I've delivered projects across the whole of the last decade's argument about how to deliver projects — and it really was an argument, often a heated one. Watching it play out from inside conservative, governed organisations gave me a useful vantage point. Here's how the industry went from tribal warfare over method to a quiet, sensible consensus.
1. 2016: the methodology wars
Ten years ago you were expected to pick a side. Agile was the insurgent, waterfall the establishment, and the debate had the tone of a religious dispute. Agile advocates treated any plan as heresy; traditionalists treated any iteration as chaos. It was exhausting and, in hindsight, mostly beside the point — because the real question was never "which method is correct?" but "which method fits this work?" Very few people were asking that in 2016.
2. The scaling experiments
As agile moved from small teams into large enterprises, the industry tried to scale it — frameworks to coordinate many agile teams under enterprise governance. Some of it helped; a lot of it became heavyweight process wearing agile vocabulary, which produced a wave of disillusionment. The useful residue was an honest realisation: at enterprise scale you cannot escape governance, budgets and coordination. The interesting work is blending them with iteration, not pretending you can abolish them.
3. 2020: remote work forced the issue
Then 2020 settled a lot of arguments by force. Overnight, every team became distributed, and the practices that depended on everyone being in a room stopped working as written. The co-located stand-up, the wall of sticky notes, the energy of a shared space — all of it had to be rebuilt for a screen. What survived the transition was telling: the disciplines that were about structure and communication held up; the ones that were about ritual and proximity struggled. It was an accidental, industry-wide experiment in what actually mattered.
Remote delivery punished ceremony and rewarded clarity. The teams that thrived weren't the most "agile" — they were the ones with the clearest written agreements about who decides what, and by when.
4. PMI's own pivot
The most telling signal came from the establishment itself. On 2 January 2021, PMI overhauled the PMP exam — collapsing the old five process-heavy domains into three (people, process, business environment) and, crucially, weighting the exam so that roughly half of it now covers agile and hybrid approaches rather than pure predictive planning. When the body most associated with traditional project management rebuilds its flagship exam to be half-agile, the war is over. Hybrid won.
5. ITIL grows up
Service management modernised in parallel. ITIL 4, launched in February 2019, was the first major refresh of the framework since 2007, and it deliberately made peace with agile, lean and DevOps rather than standing against them. It reframed rigid processes as flexible "practices" and put value, flow and continual improvement at the centre. ITIL stopped being the thing DevOps teams rolled their eyes at and started being a sensible partner to it.
6. Where it settled
A decade of argument produced a boring, correct answer: projects are hybrid by default until proven otherwise. You use predictive structure where the work is well-understood and the stakes are fixed, and iteration where the work is uncertain and discovery is the point — usually both, on the same program, at the same time. The mature practitioner isn't loyal to a method; they're fluent in choosing. And now AI is entering delivery as the next chapter, taking the administration so we can spend more time on the judgement. Through all of it, the two things that never stopped mattering were governance and honest iteration. If you're trying to get that blend right in an organisation that still argues about method, I've probably been there.