Rewriting Harmful Analogies While They’re In Use
This project is difficult, y’know, because it’s like we’re rebuilding the plane while it’s in flight.
For those who’ve spent any time in tech, you’ve likely heard this phrase.
It usually refers a project that seeks to update (or, more often, completely replace) a business-critical, “legacy” piece of software or infrastructure, but with the added challenge of doing so without disrupting whatever’s in production (and, more importantly, its ability to generate revenue).
It is also one of the stupidest phrases I’ve ever heard to describe the actual work we’re trying to do.
“Rebuilding planes in flight” is metaphorically problematic for a number of reasons:
It’s not a clear, well-defined analogy: digging through various references, some thought it meant “a professionally acceptable way to say ‘we have no idea, we’re making it up as we go.’” While I admire that interpretation — it’s probably more often true than not — it’s not one I’d heard before.
“All models are wrong, some are useful,” as they say. But if we don’t agree on what the metaphor even means, it’s can’t serve as a “useful” model to address the challenges these types of projects face.
“Rebuilding the plane while in flight” refers to something we — and I mean all of humanity — does not do.
So, if the goal of using this phrase is to try to create a common context about where we are, what work we need to do, how we’re going to organize and coordinate it, and what effective communication channels need to be built, it’s not useful to analogize these very real problems to an activity humanity isn’t doing, has never done, and is unlikely to ever attempt.
Sure, pontificating on methods to rebuild airships in the sky might provide an interesting afternoon of debate with Theseus, but when you’re responsible for managing actual people, doing actual work, with real budgetary and schedule constraints, on a high profile, existentially-relevant project, a more concrete comparison is sure to serve us better.
Precisely because of this phrase’s pervasiveness, we shrug, throw it around instead, and blindly keep going instead of investigating what other industries actually do in similar situations, so we can learn practices that could apply to our context. There are a number of examples, but if you want an aviation-based one: adding a runway to a major international airport.
An active airport represents a huge, dynamic complex system. Maintaining any level of safety requires establishing, maintaining, and repairing shared joint cognition of that system. A huge amount of planning and coordination goes into the process of mixing heavy construction equipment and materials with multi-million dollar planes operating all around. And this is done with a focus on minimizing disruption to existing operations until the runway is ready for use.
Now does that sound… familiar to any “rebuild the plane in flight”-projects you’ve seen?
An inevitable response to my particular commentary has been “Ok, maybe we’re not rebuilding planes in flight, but… we’re… changing a car’s tires while driving it?”
Yes, I’ve seen that video too.
But no. No, we’re not. And we shouldn’t be.
First, changing the tires on a moving car is not the same as rebuilding and replacing the entire functional system, which is what we’re poorly trying to describe in tech circles using the plane analogy.
Secondly, we don’t analogize our work with this practice because while it can clearly be done, it is incredibly risky. No project manager would purposefully induce that level of risk in such a project, no executive would abide that level of risk, and no board — if they knew that amount of risk to their existing revenue streams, available R&D capital, intellectual property, or engineering teams — would accept that amount of risk, at least in a fiduciary capacity. Stakeholders would excoriate any leader acting so publicly and obviously cavalier.
In other words: yes, humans — with questionable judgment — change tires on moving cars, but we don’t use that analogy because we specifically do not want to suggest appetite for that type of risk (at least, in any way that can be deposed).
So instead, we rely on a ambiguous phrase describing nothing humanity has ever done (nor has any use for or appetite to do), and hope that our contemporaries marvel at the wonders of the massive endeavor we’ve undertaken (organizational change or infrastructure transformation practices from other industries be damned)!
Tech needs to teach itself a new phrase to describe this common type of project we pursue.
Because when a project gets described to me as “We’re rebuilding the plane in flight!” I hear “We not really sure what we’re doing, we haven’t looked at other industries’ operational practices in this area for inspiration and guidance, and we haven’t realized that physics dictates the effort will crash.”
