phase of transaction type T2 is started from the E-phase of transaction type T1, and that the O-phase of transaction type T1 is started externally (this is indicated by the small circle). The dotted arrow from T1/R to T2/E represents a conditional relationship. It means that T2/E can only be completed after T1/R has been completed. The Process Model expresses that the request for payment is issued in parallel with the statement that the order is delivered, and that the payment is actually performed after the order delivery has been completed. The interesting point in the Ford case is, that the essential model is the same before and after the radical(!) change. In the terminology of DEMO, this means that there has only been reengineering, not redesign. Redesign would imply that (part of) the essential model, e.g., the Process Model, would have been changed. In this respect, it sounds misleading to say that Ford has become invoiceless. This is only true at the documental level, meaning that there are no paper invoices sent anymore. The essential meaning of an invoice however is that it is a request for payment, and this request is still being performed. The dierence between the new situation and the old one, is that in the new situation the delivery of goods (also) counts as requesting for payment, while in the old situation the paper invoice conveyed this request.
Fig. 11. The essential model of the Ford Case.
362
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363
The essential model as presented in Fig. 11 (and including the transaction process models of each of the transaction types as exhibited in Fig. 8), ®rstly oers the right redesign freedom for the informational and the documental level, because it only shows the what instead of the how. It so triggers the mind for creating new solutions for the how, like the proposed change. Secondly, it oers strong support to the proposer of a change for assuring himself and others that the change is allowable, since it leaves the essential level unaffected. The sceptic reader would object, of course, that the change could also be found by applying common sense (as was done in the Ford Case). Common sense is however not always reliable, and is unevenly distributed among people. Moreover, it cannot be taught whereas a systematic approach like DEMO can. 7. Conclusions To bridge the gap between the O-people and the I-people requires more than the willingness to listen to each other. A new, engineering, perspective on business processes appears to be needed. The DEMO theory provides an appropriate framework for understanding business processes in this perspective. The essence of an organisation lies in the entering into and the complying with commitments by authorised and responsible subjects. This constitutes the working principle of any organisation. The DEMO transaction is the elementary building block of every business process, irrespective of the nature of the business, i.e., of the kind of the objective actions (material or immaterial). At the same time it becomes clear that a business process diers fundamentally from a production or a logistic process, and that so-called information intensive organisations (banks, insurance companies etc.) do have business processes like all other organisations, they only do not have production and logistic processes (at least at the essential level, they may have documental ones). Modelling business processes is a prerequisite for (re)designing and (re)engineering them. Most approaches to modelling business processes how-
ever either rely on black-box models or on ¯ow models, and therefore do not embody an appropriate understanding of the notion of business process; consequently they do not provide eective support. It has been illustrated that the DEMO methodology can be an eective help in various activities concerning the analysis and optimisation of business processes, thereby applying in an optimal way modern information and communication technology. Experiences with the methodology up to now sustain unequivocally the phrase that nothing is more practical than a good theory. In providing this theory, DEMO takes a step in the direction of the development of the badly needed discipline of Organisation Engineering.
References Auram aki, Lehtinen, E.E., Lyytinen, K., 1988. A speech act based oce modelling approach. ACM TOIS 6 (2). Austin, J.L., 1962. How to Do Things with Words. Harvard University Press, Cambridge, MA. Bunge, M.A., 1979. Treatise on Basic Philosophy, vol. 4. Reidel, Dordrecht, Netherlands. Dietz, J.L.G., 1994. Modelling business processes for the purpose of redesign. In: Proceedings of the IFIP TC8 Open Conference on BPR. North-Holland, Amsterdam. Dietz, J.L.G., Mulder, J.B.F., 1996. Realising strategic reengineering objectives with DEMO. In: Proceedings of the International Symposium on Business Process Modelling. Springer, Berlin. Dietz, J.L.G., Widdershoven, G.A.M., 1991. Speech acts or communicative action? In: Proceedings of the Second European Conference on CSCW. Kluwer Academic Publishers, Boston, MA. Dignum, F., Dietz, J.L.G., Verharen, E., Weigand, H., 1996. Communication Modeling-the Language/Action Perspective. Electronic Workshops in Computing. Springer, Berlin (available at http://www.springer.co.nk/ewic/workshops/ CM96/). Dignum, F., Dietz, J.L.G., 1997. In: Proceedings of the Second International Workshop on Communication Modelling. Computing Science Reports. Eindhoven University of Technology. Flores, F.M., Ludlow, J.J., 1980. Doing and speaking in the oce. In: Decision Support Systems: Issues and Challenges. Pergamon, Oxford. Goldkuhl, G., Lind, M., Seigerroth, U., 1998. The language action perspective on communication modelling. In: Proceedings of the Third International Workshop, Department of Informatics, J onk oping International Business School, Sweden.
J.L.G. Dietz / European Journal of Operational Research 128 (2001) 351±363 Habermas, J., 1981. Theorie des Kommunikatives Handelns. Erster Band. Suhrkamp, Frankfurt am Main. Hammer, M., 1990. Reengineering Work: DonÕt Automate, Obliterate. Harvard Business Review. Honneth, A., Joas, H. (Eds.), 1991. Communicative Action ± Essays on J urgen HabermasÕ ÔThe theory of Communicative ActionÕ. Basil Blackwell, Oxford. Medina-Mora, R., Winograd, T., Flores, R., Flores, F., 1992. The action work¯ow approach to work¯ow management technology. In: Proceedings of the Fourth International Conference on CSCW, ACM, New York. Searle, J.R., 1969. Speech Acts, an Essay in the Philosophy of Language. Cambridge University Press, Cambridge, MA. Searle, J.R., 1995. The Construction of Social Reality. Allen Lane, Penguin, London. Taylor, J.R., 1993. Rethinking the Theory of Organisational Communication: How to Read An Organisation. Ablex, Norwood, NJ.
363
Van Reijswoud, V.E., 1996. The Structure of Business Communication ± Theory, Model and Application. Doctoral Thesis. Delft University of Technology. Van der Rijst, N.B.J., Van Reijswoud, V.E., 1995. Comparing speech act based modeling approaches for the purpose of information systems development. In: Proceedings of the Third European Conference on Information Systems, Athens, Greece. Van Reijswoud, V.E., Mulder, J.B.F., Dietz, J.L.G., 1999. Communicative action-based business process and information systems modelling with DEMO. Information Systems Journal 9, 117±138. Winograd, T., Flores, F., 1986. Understanding Computers and Cognition: A New Foundation for Design. Ablex, Norwood, NJ.