The debate between Agile and Waterfall is usually framed as a battle between two opposing camps, when in the mature practice of project management they are simply two tools with different contexts of application. The right question is not "which one is better?" but "which approach fits the nature of this project, its uncertainty and its constraints?" This article explains both frameworks with precision, presents the hybrid model that combines the best of each, and sets out clear criteria for choosing between them — aligned with the PMI's PMBOK standard and frameworks such as PRINCE2.
Waterfall: the predictive, phase-based model
The waterfall approach is a predictive, sequential model. The project advances through well-defined phases (requirements, design, construction, testing, deployment and closure), each completed before the next begins. Its strength is predictability: scope, schedule and cost are fixed at the outset, which facilitates fixed-price contracting, resource planning and documentary traceability. It is the natural approach for civil construction, manufacturing or regulated deployments where the scope is clear and the cost of late change is very high.
Its weakness is rigidity. If requirements change mid-project — something common in software and in innovation — the waterfall model absorbs those changes poorly: each modification forces a return to earlier phases and drives up cost. The well-known "big bang effect" — delivering everything at the end — concentrates risk at the most expensive moment to correct it.
Agile: the adaptive, iterative model
The Agile approach, formalised in the Agile Manifesto of 2001, is iterative and incremental. Instead of planning everything upfront, it delivers value in short cycles and adjusts direction based on customer feedback. Its four values prioritise individuals and interactions over processes, working software over comprehensive documentation, customer collaboration over contract negotiation, and responding to change over following a fixed plan.
The most widely adopted Agile framework is Scrum, which organises work into fixed-length sprints (typically one to four weeks), with defined roles (Product Owner, Scrum Master, development team) and structured events (sprint planning, daily stand-up, review and retrospective). The prioritised backlog ensures that work always focuses on what delivers the most value. Kanban, by contrast, manages a continuous flow of work by limiting work in progress (WIP) to reduce bottlenecks, without fixed sprints.
The hybrid model: the best of both worlds
The reality of most complex projects is hybrid. A common approach, sometimes called "Water-Scrum-Fall," applies waterfall to the initial definition and final deployment phases — where predictability and documentation are critical — and Agile to the construction phase, where learning and change are the norm. This makes it possible to close a contract with a defined scope and budget while executing development through increments validated with the customer.
The PMI explicitly recognises this continuum. The seventh edition of the PMBOK abandoned the prescription of rigid processes and adopted a principle-based approach centred on selecting the lifecycle — predictive, adaptive or hybrid — best suited to each project. The PMI-ACP certification and PRINCE2 Agile itself both reflect the fact that the boundary between frameworks has become permeable, and that professionals must command the full repertoire.
Scaling Agile: when one team is not enough
Scrum works well with a team of three to nine people, but large projects involve dozens of teams that must coordinate without losing agility. Scaling frameworks were developed for exactly this challenge. SAFe (Scaled Agile Framework) organises multiple teams around "Agile Release Trains" with joint planning by programme increments. LeSS (Large-Scale Scrum) extends Scrum to multiple teams while maintaining a single backlog and a single Product Owner to preserve simplicity. Scrum of Scrums is the lightest mechanism: representatives from each team meet periodically to synchronise dependencies.
The choice of scaling framework is not trivial and must reflect the organisation's size and culture. Adopting SAFe in a small company typically introduces unnecessary bureaucracy, while relying solely on Scrum of Scrums in a programme with fifty teams leaves critical dependencies ungoverned. The guiding principle is the minimum process required to coordinate without suffocating team autonomy.
Estimation, risk and earned value management
Each approach measures progress with different instruments. In predictive projects, earned value management (EVM) compares planned value, earned value and actual cost to calculate cost performance index (CPI) and schedule performance index (SPI) — metrics that flag deviations before they become irreversible. In Agile projects, relative estimation with story points and team velocity allow forecasting when the backlog will be complete, while burndown charts reveal whether scope is growing faster than the team can absorb it.
Risk management, for its part, is transversal to any framework. A live risk register — with probability, impact and a response plan for each threat — prevents the project from advancing blindly. In Agile environments, a large part of risk is managed implicitly by delivering early and validating with the customer, which reduces the exposure to building something nobody wants. In predictive environments, risk is addressed explicitly in the initial planning and revisited at each milestone.
| Dimension | Waterfall | Agile | Hybrid |
|---|---|---|---|
| Scope definition | Complete upfront | Emergent and evolutionary | Initial frame + iterative detail |
| Change management | Costly and formal | Expected and welcomed | Bounded by phases |
| Value delivery | At the end of the project | Incremental and continuous | By milestones and increments |
| Ideal context | Stable requirements | High uncertainty | Mixed constraints |
| Key metric | Plan adherence | Velocity and value delivered | Milestones + velocity |
How to choose the right approach
The choice depends on measurable variables. If requirements are stable and well understood, and the cost of late change is high, waterfall reduces risk. If requirements are uncertain or the market is evolving rapidly, Agile enables learning and adjustment. The contracting model also matters: a fixed-price contract pushes towards waterfall, while a continuous collaborative relationship enables Agile. Finally, client maturity counts: Agile requires a client who is available and committed to ongoing decision-making; without that involvement, the Agile approach degrades.
Supporting tools and artefacts
Regardless of the framework, certain tools structure the management effort. Gantt charts and the critical path method (CPM) organise predictive projects. Kanban boards and burndown charts visualise Agile progress. The Work Breakdown Structure (WBS) serves both. Project management platforms integrate backlog, sprints, dependencies and reporting, but it is worth remembering that the tool does not replace the method: a digital Kanban board used by an undisciplined team does not produce Agile results.
Steps to implement a project methodology
- Characterise the project: assess requirement uncertainty, schedule criticality and the contract model before choosing a framework.
- Select the lifecycle: predictive, adaptive or hybrid, justifying the decision against those criteria.
- Define roles and governance: who decides on scope, who prioritises, and how blockers are escalated.
- Establish the cadence: milestones in waterfall or sprints in Agile, with clear review events.
- Measure and retrospect: progress and value indicators, plus continual improvement of the management process itself.
Common mistakes
The first mistake is cardboard Agile: adopting Scrum vocabulary (sprints, dailies) without its principles, while in practice running a waterfall disguised with micro-management. The second is forcing Agile onto fixed-scope, fixed-price projects, where it generates contractual friction. The third is the opposite: applying waterfall to a high-uncertainty project and discovering at the end that the wrong thing was built. The fourth is forgetting the retrospective: without improving the process itself, any methodology stagnates.
Frequently asked questions
Is Agile only for software?
No. It originated in software development, but its principles apply to marketing, R&D, events and any project involving uncertainty and the need to iterate. What matters is not the sector but the degree of uncertainty in the scope.
Can I combine Agile and Waterfall in the same project?
Yes, and it is the most common approach in complex projects. The hybrid model applies waterfall to definition and deployment and Agile to construction, combining contractual predictability with the ability to adapt during development.
What does the PMBOK say about these approaches?
Since its seventh edition, the PMI's PMBOK does not prescribe rigid processes but is principle-based and recognises predictive, adaptive and hybrid lifecycles. The project manager selects the approach according to the project context, rather than applying a single recipe.
Scrum or Kanban?
Scrum fits when work is planned in cycles with defined goals and a stable team. Kanban suits continuous flows of requests with shifting priorities, such as support or maintenance work. Many teams combine both in what is known as Scrumban.
Conclusion
No project management framework is universally superior: there is only the right framework for each combination of uncertainty, contract type and client maturity. Waterfall offers predictability when scope is stable; Agile offers adaptability when scope is uncertain; and the hybrid model — now recognised by the PMI itself — allows commitments to be fixed without sacrificing the ability to learn along the way. The true differentiator of a good project manager is not the defence of a dogma but the ability to diagnose the project and select the lifecycle that minimises risk and maximises the value delivered. At Summum Consultoría we help organisations characterise their projects and implement the methodology — predictive, adaptive or hybrid — that best fits their reality, training teams so that the chosen method endures over time.