Designers — at least the good ones — have a rare superpower: they can leap several levels of abstraction in a single bound. Among other things, this allows them to look at the form of a thing (a building, a kettle, a website) and get a sense of its constituent parts, the relationships between those parts, how those relationships help serve particular functions, and what those parts, relationships, and functions say about the goals of the entity that commissioned the thing — all without getting hung up on its “look and feel.” In other words, good designers can look beyond the tangible forms of things to the conceptual models they manifest. Some designers can even map out these models in ways that make sense to the rest of us.

Working with models is a key skill — perhaps the key skill — in 21st Century design. Today’s most important design challenges deal with complex, evolving systems. The parts of these systems we see and interact with (such as user interfaces) are only surface manifestations of deeper structures. It’s essential to understand the connection between these structures, the forces that call them forth, and the user interfaces that manifest them. You can’t skillfully intervene by acting solely on the surface.

However, many designers (and most stakeholders) want to work with screens, not models. Screens are things they can test and critique. Models? Not so much. It takes practice to see the tangible forms latent in an abstract diagram. Most of us lack the patience to acquire the practice. We’re drawn to screens because we can draw screens; they’re familiar, things we deal with every day. Models, on the other hand, are abstractions. They can be ambiguous and subjective and unfamiliar. This makes them hard to communicate. How do we begin to draw such a thing?

And yet, draw them we must. Only by understanding models can we effectively deal with fitness-to-purpose and second- and third-order effects, and thereby ensure design directions are strategically and ethically sound. Of course, this doesn’t mean we won’t work on screens at all. As I said, we can’t test models without manifesting them as tangible artifacts. But these artifacts must be rooted in a clear understanding of the underlying models, not the other way around. The ability to jump back-and-forth between models and their expression as UI requires training and practice. It’s an essential skill for today’s designers, and one I’m increasingly focused on learning and teaching.