Agile's not dead. Enterprise Agile is.
I'm seeing more and more posts claiming that agile development and in particular the Scrum process are in some way "dead". And there's something in common with all of them. They don't describe the lean, collaborative people-focused ideas I know and love. They describe the cargo cult rituals, flaccid scrums and lack of commitment inherent in the bastardisation I call Enterprise Agile.
Doing things badly is dead
I'd feel better if the complaints were about transparency of process, or co-located teams collaborating and communicating closely, or reflecting and improving on what you're doing. But they're not. The inevitable complaints are about well-known but actually quite incidental parts of the process:
- Stand ups don't work for us
- Story points don't help us plan detailed release dates and costs
What's really being said here is, "standing in a control tower made of sticks wearing headphones made of coconut shells isn't bringing us any aircraft". Which is correct. You get aircraft by building an aircraft factory. Control towers are merely something you use to help manage aircraft once you have them.
(It's worth pointing out here that some of the people on the frontier of agile methods have become so good at communication they don't have stand ups any more, and so good at creating coherent iterations that they don't use story points to estimate. What they're doing is still agile development, and in many cases still recognisably based in the Scrum framework. If you're pinning your entire adoption on just these rituals, then you're well in the realm of the cargo cult.)
It doesn't help that stand ups and estimations in Enterprise Agile tend to be rickety structures anyway. The stand up stops being about the team telling each other what's going on, and more about the project manager telling everyone what to do. Estimates cease to be loose ideas of difficulty to aid in planning, and start to be hard and fast figures used to obtain budget. That's not a good control tower, it's a bunch of sticks lashed together.
But these things going wrong isn't Agile being dead. It's glomming on to a couple of rituals and failing at even that being dead... which has always been the case, in everything from the waterfall model to the lean startup ethos.
What's not dead
Things that are actually important:
- Servant leadership
- Self-organising teams
- Reflecting upon and improving your process
These things are not going away. Yes, the aims and ideas of the Agile manifesto might be suffocated by firms who don't have the courage or commitment to do it properly. Negative articles based on failures due to that may sully the image of the movement. But it doesn't matter: organisations who truly value those three ideals, no matter what the name of the process or the precise rituals thereof, are going to outmanoeuvre and outcompete organisations who still believe in top-down fiat project management, foster learned helplessness and never see any need for change.
The values espoused in the Agile manifesto will stand in the same way that decades of new ideas and processes have never changed the basic message of The Mythical Man-Month. And these values are what "agile" really means - you can see them permeate even notionally post-agile ideas like Lean Startup. (Lean is mostly agile development done right, combined with some of the useful technical guidance from XP baked in; it only looks so different because one of the first things watered-down Enterprise Agile stripped out was having the courage to release and get feedback often.)
Enterprise Agile is, indeed, dead. But genuine agility is alive, well, and thoroughly outperforming it in every way.
Image by Matt Kimber CC-SA 3.0