Levels of Design

Most work that disappoints was never allowed to succeed. An organization feels a problem, positioning that will not hold, trust that keeps leaking, customers who do not behave as hoped, and it commissions an object to fix it. A better brochure. A new logo. A refreshed page. A prettier booth. The object arrives, polished and approved, and the problem is still there, because the problem was never inside the object. This is the most common and least examined failure in creative work. Organizations routinely ask isolated artifacts to solve systemic or experiential problems, and then conclude the work was weak when the work was only small.

The way out of that pattern is to see design as operating at four widening levels of responsibility. The levels are not a scale of beauty and not a ranking of talent. They do not measure how good the work is. They measure the breadth of conditions the designer accepts responsibility for, and that breadth, not the craft, is what determines how much of a problem the work can reach.

Craft and scope are independent.

Before the levels, one guardrail, because without it the model gets misread as a hierarchy. A higher level is not a better piece of work. It is a wider one. An artifact designed at the first level can be excellent on its own terms, and an experience attempted at the fourth can fail, more expensively, because the stakes rose with the scope. Craft is how well the work is made. Scope is how much of the problem it was made responsible for. They are independent, and the discipline is not to climb as high as possible. It is to match the level of the response to the level of the problem. A systemic problem handed an artifact stays unsolved. An artifact problem handed a system is overbuilt. The failure is the mismatch, in either direction.

Take a single assignment through all four levels, a company preparing for a trade show. The raw material is the same at every level: a booth, a backdrop, banners, brochures, a demonstration, the people staffing it, the giveaways, the way leads are captured and the follow up afterward. What changes is how much of it the design is responsible for.

Level one, the artifact.

At the first level the requested object is treated on its own. The assignment begins and ends with the deliverable. Make the brochure better, refresh the banner, redesign the giveaway. Each piece can be made attractive and each can be approved, and none of them are asked to account for anything beyond themselves. The weakness of this level is not quality, it is context blindness. At the trade show, the booth is handsomer and the brochure is cleaner and the encounter is exactly as scattered as it was, because a set of improved objects is still a set of objects.

Level two, functional form.

At the second level each object is shaped around its purpose. The banner is designed for the distance it is read from and the two seconds it gets. The brochure is built for what a visitor does with it after they walk away. The demonstration is sequenced for how attention actually moves. Each touchpoint becomes good at its own job. The limit of this level is that a collection of individually effective pieces can still add up to a fragmented whole. Every object works, and the encounter still does not cohere, because nothing was made responsible for the relationships between them.

Level three, the system.

At the third level the designer accepts responsibility for how the parts relate, which means accepting responsibility for the operating structure that produces them, not only for the touchpoints a visitor sees. Message hierarchy, a shared visual language, the roles each piece plays, the sequence a visitor moves through, the repetitions that make an idea stick, and the decisions and defaults upstream that generate all of these. The booth, the banners, the demonstration and the conversation stop being separate objects and start reinforcing one idea, so that a visitor who saw only part of it still leaves with the whole. System creates coherence. It produces clarity, orientation, consistency and comprehension. What the system does not yet guarantee is that anyone will care.

Level four, the experience.

At the fourth level the coherent system becomes something lived, the coherence of meaning a person feels and carries over time rather than a staged event that ends when they leave the booth. The designer takes responsibility for anticipation and pacing, for participation and sensory cues, for surprise and resolution and for what a person remembers a week later. Experience creates resonance. It does not replace the levels beneath it. It contains them. An experience is coherent function and a coherent system made to matter, and it cannot be reached by skipping the levels under it, because there is no resonance in a system that is not yet clear and no clarity in touchpoints that do not yet work. When resonance runs deep enough it can change the person who felt it, but that transformation is a consequence of the experience, not a fifth level above it, because no designer can be handed responsibility for who another person becomes.

One caution about how this coherence is made. A higher level of responsibility is not a license to dictate the whole from above. Resonance often assembles from many aligned local acts, and the designer's work at the upper levels is to create the conditions in which the parts cohere, not to impose a coherence onto them from the outside. The responsibility widens. The control does not widen with it.

The model in a longer conversation.

None of the four levels is new by itself. A whole is built from the relationships among its parts and can be alive or dead at any scale, which is the lesson of Christopher Alexander. Some problems are symptoms of other problems, which is why an artifact cannot solve them, the finding of Rittel and Webber. What is the contribution here is the combination: four widening levels applied to the same assignment, sorted on the axis of accepted responsibility rather than domain or craft or organizational maturity, with craft held independent of scope, and with the diagnostic that names the recurring failure, a problem at one level handed an intervention authorized only at a lower one.

Responsibility is not authority.

There is a distinction inside that last word the model has to make, because the whole diagnosis turns on it. Responsibility is what a designer accepts, the breadth of the problem they agree to answer for. Authority is what an organization grants, the range they are actually permitted to change. The two come apart constantly, and the gap between them is where the better brochure comes from. A designer can accept responsibility for the experience and be granted authority only over the artifact, and the result is a beautiful object that cannot reach the problem, not because the designer misjudged the level but because the permission stopped below it.

The strategic implication.

The model resolves into two questions a leader can ask about any brief. At what level does the problem actually exist. And at what level are we authorizing the work to intervene. When those two answers do not match, the money is already lost, whatever the quality of what gets made. The seniority of design is not measured by the beauty of its artifacts. It is measured by the level of responsibility the work is trusted to carry, and trust here means authority as much as skill. Design cannot resolve a problem beyond the level of authority it has been granted, however wide the responsibility it is willing to accept, so the most consequential decision about any piece of work is made before the work begins.