Journal of Information Technology (1990) 5, 168-174

Systems thinking about information systems and strategies IAN O. ANGELL Professor of Information Systems, London School ofEconomics

Abstract: The lack of strategic thinking about IT is discussed and the advantages of adopting systems thinking are detailed. The nature of strategy and tactics with regard to information systems are investigated. The importance of continually realigning information resources with organizational and environmental requirements is highlighted.

Strategic thinking about information technology The contagion of both local and global change is generating severe insecurity among managers. That information technology (IT) will come to the rescue is a vain hope. On the contrary IT is now seen as another agent and source of this change. We are quite categorical in the belief that the present use of this technology is inappropriate for coping with such change. Organizations are now at the end of that stage of development when, in a feeding frenzy, they have bloated themselves on information technology. In the future, they must be more careful in their choice of IT diet. Without coherent ways of thinking about the place of information systems in respect of the perpetual changes around them, companies can no longer place

their faith in technological solutions. For in so doing they surrender responsibility for IT to technocrats, who indulge their fascination for irrelevant technology, without considering the wider needs of the organization. A balance must be found between the present emphasis on simple functional issues and short-term advantages of IT, against the application of information technology both to simplify and enhance a company's long-term competitive position, in the light of profound uncertainty. Organizations must be flexible in the face of the unknown, while at the same time optimizing their response to the opportunities and risks at hand. Decisions must not be based on seductive technological fashions and simplistic numerical models. Aggrandized management information systems, that are little more than computerized executive toys, must not be allowed to replace effective qualitative management with an alien quantitative imperative. The real benefits of systems simplification must not be overlooked in a confusion of complexity,

supported by an explosion of facile anthropomorphic classification, as decision support systems vie with expert systems, which vie with expert decision support systems and executive support systems. The question managers confront is not just how to integrate IS strategies with corporate strategies, but how to relate information systems to a continually changing organizational context.

Systems thinking In this situation, it is tempting to shrug our shoulders at the messiness of organizational reality and deny the relevance of any theory. Similarly, it is very easy to treat the management of information systems as a of half-remembered anecdotes and voluminous lists of guidelines. Such an approach is fundamentally futile, failing to provide any explanation or basic understanding ofinformation systems in organizations. series

Whilst there are a number of approaches in organization theory that explain, to a greater or lesser extent, the behaviour of organizations in practice, a particularly useful approach for the use ofinformation systems in management is to consider not only the information system, but also the organization, in systems terms. Information systems, and the organizations they serve, are social systems, and certain systems principles, suitably interpreted, can provide a limited understanding of their behaviour. The approach stresses the need to treat any system as a whole, hence the difference in meaning of the words systematic (using a method or rational procedure) and systemic (interpreting from the holistic viewpoint). Any intuitive understanding of management will, explicitly or implicitly, include a mixture of systematic and systemic thinking, but it

Systems thinking about information systems and strategies

must also embrace the limitations of both approaches. System is an elastic and often misused word, and the more we try to define it, the less we understand it. Successful systems thinking will involve the ability to grasp the essence of the idea, whilst accepting the implied vagueness and limitations of weak definitions such as: a system is a perceived complex assembly of parts connected together in an organized way, so that it has a coherent identity and supposed objectives, and is self-protecting and purposeful. A system is understood to have a boundary, chosen according to the observer's particular purpose and priorities, which separates it from its environment. For most mechanical devices and biological organisms it is not a problem to achieve a consensus on the choice of a boundary, however, it is highly likely that the choice of boundaries for social systems will be a source of doubtful classification (ambiguity). Two pairs of concepts are fundamental to this systems approach: emergence and hierarchy, communication and control. Within a system of organized complexity there is a hierarchy of levels of stable sub-systems, each level being a simplified interpretation of the complex behaviour of subsystems at the level below. The elements at each level being identified by emergent properties that do not exist at lower levels. Complexity of components at a lower level cannot, on their own, explain the emergent properties of a higher level. Communication between elements of a system is needed in order for it to regulate and to control itself, and to maintain its identity while interacting with its environment. It is information that is perceived as the stuff of this meaningful communication, and it is an information system, as a sub-system, that makes this communication possible. A proficient information system is one that collects and communicates sufficient information concerning changes in the environment to enable a timely, effective response by the system. The raison d'être of computerized information systems is the belief that computers can enhance the effective communication of information, that in turn enables managers to be in control of their actions and to make sensible decisions, and hence to run the organization successfully.

Causal mechanisms for control and communication are at work within any system, and feedback from a system's own outputs continuously modify elements, processes, or sub-systems within itself. Positive feedback increases the discrepancy between the actual state and some reference state chosen to identify the system; and negative feedback decreases that discrepancy. We must differentiate between on the one hand algorithmic causality and predictable mechanisms and on the other non-deterministic and emergent causal mechanisms working within uncertainty and ambiguity, that are only apparent in

169

hindsight. It is too easy to treat uncertainty and ambiguity naively as either misunderstanding or a degree of randomness, differentiating between the deterministic (predictable cause/effect relationships between input and output) and the probabilistic (with a degree of uncertainty in predicting outputs, which nonetheless do conform to some concept of a statistical distribution) and the truly random. However, even the word random has chaotic connotations, which may not be valid when considering non-systemic situations, and certainly cannot be used to describe the genesis of unpredictable emergent properties phenomena that are ordered, far from random and yet totally unpredictable. By the very nature of a consensus demarcation of a boundary, any system will have an identity, which must be maintained in a dynamic yet recognisable reference state. The system needs to be negentropic in order to maintain this state. This system must be able to adapt to the continual, predictable and unpredictable changes in the environment, and survive, reproduce, be purposeful, grow, colonize, cooperate, achieve intended results in different ways and from different initial conditions, and do all the things that physical, biological and social systems do. That the environment itself contains a thriving bubbling mass of systems, some similar, some totally different, often means that there is conflict between the individual processes that maintain communication and control, emergence and hierarchy within a given system. Some sub-systems will combine with others from the environment to form new systems, which may or may not peacefully co-exist with the original systems. The ability to maintain a relatively stable state, is not rigidity, and the negative feedback that stabilizes a system in one environment, can be positive and destructive in another. In some companies (such as Hewlett Packard) workers have agreed to cuts in pay to help it through a bad patch, whilst in others they will aggressively push the company to bankruptcy. Environments are continuously changing, and to be effective a system must be capable of a variety of responses to match those changes in the environment (Ashby's Law of Requisite Variey). It is because of this dynamic reality that we differentiate between aeterministic/probabilistic mechanisms (with parameters based on complete or partially complete information) that can be the basis of planning and tactics, and unknown processes in the environment, apparent only in hindsight, that can only be dealt with strategically. In a deliberately vague way, mechanisms are considered to be an initial state of the system, but the accrual of other components and sub-systems to this state, feedback, the reality of emergence and synergy, the development of new systems, and the breakdown of old systems from interactions within the environment,

170

Angell

imply that there can be no permanent control over a which are transferred to the brushes where they breed system that is continuously evolving and emerging among the bristles. Vigorous brushing often causes the from the mechanism. Deterministic/probabilistic gums to be scratched allowing the bacteria to enter the mechanisms, programs and procedures are designed, bloodstream directly, by-passing the natural mucous systems are not! Mechanisms can be engineered, defences in our mouths. So why should information systems cannot! An effective information system helps systems behave any less perversely? steer an organization through this environmental turbulence, taking tactical advantage of the Definitions of strategy environment, but only within the constraints of Before considering the role of information systems in strategies that will enable it to deal with the unknown. In social systems, sub-systems can be strategies, we should first ask, what is a strategy? Is it simultaneously part of the system and part of the just another trendy buzz-word that is sloshing around environment. This immediately causes a problem with the international business community? Can it be the analogy of biological organisms and introduces defined, or is it another of those nebulous concepts, doubt about an agreed perception of a boundary and the harder we try to understand it, the less we between the system and its environment, and about grasp its essence? The Oxford English Dictionary each sub-system. Such doubts about boundary states that strategy is "the art of the commander-inclassifications imply that social systems are chief; the art of projecting and directing the larger intrinsically ambiguous! We have to live with this military movements and operations in a campaign. ambiguity and make a rational analysis to isolate it Usually distinguished from tactics, which is the art of within a chosen systems boundary, and thereby handling forces in a battle or in the immediate resolve it or ignore it. Recognizing the inevitability of presence of the enemy". Most business writers agree such ambiguity, means that even the very with Von Clausewitz's observation that strategists identification of the coherent whole we call system don't achieve certainty, they only have an edge on the involves purpose-driven choice. So there can be no competition; and with Napoleon's notion of strategy as absolute analysis, design and implementation of planned flexibility, although they may not concur with systems; we are left only with perpetually changing his method of selecting his generals, 'I choose the lucky questions of appropriate or inappropriate choice. ones'. This vagueness implicit in the military meaning Within an organization (a system with its own of strategy has permeated business management, and purposes), IT professionals (sub-systems) often form materialized as numerous observations and taxonomies. Strategies are "aggregates of philosophies or peer groups with individuals in other companies, perhaps with trade unions, to compare notes and agglomerations of programmes"; they are abstracexperiences, and to optimize career opportunities. It is tions, "concepts in the minds of interested observers". not uncommon that their loyalty is to their trade (an According to Quinn and colleagues (1988) "strategy is emergent system with totally different purposes) and the pattern or plan that integrates an organization's not to their employer. And of course these individuals major goals, policies, and action sequences into a have family ties, political, philosophical, and religious cohesive whole. A well-formulated strategy helps to beliefs, all adding to the ambiguity. These systems marshal and allocate an organization's posture based may coexist peacefully or they may be involved in on its relative internal competencies and shortcomings, anticipated changes in the environment, and conflict. Schemes, based on a grand overview, are proposed contingent moves by intelligent opponents". In the and accepted within organizations, in the belief that all same article, he claims that strategic decisions are that is left is `to dot the i's and cross the t's' on the those that determine the organization's viability in programme of course nobody accepted that they light of "the predictable, the unpredictable and the were making choices, nobody considered the unknowable", leaving plenty of scope for ambiguities, and nobody told the other systems in the interpretation. environment to stay away. The joke is that these schemes may even solve the problem as intended, only to create worse problems. The 'road to hell is paved Systems, disposition, strategy, and tactics with good intentions'. We even have the situation where the system not only fails to solve a problem, it From our theoretical basis, it is now possible to give an actually achieves the exact opposite: systems interpretation of the words strategy and tactics. To inversion. One uniquely perverse example of system prosper, a system has to evolve a disposition that will inversion is from Professor Glass of the University of satisfy the Law ofRequisite Variety; that is a system must Oklahoma, who gives dire warnings in respect of have the facility, by sufficient internal variety, to toothbrushes: intended to help oral hygiene, they can respond to perpetual changes in its environment, and actually spread diseases. Mouths are full of bacteria, hence to survive and flourish. A disposition is an

Systems thinking about information systems and strategies

implicit property of every system, perhaps even the very essence of being a system, and the clue to the identity of its reference state by which we choose to classify it as a system. However, such a disposition is in no way fixed and inflexible. It is inherently ambiguous and constantly being refined and redefined via feedback from its interaction with the risks and opportunities within its changing environment. We can now introduce the adjective 'strategic' to describe any influence that has a lasting effect on the disposition of a system. This influence will have emerged from human actions, but not necessarily human designs. It will be perceived as a benefit if it increases the system's ability to generate variety, and a loss otherwise. The optimization of short term or even long-term results is not strategic, but the ability to optimize within the system under varied conditions is. A 'strategy' then is a purposeful human decision and action that is intended to be strategic. But, not all strategies are strategic and not all strategies achieve their intended strategic aims, and just as importantly not all strategic effects are the result of strategies, they can be the result of unintended actions! There can be no guarantees: with its disposition, a system may prosper in its environment to a greater or lesser extent; failure may eventually mean extinction. Time constraints cannot be imposed on a strategy, the time factor only refers to the duration of the effect on the disposition. So strategies are not identified with longterm plans! Neither are they identified with beneficial changes, but only with intended change, and that may not be achieved. We do not comment on whether strategies increase or decrease internal variety, nor whether they ultimately are for the success or the good of the system. Nor do we believe that strategy is solely the domain of top management, it is not just the purposiveness expressed by the powerful; the actions of everyone in the organization will influence its disposition to a certain extent. This view of strategy has a different emphasis from that of the business writers. They concentrate on the conscious attempts of strategists, to change, to influence and to graft preconceived notions of generic strategies onto the apparently blank sheet of an organization. But organizations have complex predispositions before the strategists even start, and these dispositions will evolve and emerge from human actions, with or without, even contrary to, the considered attentions of these strategists. System inversion is probably the most extreme situation, where the emergent disposition of the system will finesse and even diametrically oppose the original intentions of the strategists, designers, and planners. A proposed strategy will only be accepted when the receiving system is well-disposed towards it, and even then it seems reasonable that the more radical strategies be introduced incrementally (Quinn and

171

colleagues, 1988); evolution rather than revolution is needed, to avoid the ill-effects of a shock to the system. Design only works if it is not contentious to the present disposition. Misapplied and inappropriate strategies will result in tension, disruption and constant and perpetual reorganization within an organization, with highly detrimental effects. A deliberate strategy, with all its camp-followers, is itself a sub-system, and the problem of its implementation is exacerbated by the conflict and dubious classification of other internal sub-systems, each with their own dispositions, which can be ill-disposed to the expansionist intentions of this new strategy. By its disposition, a system must attempt to profit from its environment. To do this it must react to and interpret events occurring in the environment. Events that are transmitted to the system's control by an information system, where it perceives and formulates predictions of trends, models the data, derives plans and transmits tactics (via the IS) for taking advantage of a situation. So from this interpretation, competitive advantage is seen as tactical and not primarily strategic. The tactics used to achieve competitive advantage may turn out to be of lasting effect on the organization's disposition, but whether this strategic influence turns out to be beneficial is by no means certain. It is perfectly feasible for a tactically advantageous approach to reduce internal variety and so restrict future capacity to innovate. Information systems, therefore, can be used both for tactical and strategic applications, which could be mutually contradictory. Given the uncertainty and continual change in the environment, any tactical gain will be transitory; consequently changing trends will require changing tactics. The necessary ability to change perception and tactics will require much internal variety. This demonstrates the paramount importance of strategic management over short-term tactical manoeuvring; without continuous innovation and generation of the required internal variety, an organization will be in a permanent state of crisis management. An organization and its environment are both seething broths of interacting systems, and to formulate tactics we find it necessary to choose, observe, systematize, measure and interpret. This will involve the mapping of observations onto analogies and models based on previous personal experience or onto formal methods. Each system, measure, model, analysis and an' alogy is based on choice. Each is a distorting and distorted filter, and any structure we sense in the filtrate canjust as easily be a product of the filter as of the observed event. It may even be a sensed transient phenomenon which biased the choice of filter in the first place. Every model will involve arbitrary and simplistic comparisons, classifications, measures and syntheses; every analogy will involve a conscious

172

Angell

association between different events; and all danger of imposing systems thinking at levels where compounded by the chosen systematic analysis. ambiguity and disaggregation make it inappropriate, Although we have no choice but to use this approach, just as the arbitrary application of scientific thinking is we must nevertheless be aware of, and question, its unjustifiable. Joseph Weizenbaum (1986) tells a appropriateness, relevance, validity, and permanence. highly relevant story: "a policeman sees a drunk Is it any wonder then, that our interpretation of scrambling around on his hands and knees under the words such as organization, strategy, tactics, and street lamp. 'I'm looking for my keys, I lost them over system is so vague and confusing, and that the end there', the drunk says, pointing into the darkness. 'So product of systematic thinking about strategies and why look here?' asks the policeman. 'Because this is tactics so often results in no more than recipe lists, where the light is!" The particular light (scientific or agenda, or grids. Can we really expect such systems thinking) may not be capable of illuminating methodologies to give anything other than the palest the problem. shadow of organizational complexity? The dynamic Management too often deals with mechanisms and and ambiguous complexity of an organization's future functionality and not with the systemic or the just cannot be reduced to simplistic data structures, transcendental (non-systemic). They do not take into which imply a tidy and convenient homogeneity in account the magnifying effect of positive feedback in organizations that is just not there. These approaches ill-structured and turbulent environments such as the are necessarily restricted and restrictive, focusing as business world, and so ignore the butterfly effect where they do on very specific and unambiguous features of the marginal and insignificant trends of today become systems. In the process they lose all sense of the true the major business opportunities/risks of tomorrow. complexity, of wholeness and endemic uncertainty. Furthermore, management techniques based on rigid Ultimately all of these approaches are merely deterministic thinking precipitate further problems. guidelines, which come to nothing without the input of The tension, between control via rigid mechanisms and a quality individual who fully understands the the dynamic reality of systems emerging from the fluid organization and relates to the compromises being environment, is itself a source of many problems. made within the system in order to cope with its Modern management must learn from human experiinherent ambiguity. ence, and learn to live with the inevitable ambiguity and vagueness, not to deny it. The automation of such deterministic and Models and mechanisms mechanistic management processes may introduce a further level of complexity that can obscure our lack of The construction of a model of any system necessarily understanding, and serves to deny our uncertainty in places an artificial boundary around that system, and those processes. Computer models can only be a pale freezes it with a misinterpretation based on a shadow ofwhat actually happens, they cannot emulate mechanistic classification that is at best a good analogy 'being there', with all the subtle, and not so subtle, of past performance. The recent theory of checks and balances, and the unknown and Deterministic Chaos implies that even physicists no unknowable interactions. Many 'technocrats' fail to longer accept the universality of 'similar systems in comprehend that their simplistic data structures similar environments will undergo similar changes', searching for "Perfection of planning [,] is a symptom and they, after all, work in a reasonably well-behaved of decay" (Parkinson,1986). Uncertainty is not just a universe of discourse, where the meaning and choice of matter of incomplete information. It thus makes little similarity can be justified. The problem is exacerbated sense to submit the decisions that effect commercial as deterministic or probabilistic mechanisms deny the success to simplistic rules and measurement, all dynamic nature of systemicity, and misrepresent its implemented on a 'glorified adding machine'! When uniqueness born of uncertainty. This highlights the used to enhance (rather than replace) intellect, fallacy of creating a machine/model to simulate a information systems can be useful tools, and can form system; the implied ambiguity, uncertainty, and an opportunistic framework for coping with future emergence means that the model returns just one of an innovations, trends, and requirements, as well as infinity of possible universes. Lip-service may be paid minimizing dangers. But unquestioning acceptance of to open systems (where the environment has an effect), their models and simulations of open-ended human but there is an implied closed system (where nothing experience is sheer folly. outside, except pre-chosen parameters, can cross the Every mechanism and measure we use, whether systems boundary). Paradoxically even systems deterministic or probabilistic, has limitations and a thinldng can fall into this trap. We must not assume hidden agenda of choice. Their interaction with that everything is systemic; it is perfectly reasonable to elements or properties in a system reflect an act of faith accept the non-systemic groupings of systems which in the existence of the element, the property and the may (or may not) evolve later into systems. There is a system. This existence has to be perceived, a boundary

Systems thinking about information systems and strategies

imposed, and no emergent property can be measured or manipulated before it has emerged from the unknown. The measures have to be ergodic, they are numerical samples of a property of the system which must reflect the behaviour of that property through the system as a whole, and they must conform to some sense of statistical distribution. This implies that the system must be relatively stable within a given time scale and not in an unpredictable or chaotic state of flux. Just because these techniques have found universal acceptance in physics and biology, does not mean that any of these assumptions are selfevident in social systems, and so the behaviour of measures and mechanisms must be well understood and seen to be relevant to the problem at hand before they are applied.

The IS enviromnent Nevertheless, we do have to consider the place of an information system within an organization, given that so much time and capital is tied up in this new technology. It is important to stress that an information system is itself a semi-autonomous system, which has the organization as part of its environment, and so it will evolve a disposition towards the organization, as well as to the rest of its environment. We must also consider the dispositions of the environment, the organization, and its subsystems, with all their intrinsic ambiguities, towards the information system. We cannot assume that either disposition will be harmonious; this may have been the original intention, but the result can just as easily be one of major disruption. "Today, information technology must be conceived of broadly to encompass the information that businesses create and use as well a wide spectrum of increasingly convergent and linked technologies that process information" (Porter and Millar, 1985). Information technology is changing the very structure of industry, and with it the basis of competition. It spawns new products and even new businesses, often from existing operations, but it also kills off old products and companies. It changes the balance of power and creates new ways of outperforming or being outperformed by rivals. It creates and destroys barriers to entry, raises and lowers switching costs and it stresses the possibility of competitive advantage and disadvantage. Information, the very stuff of systemic communication, is essential for the cohesion and the control of the organization facing a future of perpetual change and uncertainty, which is exacerbated by the existence of IT. An understanding of the value of information within an organization must be fundamental to corporate and IT strategies. Recognition of this fact in the business community has spawned a large consultancy industry. However, there is still much

173

nonsense around, purporting to be strategy. A brochure of one well-known consultancy claimed they had produced "a five-year strategic systems/ development plan for a leading investment fund manager which justified a doubling of its spending on IT". A strategic plan may require a doubling of spending, but it certainly doesn't justify it. Furthermore, a strategy may require a certain time for its formulation and implementation, but time constraints are in the realm of the mechanistic, they cannot be imposed realistically on strategic thinking. Accepting the existence of systems and their behaviour can cause some major problems with the standard ways of dealing with information systems within organizations. An IS manager must accept that real systems are messy, vague and, more often than not, demonstrate the marginal relevance of any mechanistic interpretation of information. Much of the management of information systems is about dealing with change, and that change must be seen as a systemic reality and not necessarily systematic. Thus IS management should be driven by an understanding of the behaviour of these systems, and avoid the pitfalls of the anthropocentric view, where human intent is seen as the ultimate driving force. They may start off with pre-programmed causal mechanisms, but the outcome when introduced into the wider business environment is often uncontrollable. So much of uncertainty must be recognized as emergence, and IS managers must learn to cope with it. The only sensible long-term stance is to deny the omnipotence of efficiency and effectiveness and learn to recognize information systems within the business environment for what they are. Ashby's Law of Requisite Variety implies that IS managers should ensure their organizations have enough internal variety of responses to be able to react tactically in an appropriate manner to changes in the business environment. This calls for strategic management, a recognition of short-term tactical thinking, and the preparation and disposition of information resources within an organization, enabling it to function and to contend in a preferred way with, not only the commercial competition, but also the uncertainty, complexity and ambiguity inherent in a unknown future. The concept of predisposition within an organization, feedforward in systems terms, will involve alertness, patience, tolerance, perseverance, imagination, innovation, inventiveness, quality, commitment, motivation, productivity, experience and many of the other unmeasurable aspects. Unless the strategic context of information is fully appreciated, and unless the managers of information systems are aware of the true nature of crisis, complexity and uncertainty that awaits them, then strategies will become mere wishful thinking. Then the proponents of proactive manage-

174

ment, a fatuous misnomer, will launch deterministic mechanisms into the unknown, without being able to recognize what emerges, or to react sensibly, or to follow the feedback effect of these reactions. Grasping the place of information systems within the business reality is a never ending process.

REFERENCES Parkinson, C. N. (1986) Parkinson's Law Penguin, London. Porter, M.E. and Millar, V.E. (1985) How information gives you competitive advantage Harvard Business Review, 63, 4, 149-60. Quinn, J.B., Mintzberg, H. and James, R.M. (1988) The Strategy Process. Prentice-Hall International, London. Weizenbaum, J., (1976) Computer Power and Human Reason: From Judgement to Calculation Freeman and Co., San Francisco.

Biographical Notes Ian Angell obtained his first degree in Pure

Angell

Mathematics from the University of Wales, and his doctorate, in Computer Applications to Algebraic Number Theory, from the University of London. All of his academic career has been spent at the University of London; he was a lecturer in computing at Royal Holloway College, a senior lecturer at University College, and since 1986 he has held the Chair of Information Systems at the London School of Economics. Angell is probably best known for his work in computer graphics: he has published nine books on the subject, the latest High Resolution Computer Graphics Using C, appeared in June this year. However, for the last six years he has been concentrating on strategic information systems and on organizational and national IT policies. He is at present acting as a consultant for one Eastern European and one Middle Eastern government. With a colleague, Dr Steve Smithson, he is about to publish a book on this topic, Information Systems Management: Opportunities and Risks.

Address for Correspondence: Professor Ian Angell, London School of Economics and Political Science, Houghton Street, LONDON WC2A 2AE, UK.