How To Design A Perfectly Terrible Design Studio

Thursday, September 24, 2026

Begin by making it standard for all projects to be group projects while keeping the precise number of students unclear. Adding whim to the structure shifts uncertainty away from the studio onto the students. Students who are familiar with the environment will plan their groups long in advance of choosing the studio. This segregates clusters of “high achievers” (as the system previously decided), “low achievers”, and outsiders. Outsiders will then form monocultural groups. This ensures that the environment has already been optimized before they enter. Some early panic can be a good motivational tool.
The groups will inevitably be dysfunctional. Disregard their concerns as “learning experiences for the work environment”. They don’t know that there will never be a work environment with 4–5 equal parties and no designated leader. This can be especially fun to say if you’re an academic architect and not even currently practicing. Evaluate all work as though the full number of people completed the work equally to ensure the groups do not fracture potentially increasing labor for yourself.
Studio consultations should have no fixed order nor time. This encourages students to self-generate a system of writing their group names (not real names, that would break the hierarchy) on a board/page upon entry to the studio. Students familiar with the environment will optimize by sending one member early in the morning to ensure a prime time. This will generate scarcity. Latecomers may not get seen at all. Students should learn early that your time and attention is something to compete for.
To ensure that there is no continuity in meeting with students, treat consultations as interchangeable. Whichever studio leader or assistant is there is simply the person they meet on the day. This successfully guarantees a lack of familiarity between teacher and student. You won’t need to learn their names or even actually know if they can speak the same language as you.
Keep all feedback vague and mundane and always ask for reference pictures instead of sketches or ideas. Some useful examples include “where is the user-flow diagram”, or “describe the tectonics of this project”, or “which products are these composed of?” (point vaguely) or if you’re feeling especially mischievous flatly say “this project won’t work.” This can save you a lot of energy by avoiding involving yourself in any of the decisions they didn’t make in order to select the picture they are showing you. As an added bonus you might get to see some nice pictures. If a student is naive enough to ask how you might tackle this problem, just smile knowingly at them and say you don’t wish to impede the development of their design process. They will never know you’re smiling because you know there was no process to begin with.
Your list of deliverables and deadlines should be strict but keep your directions vague. This will again push the focus towards the product and not the process. When the product is the focus, students will hide or mask knowledge from each other. The more time they spend focusing on their competition, the less time they will spend thinking about your infrequent appearances. Instead of designing and attempting anything, students will spend endless amounts of time trying to guess what you’re thinking without realizing you’re not even thinking about anything at all.
Grade evaluations should be vague, but announce them publicly so there is an element of shaming. For this, you can choose any arbitrary number of categories or ranking systems. Don’t push it too far: grading them out of 3 or 5 will be seen as logical, but grading them out of 3.7 might tip them over their limit (keep in mind that many of them will not have slept prior to this public but necessary ritual).
Don’t ask, don’t tell. Completely ignore the existence of generative AI. Even though it has infiltrated every aspect from pre-design analysis to concept, to development, to deliverables, steadfastly continue to focus only on the product. You do not need to know how it was made. Treating two visually similar pieces of work as being equal can be delightfully demoralising if you know that one group worked manually and the other with generative AI. This ensures that any labor or cognitive offloading that the students are doing is partially a search for a competitive edge. They will hide all learning from each other once they understand that they are being evaluated directly in relation to each other.
If at all possible, reduce all final exams and grading to a 10-minute group presentation or, if you can get away with it, just the deliverables with no students present. Again, keep the grading system as arbitrary and whimsical as possible, as it helps to generate and perpetuate your position of authority.
If you’re concerned about any backlash, don’t.

Trust in the sunk cost fallacy.

By the time they’ve reached the end, they will have worked countless hours. For them to turn around and criticize the studio might make them question the validity of their efforts and perhaps the validity of them as people. Repeat this enough times and this becomes the “studio culture”. You can repeat this process indefinitely.

Best of luck!