Common Errors in Requirements Analysis

In the traditional waterfall model of software development, the first phase of requirements analysis is also the most important one. This is the phase which involves gathering information about the customer's needs and defining, in the clearest possible terms, the problem that the product is expected to solve.

This analysis includes understanding the customer's business context and constraints, the functions the product must perform, the performance levels it must adhere to, and the external systems it must be compatible with. Techniques used to obtain this understanding include customer interviews, use cases, and "shopping lists" of software features. The results of the analysis are typically captured in a formal requirements specification, which serves as input to the next step.

Well, at least that's the way it's supposed to work theoretically. In reality, there are a number of problems with this theoretical model, and these can cause delays and knock-on errors in the rest of the process. This article discusses some of the more common problems that project managers experience during this phase, and suggests possible solutions.
Problem 1: Customers don't (really) know what they want

Possibly the most common problem in the requirements analysis phase is that customers have only a vague idea of what they need, and it's up to you to ask the right questions and perform the analysis necessary to turn this amorphous vision into a formally-documented software requirements specification that can, in turn, be used as the basis for both a project plan and an engineering architecture.

To solve this problem, you should:

• Ensure that you spend sufficient time at the start of the project on understanding the objectives, deliverables and scope of the project.
• Make visible any assumptions that the customer is using, and critically evaluate both the likely end-user benefits and risks of the project.
• Attempt to write a concrete vision statement for the project, which encompasses both the specific functions or user benefits it provides and the overall business problem it is expected to solve.
• Get your customer to read, think about and sign off on the completed software requirements specification, to align expectations and ensure that both parties have a clear understanding of the deliverable.

Problem 2: Requirements change during the course of the project

The second most common problem with software projects is that the requirements defined in the first phase change as the project progresses. This may occur because as development progresses and prototypes are developed, customers are able to more clearly see problems with the original plan and make necessary course corrections; it may also occur because changes in the external environment require reshaping of the original business problem and hence necessitates a different solution than the one originally proposed. Good project managers are aware of these possibilities and typically already have backup plans in place to deal with these changes.

To solve this problem, you should:

• Have a clearly defined process for receiving, analyzing and incorporating change requests, and make your customer aware of his/her entry point into this process.
• Set milestones for each development phase beyond which certain changes are not permissible -- for example, disallowing major changes once a module reaches 75 percent completion.
• Ensure that change requests (and approvals) are clearly communicated to all stakeholders, together with their rationale, and that the master project plan is updated accordingly.

Problem 3: Customers have unreasonable timelines

It's quite common to hear a customer say something like "it's an emergency job and we need this project completed in X weeks". A common mistake is to agree to such timelines before actually performing a detailed analysis and understanding both of the scope of the project and the resources necessary to execute it. In accepting an unreasonable timeline without discussion, you are, in fact, doing your customer a disservice: it's quite likely that the project will either get delayed (because it wasn't possible to execute it in time) or suffer from quality defects (because it was rushed through without proper inspection).

To solve this problem, you should:

• Convert the software requirements specification into a project plan, detailing tasks and resources needed at each stage and modeling best-case, middle-case and worst-case scenarios.
• Ensure that the project plan takes account of available resource constraints and keeps sufficient time for testing and quality inspection.
• Enter into a conversation about deadlines with your customer, using the figures in your draft plan as supporting evidence for your statements. Assuming that your plan is reasonable, it's quite likely that the ensuing negotiation will be both productive and result in a favorable outcome for both parties.

Problem 4: Communication gaps exist between customers, engineers and project managers

Often, customers and engineers fail to communicate clearly with each other because they come from different worlds and do not understand technical terms in the same way. This can lead to confusion and severe miscommunication, and an important task of a project manager, especially during the requirements analysis phase, is to ensure that both parties have a precise understanding of the deliverable and the tasks needed to achieve it.

To solve this problem, you should:

• Take notes at every meeting and disseminate these throughout the project team.
• Be consistent in your use of words. Make yourself a glossary of the terms that you're going to use right at the start, ensure all stakeholders have a copy, and stick to them consistently.

Problem 5: The development team doesn't understand the politics of the customer's organization

The scholars Bolman and Deal suggest that an effective manager is one who views the organization as a "contested arena" and understands the importance of power, conflict, negotiation and coalitions. Such a manager is not only skilled at operational and functional tasks, but he or she also understands the importance of framing agendas for common purposes, building coalitions that are united in their perspective, and persuading resistant managers of the validity of a particular position.

These skills are critical when dealing with large projects in large organizations, as information is often fragmented and requirements analysis is hence stymied by problems of trust, internal conflicts of interest and information inefficiencies.

To solve this problem, you should:

• Review your existing network and identify both the information you need and who is likely to have it.
• Cultivate allies, build relationships and think systematically about your social capital in the organization.
• Persuade opponents within your customer's organization by framing issues in a way that is relevant to their own experience.
• Use initial points of access/leverage to move your agenda forward.