When “Waterfall” Became an Insult
A caricature is more useful to a movement than an accurate description, because you cannot rally against nuance.
The word that ends the conversation
Watch what happens in a delivery review the moment someone describes an approach as “waterfall.” The word does not describe a sequencing choice — plan the work, then build it. It delivers a verdict on the person who proposed it: yesterday’s practitioner, command-and-control, afraid of change, not one of us. And the reply, when it comes, is almost never about delivery. It is a defence of character.
That is the tell, and it is worth naming plainly at the end of a year in which the argument has only grown louder. When a technical choice can no longer be discussed without someone’s professional identity being put on trial, we have stopped debating methods and started defending tribes.
How a method became a membership card
The manifesto that started all this is not yet ten years old, and for most of that decade agile named a set of working practices — short iterations, close collaboration, working software over documentation — that a team could adopt, adapt, or ignore on the evidence. Somewhere in the last two or three years it became something else: first a movement, and then an identity.
Movements are powerful precisely because they draw a line. They need an inside and an outside, the converted and the unconverted. Waterfall is the outside. It is telling that almost no one accused of it would ever use the word to describe their own work; it survives only as an epithet, a caricature of a straw practitioner who gathers requirements for a year and then refuses to look at them again. Real sequential delivery, done by capable people, never looked like that. But a caricature is more useful to a movement than an accurate description, because you cannot rally against nuance.
The symptoms of the shift are easy to spot once you look:
- Certification has become a rite of passage rather than a measure of competence — the badge signals belonging, not skill.
- Teams announce that they are agile, in the present tense, as an identity — the way one might state a nationality — rather than describing what they actually do on a Tuesday.
- Method choices are defended in the vocabulary of conviction — “we don’t do upfront design” — where the honest sentence would begin “this particular piece of work needs…”
- Conference stages and mailing-list arguments reward purity, and purity is the enemy of judgement.
Why the tribe costs more than it clears
Here is the practical damage. The only question that matters at the start of a piece of work is what this work actually needs — how much is genuinely unknown, how fast the requirements will move, how costly a wrong step is to reverse, how many other parties depend on a stable interface. That is an engineering question, and it has different answers for a consumer-facing feature and for a settlement interface that four downstream systems must integrate against.
The tribal frame makes that question almost impossible to ask out loud. Suggest that a regulated integration programme might benefit from an agreed interface specification before a dozen teams start building against it, and you are not offered a counter-argument; you are offered a diagnosis. You are not agile. So the honest answer goes unspoken, and the team adopts the identity-correct practice instead of the context-correct one.
I have watched this play out, and the shape is always the same. A programme knitting together several established systems adopts two-week iterations and a strict no-design-up-front stance, because that is what a modern team does. For the first few months the burndown looks healthy and the demos are lively. But nine months in, not one increment has been consumable by the downstream systems, because the interface between them was never allowed to be specified in advance — specifying it felt like waterfall. The integration work deferred as “we’ll let it evolve” then arrives all at once, late, as a single painful phase. The team did agile faithfully and produced exactly the outcome that a little sequencing was invented to prevent. The method was not wrong. The refusal to choose it on the merits was.
The opposite of agile is not waterfall. The opposite of both is choosing your method to protect your identity rather than to fit the work in front of you.
The objection, and what remains of it
There is a serious case for the tribalism, and it deserves better than a wave of the hand. The zeal did real work. A decade ago much of this industry was genuinely trapped in heavyweight process that had long since stopped serving anyone — stage gates as ritual, documentation as an end in itself, plans defended long after everyone in the room knew they were wrong. Inertia that entrenched does not yield to balanced position papers. It took a movement, with all a movement’s excess, to break the grip. The purism was the crowbar.
That is true, and it is why the argument had honour in it. But a crowbar is for a door that is stuck, and this door is open. The movement has largely won the argument it set out to win; short iterations and working software are now the mainstream expectation, not the insurgent position. A movement that keeps fighting after it has won does not consolidate the gain — it begins to manufacture the very thing it was formed to oppose: an orthodoxy defended by identity rather than evidence, practices followed because they mark you as one of us. The energy that once broke a bad consensus now enforces a new one.
The useful thing to carry into the new year
So here is the one thing worth taking from the noise. Separate the values from the membership card. The values that started the movement — a bias to working software, tight feedback, respect for the people doing the work, a willingness to change course on evidence — are close to unarguable, and you can hold them completely while still deciding, on a given programme, that some sequencing and some upfront agreement is exactly what respecting the work requires. The more interesting practitioners are already doing this quietly; you can see it in the growing pull towards Kanban and flow thinking, which many teams reach for precisely because it lets them describe how their work actually moves without swearing an oath to anyone.
The mark of a capable practitioner in the year ahead will not be which tribe they belong to. It will be the ability to say “this part wants iteration and this part wants a plan” in a single breath, without flinching at either half of the sentence — and to be trusted, by everyone in the room, to have chosen both for the work rather than for the badge.