Prototype
A prototype is an interactive model of a product that lets you test and validate design concepts before development begins. It can be a paper sketch, a clickable wireframe, or a fully interactive simulation that closely resembles the final product. The goal is to uncover usability problems, validate assumptions with real users, and minimize the cost of changes later on.
Prototyping is one of the most effective ways to reduce risk and cost when building digital products. Large-scale studies back this up:
- Fixing a bug after release costs up to 100× more than catching it at the design stage. A NIST study (National Institute of Standards and Technology, 2002) analyzed the impact of inadequate software testing in the US and put the annual cost at $59.5 billion, of which $22.2 billion could be saved by catching defects earlier (NIST – The Economic Impacts of Inadequate Infrastructure for Software Testing, 2002)
- Companies that share prototypes with users early grow significantly faster. McKinsey analyzed 300 publicly listed companies across industries over 5 years (2 million financial data points, 110,000+ design actions). Companies with the strongest design practices achieved 32 percentage points higher revenue growth (McKinsey – The Business Value of Design, 2018)
- Iterative prototyping and testing improves usability by a median of 165%. Based on 4 case studies, Nielsen Norman Group found that each prototype iteration tested with users delivers a median usability improvement of 38% (NNGroup – Iterative Design, 1993)
- 5 users are enough to uncover 85% of problems. Jakob Nielsen and Tom Landauer published a mathematical model (ACM INTERCHI’93) showing that a qualitative prototype test with 5 participants reveals most key usability problems (NNGroup – Why You Only Need to Test with 5 Users, 2000)
FAQ
Which type of prototype should you choose—low-fidelity or high-fidelity?
It depends on the project phase. Research by Virzi, Sokolov, and Karis (1996) showed that paper (low-fidelity) prototypes uncover a comparable number of usability problems as fully functional high-fidelity prototypes. NNGroup recommends starting with low-fidelity prototypes, which allow fast iteration—one cycle takes about a day. High-fidelity prototypes are better suited to later stages (NNGroup – UX Prototype Fidelity).
How many prototype iterations do you need?
NNGroup recommends at least 3 versions (2 iterations), ideally 5–10 iterations. Their research shows a median usability improvement of 38% per iteration. Mozilla, for example, used paper prototypes to quickly test and iterate on a redesign of its Support page, saving significant resources (NNGroup – Mozilla Paper Prototype Case Study).
Is prototyping worth it for smaller projects too?
Yes. According to the NIST study, the cost of fixing defects grows exponentially with each development phase—from 1× at the design stage to 6.5× during implementation and 15× during testing. Even a simple paper prototype tested with 5 users can reveal critical problems before they turn into costly development mistakes (NIST, 2002).
Sources:
- NIST – The Economic Impacts of Inadequate Infrastructure for Software Testing (2002) – a study of the economic impact of inadequate software testing, $59.5 billion a year
- McKinsey – The Business Value of Design (2018) – a five-year study of 300 publicly listed companies, 2M financial data points
- Nielsen Norman Group – Iterative Design (1993) – research on iterative design, 4 case studies, median improvement of 165%
- Nielsen Norman Group – Why You Only Need to Test with 5 Users (2000) – the Nielsen & Landauer mathematical model: 5 users uncover 85% of problems
- Nielsen Norman Group – UX Prototype Fidelity – a comparison of low-fidelity and high-fidelity prototypes
%201.webp)