process & strategy
Prototype & fidelity
What is a prototype and what does fidelity mean?
A prototype is a working model of a design made to be tested, and fidelity is how closely it resembles the finished product. Low-fidelity prototypes (paper sketches, clickable wireframes) are rough, fast and cheap; high-fidelity ones (polished, interactive, near-real) look and behave like the real thing. Choosing the right fidelity for the question you're asking is the whole skill.
Also known as: prototype and fidelity, prototype fidelity, lo-fi vs hi-fi
Prefer to watch?
Watch the recap 1:14
The demo
Drag the fidelity dial and watch the same screen go from a paper sketch to a polished, interactive design. Notice what each level is good for - and why higher isn't always better.
What this demo shows (text version)
A fidelity dial takes the same screen through four levels: a rough paper sketch, a grey-box wireframe, a styled mock-up, and a polished interactive design. As you slide up, the screen looks and behaves more like the finished product. Each level is suited to a different question - the rough end for testing concept and structure early and cheaply, the polished end for testing visual design, detailed interaction and realistic usability later.
That's prototyping and fidelity: a prototype is a testable model, and fidelity is how closely it resembles the real thing. Match fidelity to the question and the stage - stay low and disposable early, raise it as decisions get specific. Going high too soon wastes effort and dampens honest feedback; staying low too long never tests the real feel.
Match the fidelity to the question, and to how early you are. Low-fidelity prototypes are fast and cheap and people happily critique them, so they're ideal early for testing concepts, structure and flow - and you won't grieve over throwing them away. High-fidelity prototypes feel real, so they're right later for testing visual design, detailed interactions, micro-copy and realistic usability, and for buy-in. The two traps: going hi-fi too early (slow, premature polish, and reluctance to change something that looks "done" - plus testers nitpick the visuals instead of the idea), and staying lo-fi too long (you never test the real feel). Start rough, raise fidelity as questions get more specific and confidence grows.
Slide from paper to polished and you can feel what each fidelity is for: the rough version is fast and invites big-picture critique, while the realistic one tests the actual look and feel. There's no single "right" fidelity - only the right one for the question you're asking and how far along you are. Go high too early and you waste effort and chill honest feedback; stay low too long and you never test the real thing.
Fidelity is a spectrum across several dimensions - visual (sketch → wireframe → full design), content (placeholder → real), and interactivity (static → clickable → fully functional). Low-fidelity prototypes (paper, lo-fi wireframes) are quick to make and change, cheap to discard, and signal "unfinished", which invites bold structural feedback. High-fidelity prototypes are realistic and interactive, better for evaluating aesthetics, detailed interaction and true usability, and more persuasive for stakeholders - but slower and costlier.
The classic mistake is jumping to high fidelity too soon. It wastes effort polishing ideas that may change, makes people reluctant to alter something that looks finished (a sunk-cost and "looks done" bias), and shifts test feedback toward surface details when you still need to validate concept and structure. Conversely, only ever testing lo-fi means never checking the real visual and interactive experience, where different problems surface.
Choose fidelity by the question and stage: early, to test whether the concept and flow work, stay low-fidelity and disposable; later, to test visual design, motion, copy and realistic usability, raise fidelity. Be willing to throw prototypes away (they're for learning, not keeping), use the lowest fidelity that genuinely answers your current question, and increase it as uncertainty narrows and decisions get more detailed.