How Many Test Respondents Do I Need?

“How many test respondents do we need for a good user test?”

We get asked this question all of the time. Noted User Researcher Jakob Nielsen already shares a great answer to this question on his blog¹, so we will just summarize his thoughts here.

1. Basic Usability Study

New task validately

  • Definition: A basic test of whether a respondent can complete a specific task.
  • # of respondents needed: 5
  • This could be higher if you have vastly different customer segments. However, you do not need to multiply the number of segments by 5. Multiplying the customer segments by 3 should be fine unless you see something very different in the results. For example, if you have 3 different customer segments, run 9 user tests (3 for each segment) and look for differences.

2. Quantitative Metrics (eg Utility or Satisfaction)

 

Zrzut ekranu 2014-10-06 13.22.06

  • Definition: Usability metrics such as Utility (does the customer need this feature) and Satisfaction (How pleasant is it to use the design?).
  • # of respondents needed: 20

3. Advanced Analytics (eg Heatmaps)

 

advanced

  • Definition: Looking for high usage patterns across user tests.
  • # of respondents needed: 39

Conclusion: A statistically significant user test requires 5 respondents for basic task usability, 20+ respondents for quantitative metrics and 39+ respondents for heatmaps. Tweet this


¹Source: http://www.nngroup.com/articles/how-many-test-users/  and http://www.nngroup.com/articles/quantitative-studies-how-many-users/ and http://www.nngroup.com/articles/usability-101-introduction-to-usability/ and http://www.userfocus.co.uk/articles/2-kinds-of-usability-test.html


No Comments

Leave a Comment

Only registerd members can post a comment , Login / Register