The Worst Interface
A working interface you make deliberately worse on purpose, with the receipts for exactly why
Due: 29 March 2027 · Weight: 25%
The brief#
Take a simple, ordinary interaction and redesign it to be worse, on purpose, without breaking it.
Checkout, sign-up, booking, a subscription, a settings page, a survey, a login, a cancellation — pick something ordinary enough that everyone already knows how it’s supposed to work. Then decide, before you touch anything, exactly one behaviour you want to produce: users picking the pricier option by accident, or cancelling taking five times longer than subscribing did. Every change afterwards should serve that one goal — poor hierarchy, a misleading default, an extra step, wording that’s technically true and practically confusing. None of it counts if the interface stops working; deleting the submit button isn’t design, it’s damage.
What you submit#
A working prototype of your redesign, the original (or a neutral baseline) so a marker can see exactly what changed, and 600–800 words explaining your target behaviour, what you changed, why you expect each change to work, and which course concepts you’re leaning on. Testing it properly comes next assessment — for now, the argument just has to be plausible.
The spec
The brief poses the problem and leaves room for a response. The spec is the fixed part: what a marker checks. Some lines can be checked mechanically; the rest call for judgement.
- the interaction you redesign is an ordinary one (checkout, booking, subscription, cancellation, or similar) and still works end to end
- you name one specific hostile goal before you change anything, and every design decision serves that goal
- the 600–800 word write-up explains what you changed, why you expect each change to work, and which course concepts back it up
- the original or a neutral baseline is included so a marker can compare it directly against your redesign
How it is marked
| Criterion | Weight |
|---|---|
| Clarity of original vs redesign | 25% |
| Explanation of the mechanism | 45% |
| Plausibility of the intended effect | 30% |