Memorisely
UX Design

How much fidelity your prototype needs

Matching the prototype to the question you are asking.

Zander Whitehurst

9 min read

One of the most common questions I hear from students is how polished a prototype needs to be. The honest answer is that it depends entirely on what you are trying to learn. A prototype is a tool for answering a question, and the right level of fidelity is the lowest one that answers it well.

I have wasted more hours than I would like to admit building pixel perfect prototypes to test ideas that turned out to be wrong. There is a special kind of sadness in watching a participant ignore a feature you spent two days animating. These days I start rough and only add polish when the question demands it.

Start with the question

Before you open any tool, write down what you want to learn. Some examples:

  • Do people understand the overall structure of this flow?
  • Can people find the settings they need?
  • Does this layout make the key information clear?
  • Does this interaction feel natural?
  • Will people trust this checkout enough to pay?

Each of these questions needs a different kind of prototype. Structure can be tested with paper sketches. Trust and interaction usually need something much closer to the real thing.

Each question points to a different level of fidelity, so start with the question.

Low fidelity for ideas and structure

Low fidelity prototypes are sketches, paper screens or simple grey boxes. They are quick to make, cheap to throw away and great for exploring lots of ideas.

They work well for:

  • Testing whether a flow makes sense
  • Comparing different layout approaches
  • Getting early feedback before committing to a direction
  • Workshops where the whole team contributes ideas

A nice side effect is that people feel more comfortable criticising something that looks rough. Nobody worries about hurting your feelings over a sketch on a sticky note.

Mid fidelity for layout and content

Mid fidelity prototypes add real structure, realistic content and basic interactions, but keep visual design simple. Think greyscale screens with proper copy and clickable links between them.

This is often the sweet spot for usability testing. People can complete realistic tasks, and you can test navigation, content hierarchy and labels without getting distracted by colours and imagery.

Use real content wherever you can. Lorem ipsum hides problems. A product title that wraps onto three lines or a price that does not fit is exactly the kind of thing you want to discover early.

Real content exposes wrapping titles and long prices that placeholder text hides.

High fidelity for detail and feel

High fidelity prototypes look and behave much like the finished product. They include final visuals, realistic interactions and sometimes animation.

They are worth the effort when you need to test:

  • Microinteractions and animation
  • Visual hierarchy and brand perception
  • Trust in sensitive flows like payments
  • Final details before handing over to engineering

They are also useful for stakeholder presentations, where people often struggle to imagine how rough sketches will eventually look.

Mix fidelity where it helps

Fidelity does not have to be the same everywhere. You might build one screen in high detail because that is where the question lives, and keep everything around it rough.

For example, if you are testing a new checkout button, the product pages leading up to it can be simple. Spend your time where the learning happens.

Watch out for false feedback

Fidelity affects the feedback you get. Polished prototypes can make people focus on colours and imagery rather than whether the flow works. Very rough prototypes can make it hard for people to imagine using the product.

I once showed a beautifully finished prototype to a group of stakeholders hoping to discuss the navigation. We spent forty minutes debating the shade of a button. Lesson learned. If you want feedback on structure, show structure.

Consider coded prototypes

Sometimes the best prototype is real code. If you are testing something that depends on real data, performance or complex interactions, a coded prototype can answer questions design tools simply cannot.

With modern AI tools, building quick coded prototypes is easier than ever. Just remember that they are still prototypes. Keep them focused on the question and be ready to throw them away.

A focused prompt spends the detail on the screen where the question lives.

Keep it disposable

Whatever fidelity you choose, treat prototypes as temporary. Their job is to help you learn. If you become too attached, you start defending the prototype rather than learning from it.

Build only what the test needs. If participants will never click a link, it does not need to work. If a screen will never be seen, it does not need to exist.

Start with your next question

Before your next prototype, write down the single question you need to answer. Then choose the lowest fidelity that can answer it honestly. Sketch first, test early and add polish only when the question demands it. You will learn faster, and you will waste far fewer afternoons animating buttons nobody clicks.

Keep reading

View all
View all

Optimize your AI for better experiences, not AI slop.

View all

AI Prototyping Skills

Prototype with AI using clean UI patterns and guardrails so flows feel product-grade, not generic template slop.

$99

AI UI Motion Skills

Motion and micro-interaction patterns that make AI-built UI feel designed, polished, and trustworthy in real products.

$99

Fix AI Slop Skills

Spot AI slop early and fix type, spacing, and components so your product looks intentional, not like a template.

$99

AI Responsive Design Skills

Responsive layout patterns for rails, sidebars, and breakpoints so AI-generated UI adapts cleanly from narrow to wide.

$99

AI Dark Mode Skills

Dark theme tokens, contrast, and surfaces that keep AI-generated UI readable and product-grade, not flat gray on black.

$99

AI Color Token Skills

Neutral and primary ramps with clear steps so AI palettes stay systematic, not a random set of accent hex codes.

$99
View all