Don't go chasing waterfalls

 

Some of you are familiar with my silly little rants against abused or overused words in project management.  Here's a common and accepted example that I resist: "Waterfall." 

This term is used incessantly online, especially in conversations (arguments) about "Agile vs Waterfall." The term 'waterfall' is used just once in the PMBOK 5th edition, and it's in exactly that context.  It's used multiple times in the PMBOK 6th edition, which could indicate that it has gained acceptance.

My objection to the "Agile vs Waterfall" debate is that "agile" is a set of cultural values, not a project management model.  People who engage in this argument might as well compare apples to screwdrivers.

But my hesitation with the term "waterfall" is that I've never seen a real project that looked like a waterfall.  Have any of you, on the job, ever seen a perfect, serial network diagram that looked like this?

Except for some academic exercises in college, I've never seen a network diagram or Gantt chart that looked so simple and perfect.  I would be immediately suspicious of a plan that looked like this.  Either this project so simple that it has no need for a project manager, or it's in desperate need of a skilled PM to challenge the over-simplified plan!

In the real world, project managers look for ways to accomplish the most work possible in the shortest amount of time.  This is why we're so heavily involved in the planning stages of projects.  We challenge discretionary dependencies and mitigate risks associated with external dependencies.  We seek opportunities for parallel activities so that one member of our project team isn't idly waiting for another.  Our network diagrams don't look like waterfalls, they look like drainage basins, where various streams connect and flow into a final body of water. 

Trivia: Many people attribute the word "Waterfall" as a project management term to Winston W. Royce in his 1970 paper ""Managing the development of large software systems."  But Royce neither used the term, nor did he advocate for it.  Instead, he pointed out flaws in a sequential plan, stating that it is "risky and invites failure."  His diagram also flows both directions, iterating forwards and backwards.  If that still sounds like a waterfall, you should see his final diagram! 

When you're ready to use the word "waterfall," try using alternate (dare I say "old") terms like "plan-driven" or "predictive." I've noticed that this immediately changes the conversation with "waterfall vs agile." After all, a fully Agile organization can use a predictive project plan, given that the conditions support that decision (i.e. the scope is stable, the steps are well-established or repeatable). What agilists in these arguments resent is the ignorance of Agile values. When our most cherished project artifact is just a pretty waterfall, we value a tool over people, and plans over results. 

I promise not to be too big of a jerk about this; I know what people mean when they say "waterfall."  But sometimes, the wrong word choice complicates a conversation or misrepresents the value of the thing it describes.  This is why the PMBOK states "a common vocabulary is an essential element of a professional discipline," and why PMI settled on the term "predictive" and excluded "waterfall" from the project management glossary.  Does is work in some casual contexts?  Yes, but "waterfall" is too often a term that causes misunderstanding, and we'd often be better off avoiding it.

/silly rant