I have been accused of perfectionism numerous times throughout my career. I also routinely find the matter to be the unfortunate subject of a number of extremely boring and costly arguments. The truth is that I am doing something slightly different: I have found it monumentally advantageous in the early stages of my projects to consider a spherical cow and then incrementally converge this consideration with reality as I discover the nuances of my charge.

We can say that in any design problem, there is an ensemble of a context — that which exists already and that we do not alter, and a form — that which we seek and work to acquire. We can say that the two fit one another and that the purpose of design is to produce this fit.

It is entirely reasonable to suggest that there can exist the idea of a form that perfectly complements a given context, albeit with the following understanding:

  1. The context is dynamic, and therefore the form is a moving target.
  2. As is the case with dynamic systems, the effort required to approach a perfect fit increases exponentially the closer I get to it.
  3. It is therefore not reasonable to expect to produce a perfect form, I can only approach it.
  4. Because I would need an infinite amount of information about the context to define a perfectly-fitting form.

So why consider it if I can never completely achieve it?

Sketch of a graph depicting curves of design cost, value and proximity to "perfection".

The cost of design grows as a function of the design team's effort over time, making it linear. The value of the designer's work to the client behaves like an S-curve, spending a period below the cost of design before converging and overtaking it, and eventually decaying into diminishing returns. As the design team works, it accrues more information about the context, bringing the form asymptotically closer to a perfect fit.

In my experience, it doesn't cost any more to imagine a perfect fit between form and context than it does to consider one that is good enough. In fact, I find it harder to aim for good enough, because I find myself spending extra effort figuring out what constitutes merely satisfactory performance instead of choosing what I am confident is best.

And a funny thing happens on the way to production: good enough becomes the new perfect — I can't achieve it, just approach it. Given that, if I aim for a perfect fit I ought to get at least good enough. If I aim for good enough, who knows what I'll get. Consider Michelangelo:

The greater danger for most of us lies not in setting our aim too high and falling short; but in setting our aim too low, and achieving our mark.

I think it is entirely safe to say that out there somewhere there is a perfect form for our context, and that we must learn about the context in order to synthesize a fitting form. It is extremely likely that the context will shift naturally, or our perspective will change with new information, and we will have to adjust our vision of the perfectly-fitting form accordingly. When we commit to production, our vision of an ideal form will enable us to stay open to future directions, instead of committing forever to suboptimal solutions.

I anticipate at least three issues with this assertion. The first is undoubtedly a malaise among pragmatists and holders of pocketbooks of carte blanche spending of their precious resources. The others are around when to commit to production, and how to exercise the value of the designed form incrementally. I believe these subjects each deserve their own treatment, to be delivered in due time.