This blog post is about an article I read and I want to share with you some of the main points of this article.
That article has a title "Why You Only need to Test with 5 Users" and the writer is Jakob Nelsen.
From the research of Toom Landauer exported the following chart:
From the above chart we can export information:
References:
Nelsen, J. (2000). Why You Only Need to Test with 5 Users. [Internet]. Available at: <https://www.nngroup.com/articles/why-you-only-need-to-test-with-5-users/>. [Accessed on:14/3/2016].
That article has a title "Why You Only need to Test with 5 Users" and the writer is Jakob Nelsen.
From the research of Toom Landauer exported the following chart:
From the above chart we can export information:
- Zero users => gives zero insights
- Single test user => you know 1/3 of all these you need to know
- Second user => you discover same things as from the first user + something new
- Third user => many things that already observed + small amount of new data
- More and more users => you learn less and less
- After fifth user => wasting time by observing the same findings
Also Jakob Nelsen mentioned in his article that
- There is a need for multiple tests in order to achieve usability
- Engineering
- improve the design
- Not only document weaknesses
- After a new design always test again
- Nobody can design the perfect user interface
There are three main points on this article:
- Elaborate usability tests are a waste of resources.
- The best results come from testing no more than 5 users
- and running as many small tests as you can afford
References:
Nelsen, J. (2000). Why You Only Need to Test with 5 Users. [Internet]. Available at: <https://www.nngroup.com/articles/why-you-only-need-to-test-with-5-users/>. [Accessed on:14/3/2016].
Mary is very information in your post. But you can give me an example in which be presented why 5 users is appropriate for the best results?
ReplyDeleteAn example is when you evaluate a website of an online store.
DeleteThe first user that you will ask, he will may tells you that your colors need a change, the letters' size need to be bigger and also the button of the checkout is not easily identified.
The second user will may tells you that your colors needs to be changed and also the total cost of the purchasing items needs to be displayed at the end of the invoice.
The third user will may tells you about the colors of the website, the size of the letters and also the need to add a basket button.
As you move on and ask more users you will listen the same things and less new things. Furthermore, in order to ask more and more people you lose time. Thus is more preferable to ask less users and find the main comments of feedback, than to ask more, lose your time and discover only few additional comments.
Maria, this blog post is very interesting. But what do you mean by ' Test with 5 Users'? Do you mean the actual users of the system, engineers, programmers, friends, designers, experts or other people that are indirectly may have an interaction with the product/system/website? What do you think is the appropriate combiniation of those people and how many will you include if you want to evaluate sth?
ReplyDeleteI like your comment Anna.
DeleteThe groups of people that has to be a part of the evaluation, can be anyone that will have direct interaction with the product/system/website.
For example if you want to make evaluation for a company's software. The direct users of your system will be the employees of that company. Thus five employees of this company have to be in your evaluation process.
Also, experts in my opinion needs to be always in the evaluation process, because they can give more practical feedback.
I believe there is no appropriate combination of those people. Each different product/system/ website requires different combinations of those.
It's important to remember always include experts and the most direct users of your system.
Nice post Maria,
ReplyDeleteI would like to inform you that in one of my projects, I only use 6 users and experts for testing and the results was absolutely like your post.
So now I think we know, with evidence that the use of at least 5 users in the evaluation process is the most appropriate.
DeleteThank you for your comment George!