Uh-oh. it's a computer systems project... a project management approach designed to sidestep the high IT project failure rate
Getting ready for new hardware or software implementation? Here's a unique approach that combines structured planning with a more evolutionary process that takes into account that sticky “human factor.”
by Alan Bailey
IMPLEMENTING COMPUTER SYSTEMS is notoriously difficult. The management of such projects has a somewhat checkered history: they have a reputation for going over budget and being delivered late—that is, if they aren't cancelled in midstream. Although much has already been said and written on how to manage this type of project, I'd like to offer a slightly different perspective on the subject and suggest a simple methodology that takes into account some of the issues that cause computer projects to go wrong.
Having clear business objectives and then tailoring the project plan to the character of the project reduces the risk of project failure. The plan must address any impact that the project will have on people. In addition, those parts of a computer system that involve personal communication require an evolutionary method of design and development. A successful outcome is more probable when risk management techniques are applied to an appropriate blend of project styles.
Exhibit 1. The three basic requirements shown here, when applied in concert, reduce the likelihood of failure and form the foundation of a successful project.
Exhibit 2. Any project impacts people in some way or another; the resulting change may be welcomed by some and disliked by others. The result of a change can be viewed as a combination of the three characteristics described above. Each of the characteristics impacts people in different ways.
Structured vs. Iterative? Much of the debate concerning the management of computer projects has revolved around the use of structured planning versus a looser, more iterative approach. A structured approach entails management of project progress against a detailed task plan. In the looser method, the developer works with the customer to evolve the system toward some defined objective. While both approaches have their advantages and disadvantages, neither is a complete solution.
The advocates of the structured style of management often cite the success of this approach in other disciplines, such as structural engineering. If we view computer system development as a form of engineering, an engineering approach to management ought to work. Unfortunately, building a computer system is not quite the same as carrying out an engineering project. In engineering, initial analysis is followed sequentially by design and then construction. In developing a computer system, it is often beneficial to merge these steps in a series of iterations as the final product evolves.
The iterative, evolutionary method of developing systems can be successful because it enables the system developers and their customers to communicate effectively with each other about what is being built. It is very difficult to visualize in advance from a specification or design how a system will work or how it will impact the organization that uses it. The problem is compounded by the fact that implementing a system normally requires a change to the way people do things. Indeed, implementing a new system may itself change people's views of what they really need.
The problem with the iterative approach to system development is the difficulty of predicting where the system evolution will lead. Uncertainty over what is involved in the development can then lead to a loss of management control over the project. Conversely, the precision of a detailed task plan is often illusory, because of unpredictable issues and changes that occur during the course of a project.
However, management of system development work is important because:
■ We normally need to understand in advance the cost, benefit and timing of implementing a new system in order to make rational decisions about whether to make the necessary investment.
■ Having made the decision to proceed with a system implementation, we need to ensure that anticipated benefits are achieved within acceptable costs and time scale.
■ We need to plan in advance the availability of people and other resources required for the project.
The purpose of project management is to minimize the risk of failure with respect to all of these points.
This risk minimization view of project management leads to some simple guiding principles. As illustrated in Exhibit 1, there are three basic requirements that, together, reduce the likelihood of failure and form the foundation of a successful project:
■ We need clear business objectives so that everyone understands what needs to be done.
■ We need to understand the nature of the change so that we can determine how best to carry out the project.
■ We need to understand and manage the project risks by anticipating potential problems.
Let's look at each of these requirements to see how, by taking a flexible view of project planning, it's possible to leverage the benefits of both structured and evolutionary approaches to system development.
Setting Objectives. Project success is, first of all, critically dependent on clear objectives. Objectives establish direction for the project team and expectations for the project clients. They reduce the potential for misunderstandings of what the project is about and what it will achieve. They set a stable basis against which project progress and project change questions can be judged.
The objectives need to be written in business terms, rather than in technical terms. Without clear business direction, it will not be possible to establish successful project completion. A common error of this type is to express a project objective as the delivery of a new or revised system. To state, for example, that the purpose of a project is “to implement a new financial system” is meaningless unless we understand what the new system is expected to achieve.
In this case, a better objective would be “to reduce the time taken to close the company annual accounts from one month to two weeks.” A simple business statement of this type is easy to understand; it also provides the basis for determining project benefits and measuring project success. Throughout the project, everyone will understand what he or she is trying to achieve.
Exhibit 3. In computer systems we focus too often on the tangible factors at the expense of the conceptual and personal effects, resulting in a technically oriented project that disregards the impact on people. We need to characterize our projects according to all three factors. This exhibit illustrates the possible classification of a number of development projects.
Objectives need to be specific and measurable. A prime objective of developing a new intranet site may be “improved staff communications.” Although this seems a reasonable business target, it is too imprecise for effective project control. A better objective might be “to make new personnel policies available to all company staff within 24 hours of issue.” This latter statement sets clear and unambiguous scope and direction for the project work.
Reader Service Number 5133
Exhibit 4. The planning approaches used for project management can be broadly categorized as shown here. Although quite different from each other, each approach can be a valid way of doing things, since it corresponds to a change factor. Linear planning works well for tangible changes, exploratory planning works well for conceptual changes, while personal planning works for personal changes.
Project objectives should be directly linked to business benefits. Thus, by setting clear business goals we have a basis for evaluating the business value of what is proposed. In addition, by verifying the achievement of benefits, it will be possible to measure the success of the work after project completion.
Understanding and Planning the Change. Having clarified the project objectives, we need to evaluate the project in order to determine how to carry it out. By understanding the nature of the change that the project will bring about, we can decide what style of management to adopt. There is no one correct way to conduct a project; a highly structured approach suits some projects but not others. The key factor that determines the management approach is an understanding of how the project will interact with the people it impacts.
People are fundamental to any management strategy. Projects are done by people for people. This is especially true for computer systems, which impact organizations and working practices. More computer projects fail as a result of miscommunication between technical experts and their customers than fail as a result of technical problems. We forget the human side of what we do at our peril!
Any project impacts people in some way or another. The resulting change may be welcomed by some and disliked by others. The change should result in some identifiable business benefit. However, successful implementation depends on the participation of all people who are impacted.
The result of a change can be viewed as a combination of three characteristics, which I am going to call tangible, conceptual and personal. These characteristics are described in Exhibit 2. Each of the characteristics impacts people in different ways.
In computer systems we focus too often on the tangible factors at the expense of the conceptual and personal effects. The result is a technically oriented project that disregards the impact on people. What we need to do is characterize our projects according to all three factors. Exhibit 3 illustrates the possible classification of a number of development projects.
As an example, the development of an intranet site for distributing internal company information is likely to have high conceptual and tangible components; the site needs to be aesthetically and ergonomically pleasing, but it also requires some computer system development and installation. On the other hand, the implementation of a new financial accounting system has strong personal and tangible components; it will impact the way in which the accountants work as well as being a complex system from a technical standpoint.
The approaches used for project management can also be broadly categorized in three ways: linear, exploratory and personal. These approaches are described in Exhibit 4. Although each approach is quite different, each can be a valid way of doing things.
Each planning approach tends to be effective for a corresponding change factor. Linear planning works well for tangible changes, exploratory planning works well for conceptual changes, while personal planning works for personal changes. As suggested in this exhibit, personal planning tends to be characterized by communication with people, exploratory planning characterized by experimentation, linear planning characterized by a high degree of rigid structure.
Every project will involve a combination of change characteristics, so a corresponding combination of project management styles will be required. As a result, the management approaches outlined above need to be blended together to suit individual circumstances. This can be illustrated by considering the examples of the intranet site and the financial accounting system.
The intranet site is highly conceptual in nature. This suggests an exploratory style of approach in which there is an iterative development in small increments. Prototyping and experimentation are used to flesh out the design. The tangible component of the work is concerned with the practical purpose of the site and the construction of the supporting technology; these aspects of the project require a linear approach, with a detailed task plan.
By contrast, the strong personal component of the accounting system requires personal planning. It is simply not realistic to assemble a linear task plan which, for example, says “on week four the accountants will be persuaded that the new system is a good idea.” An outline of how the personal changes will be approached is needed, with a strong emphasis on personal communication. On the other hand, the technical components of the system are tangible: these will require a linear task plan for their development. The personal and the linear plans obviously need to be synchronized.
In both of these projects, a plan containing tasks and deliverables will be developed. However, the level of detail and degree of precision in each plan will depend on the blend of management styles that is being adopted. Tangible, technical parts of a project will contain detailed tasks, while personal plans will be more outline in nature.
Reader Service Number 5021
Exhibit 5. By setting clear objectives and taking into account the nature of the proposed changes, it is possible to plan computer-related projects in a way that is both realistic and practical. The planning process must take into account the influence of people on the work as well as the technical aspects of the project. By then applying some risk management techniques, project success can be assured.
Risk Management. Having set project objectives and assembled a project plan, we now need to focus on the risk of project failure. This can be done by brainstorming a list of things that can go wrong. It is often beneficial to take each component of a project plan and then question the underlying assumptions behind it: what if these assumptions are incorrect?
We can then take each potential problem or risk and decide how to manage it. Risks can be managed in two ways: avoidance and contingency. Avoidance involves carrying out the project is a way that eliminates or minimizes the probability of the problem occurring. Contingency involves having an alternative plan for use if the problem occurs. In either case, the risk management will probably require some changes to the overall project plan.
In practice, it is rarely possible to identify all potential problems at the start of a project. As a result, a number of actions need to be taken in order to limit exposure to risk.
■ When listing risks during planning, focus on the problems that could seriously derail the project. The most important thing is to foresee the major issues where project failure could result.
■ Evaluate the degree of uncertainty in the project plans in order to set realistic client expectations. It is much better to alert a client to potential problems at the outset than to allow problems to appear as a surprise during project execution.
■ Have a change control process for managing the project so that unanticipated problems and changes can be dealt with effectively. In particular, it is important to have a client sponsor or project board so that decisions can be made regarding major issues.
BY SETTING CLEAR OBJECTIVES and taking into account the nature of what is proposed, it is possible to plan computer-related projects in a way that is both realistic and practical. The planning process must take into account the influence of people on the work as well as the technical aspects of the project. By then applying some risk management techniques, project success can be assured. The complete process is illustrated in Exhibit 5.
Traditional project management tends to take a somewhat technical view of the world, assuming that issues can be completely rationalized and solved in predictable ways. By using a more flexible approach, which takes into account the uncertainty and changeability associated with people, we can plan more realistically. We sacrifice some apparent—but unreal— precision. We gain, however, a much clearer view of what needs to be done and what is actually achievable. In addition, by using the methods I am suggesting, we can maintain a rationale for how we are managing any project, regardless of its nature. ■
Alan Bailey is an independent information technology consultant. He has over 20 years of experience in designing, building and implementing computer systems, and has worked as a project manager and quality management consultant for a major international oil company.
PM Network • August 1998
PMI research shows project teams that draw from an array of perspectives and skillsets deliver powerful outcomes.