It’s pop quiz time. The quiz has one question: Which application development methodology is gaining the most popularity?
If you answered “Agile,” Blaaaaaaat! Wrong answer bucko.
If you tried to demonstrate your more in-depth knowledge of the app dev landscape by answering “Scrum,” Blaaaaaaat! Nice try, but wrongo.
Test Driven Development (TDD) or one of its variants, Acceptance Test Driven Development (ATDD) or Behavior Driven Development (BDD), you’re just showing off. But Blaaaaaaat! TDD might be a technician’s paradise, and for that matter it might be a very good idea, but it isn’t what’s gaining the most acceptance.
Want one more guess? I’ll give you a hint: What do you get when you combine a change in process with the same old attitudes?
Now you’ve got it. The app dev methodology that’s sweeping the world is (drumroll) … Scrummerfall!
Scrummerfall (not an original term) is what you get when you stitch a Waterfall head onto a Scrum body. It’s what happens when you do want iteration and incrementalism, but for one reason or another want developers to do nothing but write code to specifications — you have no interest in their understanding the context, business purpose, or user predilections.
To be fair (an excruciating exercise but I’ll try) there are good reasons for going this route. In particular, if you’re willing to trade off Agile’s high levels of team engagement, enthusiasm and commitment for the large savings in raw labor rates you get from sending work offshore, Scrummerfall might be the right choice for you.
This is especially true in organizations that consider financial measures to be the only measures that matter, because from a purely financial perspective, it’s iteration and incrementalism that drain most of the risk from Waterfall’s combination of long-range planning and short-range planning accuracy. If all you do is wait as long as possible before making design decisions, that by itself will increase your project success rate.
What do you have to lose?
Quite a lot, as it happens. The problem is, what you lose by settling for Scrummerfall is much harder to quantify, because with Scrummerfall, what you keep is form but what you lose is essence.
Another way of saying it: Scrummerfall is an excellent example of what goes wrong when you mistake a business practice for a business processes. For the difference, see “Is it a Process, or just a process?” KJR 5/17/1999, although when I wrote it I used lowercase “process” where “practice” is now my preferred vocabulary.
In any event, with a true process, following the right steps in the right order gets you to the desired result. They’re repeatable and all that. The assembly line is your model.
That isn’t true with a practice. Following the right steps in the right order is just the ante that lets you play the game.
With a process, the steps are the essence. With a practice, they’re separate, and following the steps while losing the essence means the steps generally degenerate into nothing more than a bunch of check boxes people follow because they have to, not because they add any value to the proceedings.
And so to the differences between Agile and Scrummerfall. Start with the basics: Writing user stories and estimating them by assigning story points. (If you’re unfamiliar with these terms, user stories are the Agile equivalent of requirements; story points are vaguely equivalent to function points only they’re heuristic, not algorithmic.)
With Agile, the whole team writes the stories and assigns the story points, which means the whole team understands all of the requirements and commits to their estimated difficulty.
With Scrummerfall, business analysts write the stories and assign the story points. Team members only understand the user stories assigned to them for development, and instead of assigning story points … estimates of relative difficulty … the business analysts estimate the time that should be needed for development.
Anyone who’s been on either side of any exercise in delegation knows the difference between me telling you how much time you should need to achieve your assignment and the you telling me how much time you’ll need.
What’s the financial impact of the difference? We can envision what the research needed to answer a question like this might look like, but I certainly can’t imagine who might pay for the research, let alone any business leaders making decisions based on this research.
There’s one more piece of this puzzle to mention right now, and that’s the core model for The Cognitive Enterprise — that cognitive enterprises replace the old people/process/technology model with customers, communities, and capabilities.
With true Agile, developers and business stakeholders form a community.
With Scrummerfall, they’re just cogs in a machine.