You've finished designing a website, but how can you be sure that it really suits your target, that it's effective and that it fulfills its objectives? There's only one way to find out: do a UI/UX test. How many tests to be precise? And on how many users? A whole study has been dedicated to the subject, and that's what I wanted to share with you today. But first, I'd like to have your opinion.
How many UI/UX tests should be done? And how many test users should be involved?
As little as possible! Unless you have a very high budget.
As much as possible. You might as well be sure that the site works. After all, it's an investment!
A completely different proposal? Don't hesitate to leave a comment.
In a UI/UX test, there are usually several participants (more than two, anyway). This is in order to get satisfactory results. With the first participant (test user) you can already gather as much information as possible. This first session alone is very rewarding, because by the end you have already found so many mistakes.
Of course, the second test is less rewarding than the first, but it is still very interesting. In this second test, the user in question reproduces the same behaviors as the first user. You will certainly get the same data as in the first session (except for a few details).
Since the scenarios are almost the same (in the first and second tests), the second user test helps you to find new ideas for design improvements.
So what about the next users? Well, their behaviors are repeated. The more users you have, the less you learn. So going beyond 5 users would not be a good idea. It will just be a waste of time and money. Since you're spending time and using third party people for UI/UX testing, you're spending money for nothing.
🙊Yet, it is said in various studies and reviews that it takes at least 15 users to ensure authenticity of data in a UI/UX test. Is this true?
In a way, I would say yes, it is true. If you run 15 tests with 15 different users, your site will be great (without usability and design issues)! But as I told you earlier, those 15 people will have pretty much the same behaviors.
I suggest you to limit yourself to 5 users per UI/UX test, but to perform 3 series of tests. The number of people who would have taken part in the test would therefore be 15.
You will have spent the same amount of money as with 15 users in a row, but at least your budget is spread over several small tests . With each completed study (1 UI/UX test with 5 different users), you manage to improve the site or the product you have created.
The first test usually helps detect 85% of problems. Before moving on to the second test, it must be corrected. More tests are needed, because there is no guarantee that there will be no more errors. For the second test, we will call on another person (the second test user) who will help us to see if the first improvements were satisfactory and why not, to identify new problems. 15% of usability problems are identified at this stage. For about 2% of the problems, we will see them in the remaining tests.
❗️ I would like to point out that the purpose of a UI/UX test is not only to identify errors in a site, but above all to improve its design and its interface.
I said that to save money, we should limit ourselves to 5 users per test. Yet, these 5 users will have the same behavior in each session. And also, we learn much more in the first UI/UX test than in the next four.
You may ask yourself: why waste money asking 5 (or 15) people to take the same test? If we want to save money, why not trust one user for the different sessions?
Don't worry, there are at least 2 reasons why 15 people (and thus 5 users per test) are needed:
📌The behavior of a single person is not totally reliable. Who knows if she clicked a button by accident? You have to observe at least 3 different people to be able to draw a conclusion, if a behavior is unique or if it can be generalized.
📌Contrary to what you think, using 5 users per test is more economical. Try to imagine this scenario: you limited yourself to one user, you think the site is ready, so you put it online, but unfortunately, the site doesn't bring you anything! All the effort you put into it is gone, because the data you received is wrong. But if you had used 5 different users per test, you could have avoided all this.
Absolutely! It can be done, but with one condition. The site in question must target several distinct user groups.
Here are 2 examples of different user groups for the same site:
Of course, these groups of people are quite different. But one thing is certain: there will be similarities in their behavior (as I told you earlier in the article). After all, they are all humans.
In short, you can use more than 5 users per test. With less than 5 people, it can also do the trick. Not only do you save money, but you also don't spend too much time in the UI/UX testing session.
Here's what I recommend:
NB: never less than 3 users to ensure diversity of behavior within a group.
One UI/UX test is not enough. You need several, at least 1 test in which 5 test users participate, at best 3 tests with 5 people each. In some cases, however, more users are needed (when we are talking about a site with different target groups).
However, these different test phases all have very specific roles: