Pick the research method and the confidence you need. Returns a recommended sample size based on the published research literature — Nielsen for usability testing, Sauro for benchmarking, standard proportion maths for surveys, MDE formulas for A/B tests.
Nielsen (2000) found 5 users surface roughly 85% of major usability issues; 8 surface 95%. Diminishing returns beyond 8.
There is no universal "right" sample size for UX research. The right size depends on your method, the kind of claim you want to make, and how costly it would be to be wrong. The literature offers established starting points this calculator surfaces.
Qualitative usability testing: Nielsen's 1993 and 2000 papers showed 5 participants surface roughly 85% of issues when each user has a 31% chance of hitting any given problem. 8 users surface ~95%. More than 15 produces diminishing returns.
Benchmark studies: SUS or quant usability metrics require larger samples to detect small differences. For a ±5-point margin on SUS (SD = 21) at 95% confidence, n ≈ 68.
Surveys: Standard proportion maths. For ±5% margin at 95% confidence on an infinite population, n = 384.
A/B testing: Required sample scales with baseline rate and MDE. Detecting 1% relative lift at 5% baseline needs ≈ 160,000 per group at 80% power; 20% needs ≈ 1,500.
Related: UX method finder, SUS calculator, interview sampling guidance.