Bo Strangert (RD4)

Elaborating the case approach to R&D of defense forces

– On bias in goal analysis for complex projects 

Complex projects present you with a vast amount of uncertainty. A crucial question is whether you are capable of identifying the difficulties from the start or will stay hopeful that simple solutions will appear later on. The present paper focuses critically on preliminary goal analysis as a tool to navigate among different contextual complications.

A preceding paper (1) about the context of project planning dealt with circumstances that could influence the planning process besides the focal task factors. Two management strategies to cope with unwanted constraints were suggested: Use of a formal architecture of planning and related process coordination. A project was construed formally as a control system trying to attain its goal in a complex context through a process of coordination. Therefore, the management-by-objectives must be reassessed continuously. An inventive goal analysis is one important part of this process.

How is the goal of a project related to its context of future application? Generally, projects are defined as goal-directed or results-oriented. But how should you implement a goal direction to attain a useful application of results? One important aspect is of course when and how a project goal is formulated or elaborated.

Figure 1 illustrates some steps of formulation and analysis of project goals. Assume that the project directives emanate from some kind of master plan of development. The directives, which include a goal or task description, are input to the initial project planning.

Figure 1. Steps of formulation and analysis of project goals.

We conceive of the project planning as a control system for the succeeding development work. It will depend on contextual influences of three types: (a) the institutional planning context (e.g. restrictions from authorities, access to documented knowledge and expertise, staff organization); (b) information about real events and operations in progress; (c) the predicted context of applying the final results in real-life (e.g. through formal models and mental scenarios of hypothetical outcomes). All contexts are assumed to change dynamically.

A decisive point is how the continuous goal analysis (d) generates specific objectives (i.e. a set of means-end relations) for the development work.  The outcome of this process should be that the project results will be congruent with the real context of the application. In the end, it requires that the goal analysis should generate a valid representation of the future application context. This is to state the conditions needed to ultimately predict and control the intended development effects.

Some general requirements regarding complex projects. Project directives or original goal formulations shall not improperly influence the goal analysis. For example, they should neither include premature means-end conditions nor be too vague about desirable outcomes. However, what will be a well-balanced approach changes over time and between different parts of a project. More precisely, the issue of balancing is crucial when work methods are used for tactical, organizational, and economic goal-setting, as well as for requirement specifications of military products (e.g. TOEM and TTEM, respectively).

On the other hand, the directives, as well as the project plan, should encourage repeated cycles of CD&E, and provide adequate resources for goal refinement. This includes intense interactions with the planning context and the dynamic field of operations.

All this may sound obvious, but the shortages abound in practice. The reason is possible that it is natural to stick to standard procedures and simplifying principles when confronted with complex matters, although innovative solutions are required. Accordingly, the characteristics of the planning context are vital to maintaining an innovative culture.  Cautious examination of the operational field will also give valuable hints about faulty practices and obscure suppositions.

Examples of biases in initial planning that can restrain ultimate goals

of application

Goal-directed planning presupposes that you cautiously make frequent choices of specific objectives leading towards the ultimate goals.  In complex tasks, each choice must be made by considering multiple, mostly interacting, facets. A biased early choice in the planning can be very costly and sometimes impossible to revise. Hence, there is a need to make meticulous analyses of the consequences of each choice. Ideally, this reasoning should include some kind of ultimate goal of application as a final link.

I will exemplify a few possible biases in initial planning with some examples of different projects. The type of project is obviously an important cause of either a too wide goal conception from the start or a too narrow one. Its origin lies often in the background of the project – the definition and role of the particular project in a master plan or project directives.

(a) Studies or projects aiming at long-term defense plans

These are high-level tasks. The goal conceptions are very broad, of course. This requires a top-down approach to CD. A large number of major domains and constituents are involved in planning, for example, future threat scenarios, political intentions, budget restrictions, technical development, present organization, and available resources, and more. Many subjects are parts of the responsibilities of different authorities and expertise.

Let us define the major planning domains etc as formal sets of information, A, B, C … From a logical point of view, these sets are not mutually exclusive nor do they form an exhaustive classification of the information needed to make a proper defense plan. In reality, planning requires many interactions between different authorities and experts to result in an integrated defense plan. Also, much of the information needed has to be created during the planning process. All this promotes constraints towards simplification and saving of time.

Under such conditions, the hierarchy of domains and interests will determine the planning. Economic restrictions and governing policies are two prime candidates of constrained planning. Whether they will influence the final solution beneficially or detrimentally cannot be told beforehand. However, their supposed effects on the economy and political agreement cannot be taken for granted, as the following example shows.

Assume that one project is to design a plan for defense against possible drone attacks on a Nordic country. Economic restrictions from the start may bias planning towards using existing resources and suppressing new solutions that could be both more effective and less expensive in the long run. Likewise, the present political judgments may be outdated later, and therefore could severely limit the project outcomes. In addition, there are stakeholders from the branches of the army, navy, and air force, who may disagree about the allocation of resources for defensive actions.

A top-down planning approach with too many general restrictions can be moderated by using specially designed scenarios and wargames to establish appropriate mindsets among decision-makers about possible contextual complications. It may help to avoid overgeneralization and promote an open mind towards the need for elaborate CD&E.

Obviously, but not commonly, the guiding principle should be to avoid biased planning during the beginning stages. Freedom in a project to use inventive planning from the start would later present the decision-makers with more and qualitatively better options. It may lead to long-term economic gains. The basic preconditions for this prospect have already been stated: First, devise competent project directives, and second, select a qualified project team that takes an open systems view of the task.

(b) Project. Design a new weapon system "W"

Such a project must use systematic work methods for requirement specification. It starts by analyzing and determining the tactical, operational, and economic goals of system W (e.g. a TOEM).  This specification will probably include or presuppose some complex goal features.

In the next stage of requirement specification (e.g. TTEM), a technical systems view usually takes the lead. If possible, tactical and operational objectives are translated into technical means-end specifications. The reason is, of course, the huge availability of concrete technical design solutions which can be easily verified. However, this may introduce a biased selection of design options, because some essential but complex goal features cannot be specified and verified technically at an early stage (cf. Section 7 in Applied research methods for development projects on this site).

Theoretically, a biased design process can be described as transforming a complex but vague goal image into a complicated but specific technical system through several steps of generating more precise requirement specifications. The ease and rate of transformation depend on the availability of technical solutions. Consequently, fuzzy human or tactical needs become secondary and give reduced inventive options due to the built-in technical solutions. The possible limitations of the design may not be disclosed until the complete system W is evaluated. Otherwise, in a worst-case scenario, they will show up during an operation in a real-life context.

(c) Project: Design a new organization for logistics

The design probably takes its starting point from a doctrine (e.g. Allied Joint Logistic Doctrine, AJP-4[A], 2003). A general doctrine includes general logistic principles and policies for planning. execution and support of military operations. Logistics entails major functional areas, for example (AJP-4[A]): Supply and service functions, Maintenance and repair functions, Movement and transportation function, Infrastructure function (coordination of operational and logistic purposes), Medical function.

Organization development is required recurrently due to changes in operational missions, tactics, and techniques, as well as when new logistic methods and techniques are introduced. Complicated coordination of vastly different aspects is necessary because of the need for many options and flexible solutions. For example, missions can vary from permanent territorial defense tasks to the mobilization of rapid deployment forces. Furthermore, modern warfare and operations other than war (OOTW) involve much uncertainty and dynamics that strain the organizational preconditions to extreme limits.

One conservative strategy of development is to use the present organizational design as a starting point and try to adjust it to foreseeable challenges. Its strength lies in the amount of invested skills and data that can be utilized in planning. However, a contrasting radical strategy is to try to interpret and predict future operational and logistic needs unbiasedly. It increases the probability to invent new solutions, but at the same time, it demands more resources for basic planning and calculations given the uncertainty of outcomes.

Any major strategy will probably be biased in unpredicted ways regarding the outcome. Therefore, a desirable analysis leads to a balanced solution with both safe preconditions and smart inventions. The question is how the CD&E approach could facilitate that. Scenario construction and simulation would be useful instruments. Many actual simulations and wargames tend to be either too general or too loaded with specific details. Perhaps a better strategy is to select a few functions in turn and simulate their capabilities, given different input conditions.

For example, the flexibility and optionality of transportation could be simulated and analyzed, if the operational and logistic conditions of a dynamic network of forces are manipulated. Continuous spacetime data about all forces, vehicles, supply conditions and more could be collected. This would be an essential instrument for assessing and predicting transportation efficiency; it's nothing but a model of a network-based command and control system for logistics.

Corresponding simulations for other logistic functions could be carried out and integrated into a suitable organizational design. More about scenario construction and simulation as instruments in CD&E will be discussed incoming papers.

(d) Project: Design a course for military intelligence officers

One possible bias factor in the development of a new course program is the influence of the existing one, including the staff of teachers associated with it. Normally, the course team looks upon some ultimate course goal in the light of their teaching experience, based on regulations, rules, manuals, and inherent rationales. All this is combined with a well-formed, rational administrative structure that defines the means to fulfill the course requirements – supposed to attain the operational goal of intelligence services.

The requirement specifications of a course are mostly verified by expertise in the field. The participation of experienced intelligence officers in the verification process is highly valued as proof that an end-user perspective has been used. However, the verification of requirement specifications is not the same as a validation of the course results against the ultimate goals of intelligence operations (cf. Section 6 in Applied research methods for development projects on this site).

This case of educational structure is prone to a similar general bias as the technical systems design in case (b). Matters which can be clearly described explained, controlled, and practiced prevail in the discourse. Complex and elusive matters tend to be summarized vaguely. Such a bias depends probably on insufficient feedback from a complex field with ambiguous ultimate goals. Military intelligence certainly deals with a complex subject.

____________

There are no simple or straightforward ways of handling bias in initial project planning when the ultimate goals are complex and the contextual restrictions are numerous and strong. Possible strategies will be discussed in forthcoming papers about the URAS case study (RD1).