Waterfall and the Sequential Mindset
The oldest way to organize software work is also the most intuitive: do the phases in order, finishing each one completely before starting the next. Gather all the requirements, then do all the design, then write all the code, then test it all, then release. This is the Waterfall model.
The name paints the picture: like water flowing down a series of steps, work moves in one direction, from each phase down to the next, and doesn't flow back up. You sign off on a phase, then you're in the next one.
Picture building a house from a blueprint that can't change once the foundation is poured. Plan everything, then build exactly that. It's orderly and easy to explain — and it assumes you knew everything you needed at the start.
How Waterfall Works
Each phase produces a document or deliverable that's reviewed and approved before the next phase starts. Requirements are written down and signed off; then designers work only from that signed-off document; then developers build only from the approved design. The appeal is control and predictability: at any moment you know exactly which phase you're in, and there's a paper trail for every decision.
Where It Shines
Waterfall genuinely fits some work: projects where the requirements are fixed, fully understood up front, and unlikely to change. Building software to a government contract, or embedded code for a medical device that must be certified, can suit a sequential approach — the rules are known in advance and changing course midway is costly or forbidden. When you really can specify everything correctly at the start, Waterfall's order is a strength, not a weakness.
Where It Breaks
The trouble is that most software isn't like that. Requirements change, users surprise you, and you rarely know everything up front. Waterfall's fatal flaw is timing: because testing and real use come at the very end, you discover your biggest mistakes when they're most expensive to fix — the far-right, steep side of the cost-of-change curve. Months of work can be built on a misunderstanding that nobody catches until the end.
Its Lasting Influence
Even though most teams have moved on, Waterfall's vocabulary is everywhere. The phase names — requirements, design, implementation, testing — come straight from it, and even fast Agile teams still do all of those activities, just blended and repeated instead of strictly in sequence. Understanding Waterfall is understanding the baseline that everything since has been reacting to.
Imagine the Cadence team had built the whole app Waterfall-style: a year of planning, designing, and building every feature, then one big launch. They'd have discovered only at the very end that users desperately wanted reminders — far too late to add cheaply. That painful "we built the wrong thing and found out last" is exactly the failure the next topic's approach was invented to avoid.
- "Waterfall is just wrong and obsolete." It fits some domains well — fixed, fully-known, change-resistant work. It's a poor default for uncertain projects, not a universal mistake.
- "Waterfall means no testing until the very end." Testing is a real phase in it, but feedback from real use comes late — which is the actual weakness, not the absence of testing.
- "Agile completely replaced Waterfall everywhere." Plenty of real work — contracts, regulated systems, hardware-bound projects — is still done sequentially. It's a tool with a narrower fit, not an extinct one.
- Waterfall is the baseline every other method reacts to — you can't fully understand why Agile exists without seeing the problem it was solving.
- Its weakness is a direct, concrete example of the cost-of-change curve: late feedback means expensive fixes.
- Its vocabulary — requirements, design, implementation, testing — is still the shared language of the whole field, including on Agile teams.
Knowledge Check
How does the Waterfall model organize work?
- Each phase is finished and signed off before the next one begins
- The phases run in small repeating loops with feedback each time
- The phases happen in whatever random order the team happens to feel like
- All phases happen at exactly the same time
What is Waterfall's main weakness?
- Mistakes are discovered late, when they're most expensive to fix
- It is impossible to make any plan at all
- It produces absolutely no documentation of any of the decisions made
- It can only be used by very large teams
Which kind of project is Waterfall genuinely well-suited to?
- One with fixed, well-understood requirements that are unlikely to change
- A startup that is rapidly experimenting to discover what its users actually want
- A project where the requirements change every week
- Any project, as long as the team is small enough
Why does the topic say Waterfall still matters even though most teams have moved on?
- Its vocabulary and phases are the baseline every later method reacts to
- Because Waterfall is secretly the best method for every project
- Because every single Agile team secretly runs pure Waterfall behind the scenes anyway
- Because the law requires all software to use Waterfall
You got correct