Systems theory is one of the main intellectual movements of the 20th century. It arose in response to overspecialization in the sciences as a way to find a more integrated view of knowledge and the world.1 Systems theory attempts to describe and understand the common structure, attributes and emergent properties of all types of systems - physical, biological and social - by viewing them as systems per se rather than an economy or a business or a machine, for example.

A systems theory of business intelligence (BI) would position BI in the context of its surrounding system - the organizational environment in which it operates - and would predict the impact of that context on BI design. This theory would give designers a tool that guides them on what BI technology can do and what it cannot do in a given environment and the risks involved. In this article, I want to develop a general outline of this tool.

In previous articles in DM Review, I noted that organizations are cognitive systems in dialog with their environment. In order to learn from their experience, organizations collectively need to perform several important cognitive tasks: 1) sense and monitor their environment (e.g., using customer and supplier contact channels); 2) relate the information gained from this to the operating norms that guide the business (e.g., campaign management); 3) detect deviations from these norms; and 4) initiate corrective actions when deviations exceed some preset level. If these tasks are done well, a process of cybernetic information exchange is created between the organization and its overall environment.2

Some pundits describe BI as the cornerstone of this cognitive process, but, as I have shown previously, BI is more accurately described as a technical artifact that encodes a description of the business environment (i.e., the data model).3 BI helps users to understand their environment in terms that are meaningful, such as key performance indicators (KPIs) and dashboards, and facilitates predicting and controlling the business. This is built into the front-end design of the BI system as statements of purpose, scope, functionality, objectives, outputs and so forth, all of which are intended to align the BI system with the organization's strategy.

The problem arises when the organization's environment exhibits a strategy of its own. This happens when the feedback coming in from channels, customers and the larger world (e.g., regulators and competitors) doesn't match the predefined categories of knowledge, queries and other outputs anticipated in the BI system design. If we have designed the BI system to be very specific in the types of data it collects and reports, then its functioning is vulnerable to environmental disturbance. On the other hand, if we are too general in our specifications - the "toolbox" approach - we need to design on the fly for every unique situation. What design approach is right for a given environment?

A Human Systems Model

Let's start with the general human systems model shown in Figure 1 that positions the firm in a larger context, including relationships with its environment and resource base.

The systems view of the world attempts to see individual entities in their larger, connected context; a solution is seen in its relationship to other entities. Thus, when you push at one end of a problem you can anticipate the effects that might happen at another end. This is in contrast to a reductionist view of the world that looks at problems or situations in isolation.

In Figure 1, the firm is seen as a human system in its relationship to:

  • The larger world system or the environment in which the firm, its suppliers, competitors, customers and everything are embedded.
  • The external environment, i.e., everything that the firm encounters externally in its world and with which it has relationships, but has little control over: customers, regulators, competitors, etc.
  • The resource base, i.e., everything that the firm owns, uses or buys (inputs) or has some general control over: labor, capital, suppliers, etc.
  • The boundaries of the firm itself that include: 1) business actions consisting of prespecified goals, processes, rules, procedures and actions of the firm; and 2) a business model that includes the physical plant, distribution, organization structure and mental models (presuppositions) within which business actions take place.
  • Institutional memory that holds all of this in context and is accessible to the organization's members. BI is a technical artifact of institutional memory but not the entire memory, which is much broader in scope and content.
  • Feedback from the environment is the force that drives change. Single-loop feedback only requires that the firm respond by actions within the scope of its current operating framework - there is no revision of the firm's business model, its organization, vision and mission. Double-loop feedback impacts and challenges the firm's more basic assumptions and commitments - resulting in deeper inquiry into experience to examine the basis of the assumptions by which it governs itself, and it may change those assumptions in the process.

Figure 1: General Human Systems Model

The general human systems model of Figure 1 is scalable. For the firm as a whole, it applies to both the immediate environment (i.e., its marketplace and resources) and the larger world system (i.e., its industry). For a given department within the firm, the environment is its user community, and the firm is the larger world system. This is likely to be the case with BI. For the IT department, the BI user is the environment; for the BI user, the environment is likely to be a functional department such as sales, marketing, finance and production.

Positioning BI in the Larger System

In Figure 1, BI is the component of the institutional memory of the firm that holds the results of feedback from the environment relating to the firm's business actions. This is shown in Figure 1 as single-loop feedback that tells the firm how it is doing with respect to preset goals, objectives, process measurements and so forth. A single feedback loop connects an outcome of action mismatched by expectations to trigger some reaction, similar to how a room thermostat works. A dashboard for marketing, production, sales and finance does exactly this when it is coupled with the management decisions that respond to the information the BI system is delivering. BI can represent the elements of the firm that have a tangible reality and can be measured or quantified in the process of cybernetic information exchange.

The business model of the firm gets its "orders" to change from so-called double-loop feedback. BI does not directly influence the business model functions of institutional memory such as the firm's mission, values and cultural norms except to show when they are out of step with the environment. At the business-model level, an additional learning loop connects observed events with the strategies needed to formulate change - positive feedback. The business model and even the institutional memory can undergo revision. If BI is involved here, it is to integrate BI technology in such a way as to change the firm's operating model, not just to seek incremental improvement of current processes.

The constructs in the business model are carried by the institutional memory mostly in people's minds, reward structures, how authority is distributed, who is included in the "core group" and so on. The business model tells the members of the organization what to pay attention to, how to react emotionally, where they fit in and what to do in various situations. It is embedded in the organization's conscious and unconscious knowledge and is less visible than the components of the more visible and measurable business actions.

A Systems Theory for BI

Having positioned BI in the context of the firm's institutional memory, we can see that for a BI application to be sufficiently robust to serve its user community, it must be designed with a view to the dynamics of the environment. Systems theorists will immediately recognize this as an instance of the Law of Requisite Variety which states that the degree of complexity in a controller must match the level of complexity in the environment in order for the controller to manage the environment. That's similar to saying a batter needs to understand as many of the pitches a given pitcher can throw or expect to strike out. In this case, the BI application is the controller, a technical artifact that is used to organize, analyze, recommend, predict and measure the results of business actions. It is useful to examine the range of complexity that could manifest in the external environment and within the overall system. This will determine how well the BI application does its job, i.e., how robust it is for the task at hand and what can be reasonably attempted in the design.

We can differentiate the overall system environment in terms of whether it is open or closed, focused on a few goals or more pluralistic in its intent, machine-like or organic in its operation. These ideas can be summarized as a continuum of behaviors that describe the system environment listed in Figure 2.4 I consider the environmental impact on BI design in terms of its stability and structure, how clearly design goals can be stated, data feeds to the ETL (extract, transform and load) process, complexity of the data model, outputs and metrics, how the BI application can be used to guide management decisions and implementation risk.

Figure 2: Impact of Environment on BI Design

Designers can reference this figure to identify the type of environment they are dealing with and then choose the requisite complexity and function for the BI solution. One of the mistakes a design team can make is to overspecify the BI solution in relation to the environment, e.g., build a complex and highly functional BI suite to fit a heuristic business. I think that this happens as a result of scope creep and the pressure to sell a lot of software and services versus defining the real need of the client. There is so much sales hype around BI and its customer relationship management (CRM) and enterprise resource planning (ERP) cousins that it is easy for a business to overspecify its needs and to run when it should first learn to walk.

Figure 3 graphically summarizes the relationship between environmental complexity and implementation risk, as described in Figure 2.

Figure 3: Environmental Complexity vs. Implementation Risk for BI

In Figure 3, as environmental complexity evolves from rigid and deterministic organizations, the BI implementation risk generally rises to where it is highest for the purposive organization. This tracks my experience in that purposive organizations seem to be the largest market for BI technology due, in part, to the attempts of managers to integrate and control the plurality of goals pursued by multiple business units. Risk declines for heuristic organizations and rises slightly again for purpose-seeking organizations that may have trouble staying focused. This figure should give designers a clue to measure their BI implementation risk exposure for a given type of environment.

In this article, I have proposed a systems theory of BI that positions it in the context of a larger surrounding human system: the organizational environment in which it operates. The BI application is a component of institutional memory and receives feedback from its environment in a process of cybernetic information exchange. While BI technology can do a lot, there are some things it cannot do, and the limits of its capability are more likely to be determined by the dynamics of its surrounding environment than by the cleverness of designers or the technology employed. This is explained in terms of the Law of Requisite Variety and how BI function is constrained by the complexity of the surrounding business environment that spans a continuum from rigid to purpose-seeking.


  1. The "father" of systems theory is generally recognized to be the Canadian biologist, Ludwig von Bertalanffy (1901-1972).
  2. Stalinski, S. "Organizational Intelligence: A Systems Perspective." Organizational Development Journal, 22 (2004): 2.
  3. Kurtyka, Jerry. "The Limits of Business Intelligence: An Organizational Learning Approach." DM Review, June 2003: 36-41.
  4. Banathy, B. A Systems View of Education. Englewood Cliffs, NJ: Educational Technology Publications, 2004.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access