6.1 When to Use a Predictive Approach
Key Takeaways
- Predictive (waterfall) approaches work best when requirements are well-defined, stable, and unlikely to change significantly
- Projects with high regulatory or compliance requirements often benefit from predictive approaches due to documentation needs
- Key indicators for predictive include low uncertainty, well-understood technology, stable scope, and fixed-price contracts
- Organizational structures like functional and weak matrix tend to align better with predictive approaches
- The Stacey Complexity Model helps determine the appropriate approach based on requirements clarity and technology certainty
When to Use a Predictive Approach
The CAPM exam tests your ability to determine when a predictive (plan-based) approach is the most appropriate methodology for a project. This is Domain 2, Task 1 of the Exam Content Outline.
Characteristics of the Predictive Approach
The predictive approach, also known as waterfall or traditional, follows a linear, sequential process:
Requirements → Design → Build → Test → Deploy → Close
Each phase is largely completed before the next begins, with formal reviews (phase gates) between phases.
When Predictive Works Best
| Factor | Predictive Is Appropriate When... |
|---|---|
| Requirements | Well-defined, documented, and stable from the beginning |
| Scope | Fixed and unlikely to change significantly |
| Technology | Well-understood, proven, and low technical risk |
| Uncertainty | Low — few unknowns about what needs to be built and how |
| Regulations | High compliance or regulatory requirements demand documentation |
| Contracts | Fixed-price contracts where scope must be defined upfront |
| Stakeholders | Stakeholders prefer defined milestones and deliverables |
| Organization | Traditional, hierarchical organizational culture |
When Predictive Is NOT Appropriate
| Factor | Predictive Is Risky When... |
|---|---|
| Requirements | Unclear, evolving, or subject to frequent change |
| Innovation | The product has never been built before |
| Technology | New or unproven technology with high technical uncertainty |
| Market | Rapidly changing market conditions require quick pivots |
| Feedback | Early and continuous stakeholder feedback is essential |
| Speed | Time-to-market is critical and incremental delivery adds value |
The Stacey Complexity Model
The Stacey Complexity Model (also called the Stacey Matrix) helps determine the appropriate approach based on two factors:
| Requirements are CLEAR | Requirements are UNCLEAR | |
|---|---|---|
| Technology is CERTAIN | Simple → Predictive approach | Complicated → Hybrid approach |
| Technology is UNCERTAIN | Complicated → Hybrid approach | Complex/Chaotic → Adaptive (Agile) approach |
Application to Approach Selection
- Simple (clear requirements + certain technology): Use Predictive
- Complicated (some uncertainty in either dimension): Use Hybrid
- Complex (significant uncertainty in both dimensions): Use Adaptive
- Chaotic (extreme uncertainty): Use Adaptive with rapid experimentation
Organizational Structures for Predictive Projects
| Structure | Fit with Predictive | Why |
|---|---|---|
| Functional | High | Hierarchical control, specialized resources |
| Weak Matrix | Moderate-High | Functional managers maintain authority |
| Balanced Matrix | Moderate | Shared authority works for structured projects |
| Strong Matrix | Moderate | PM has authority within structured environment |
| Projectized | Moderate | Full PM control, but less need for it in stable projects |
Predictive Approach Advantages and Limitations
| Advantages | Limitations |
|---|---|
| Clear structure, defined milestones | Inflexible to changes |
| Easier to estimate costs and schedule | Late discovery of issues (testing is late) |
| Comprehensive documentation for compliance | Customer sees product only at the end |
| Well-understood by most organizations | Assumes requirements are complete and stable |
| Works well with fixed-price contracts | Scope creep risk if requirements were incomplete |
| Straightforward progress measurement | Long delivery cycles before value is realized |
Which of the following projects is MOST suited for a predictive (waterfall) approach?
According to the Stacey Complexity Model, what approach is recommended when both requirements and technology have HIGH uncertainty?
Which is a key limitation of the predictive approach?