Deconstructing fidelity

March 10, 2015

Fidelity is the degree of exactness to which a model reproduces the real thing.

How much fidelity is enough? The greater the fidelity, the more likely it is that someone will understand your design intent. The more real something looks and feels, the more likely it is that you’ll receive actionable feedback to validate your design or inform your next iteration. Theoretically, one could gradually increase the fidelity of a model until it’s indistinguishable from the actual product. However, that’s not a practical approach in most situations.

In the field of UX, the term low fidelity is often applied to site maps, screen flow diagrams, wireframes, and paper prototypes. These models help designers explore and gather feedback about how information within a digital product might be structured and how that product might respond to interactions from users and the system at large. While some aspects of a product’s form may begin to take shape with low-fidelity models, it is not their primary intent.

The greater the fidelity, the more likely it is that someone will understand your design intent.

The term high fidelity is more commonly used in conjunction with UX models that articulate a product’s form or aesthetics. Screen layouts, style guides, and other visual design specification documents fit into this bucket. This suggests that low-fidelity deliverables address the product’s function while high-fidelity deliverables focus on form.

However, many of the UX deliverables outlined above can actually be executed at various levels of fidelity. Let’s use wireframes to illustrate the point. Loose, hand-sketched wireframes can quickly but completely capture page-level content hierarchy and navigational elements. Many people would call this low fidelity. By contrast, wireframes with pixel precision and brand-appropriate elements such as color and typography may communicate the visual layout of information. Most people would call these high fidelity. But what if these high-fidelity wireframes lack any definition around what content—text or images—should be included on the page? And what if it’s not clear where a particular page exists inside of the website’s overall structure or how users will navigate to other pages? This doesn’t sound like a model that accurately reproduces the final product, so it can’t be high fidelity, can it?

Read the full article in UX Magazine.