Skip to main content

User Research Sample Size

Calculate sample size for usability tests, surveys, and interviews. Enter values for instant results with step-by-step formulas.

Share this calculator

X Facebook LinkedIn

Worked Examples

Example 1: Mobile App Usability Study

Problem: Planning a usability study for a mobile banking app redesign with 3 user segments: daily users, occasional users, and new customers.

Solution: Research Type: Usability Testing\nUser Segments: 3\n\nNielsen Norman guideline:\n5 users find ~85% of usability issues\n8 users recommended for higher coverage\n\nPer-segment calculation:\n5 minimum × 3 segments = 15 total minimum\n8 recommended × 3 segments = 24 total recommended\n\nOver-recruitment (30%):\n24 × 1.3 = 31 recruits\n\nCost estimate:\n24 × $150 incentive = $3,600\nModeration time: 24 × 1.5 hrs = 36 hours\n\nTimeline:\n1 week recruitment\n2 weeks testing (4 per day)\n1 week analysis

Result: 24 participants recommended (8 per segment) | $3,600 incentive budget | 36 hours moderation

Example 2: Customer Satisfaction Survey

Problem: Designing a CSAT survey for 50,000 active users. Need 95% confidence with ±5% margin of error.

Solution: Survey Parameters:\nPopulation: 50,000 users\nConfidence: 95% (z = 1.96)\nMargin of Error: ±5%\nExpected proportion: 50% (conservative)\n\nSample Size Formula:\nn = (z² × p × (1-p)) / e²\nn = (1.96² × 0.5 × 0.5) / 0.05²\nn = (3.84 × 0.25) / 0.0025\nn = 384.16\n\nFinite population correction:\nn_adj = n / (1 + (n-1)/N)\nn_adj = 384 / (1 + 383/50000)\nn_adj = 381\n\nWith 30% response rate:\nNeed to send: 381 / 0.30 = 1,270 invites\n\nCost (at $5/response):\n381 × $5 = $1,905

Result: 381 responses needed | Send 1,270 invites (30% response) | $1,905 panel cost

Example 3: Information Architecture Card Sort

Problem: Conducting an open card sort to reorganize a corporate intranet with 200 items. Single user segment (employees).

Solution: Research Type: Open Card Sort\nContent items: 200\nUser segments: 1 (employees)\n\nCard Sort Guidelines:\nOpen sorts need 20-30 participants\n200 items is large - lean toward higher end\n\nRecommendation:\nMinimum: 25 participants\nRecommended: 35 participants\n\nAnalysis considerations:\n- Use similarity matrix\n- Look for 70%+ agreement on groupings\n- Consider hybrid approach: open sort with subset, then closed sort to validate\n\nAlternative approach:\n1. Open sort with 30 users on 60 key items\n2. Closed sort with 50 users on all 200 items\n\nCost estimate:\n35 × $75 incentive = $2,625\nTool license: ~$200/month\nAnalysis: 20+ hours\n\nTotal: ~$3,000

Result: 35 participants for open sort | Consider hybrid approach for 200 items | ~$3,000 total

Frequently Asked Questions

What is the Nielsen Norman 5-user rule?

Jakob Nielsen's research found that 5 users uncover about 85% of usability problems. Beyond that, you see diminishing returns—the same issues repeat. However, this applies to single user segments; multiple segments need 5+ users each.

What sample size do I need for a survey?

Survey sample size depends on: population size, confidence level (typically 95%), margin of error (typically 5%), and expected response distribution. For large populations, ~385 responses give 95% confidence ±5%. Smaller populations need correction formulas.

How do A/B test sample sizes work?

A/B tests need enough users to detect a meaningful difference (minimum detectable effect). Sample depends on baseline conversion rate, desired lift detection, confidence level (95%), and statistical power (80%). Calculator tools like Evan Miller's are essential.

How does segmentation affect sample size?

Each segment needs adequate representation. If testing 3 personas, you need 5+ usability participants PER persona (15 total), not 5 total. Surveys need sufficient responses per segment for meaningful sub-group analysis.

What's the cost of different research methods?

Ballpark costs: Usability test $100-300/participant (incentive + moderation), interviews $150-300/participant, surveys $2-20/response (panel cost), card sorts $50-100/participant (tool + incentive).

When should I use larger sample sizes?

Use larger samples when: testing multiple segments, seeking statistical significance (not just insights), stakeholders need quantitative confidence, comparing between groups, or when decisions are high-stakes and reversibility is low.

References