DonnaM » Blog Archive » Selecting scenarios for a usability test

Selecting scenarios for a usability test

I used a new method of selecting scenarios for a usability test for a client last week. The test was for a system that would help frontline staff with client questions. We had an initial list of 30 scenarios that I thought looked good, but because this was my first exposure to the users, I wasn’t sure. I would normally select 10 or so scenarios, but didn’t want to in case the ones I picked weren’t realistic or representative.

So we printed out all scenarios on index cards. At the beginning of the usability test I asked the participants to put them into 4 piles according to how often they were asked that question – frequent, sometimes, rarely and never. I also asked participants if there were questions that were asked frequently but were not covered. This gave me extra information about what clients ask about and checked that the scenarios were realistic. I chose 4 scenarios from the ‘frequent’ pile and 4 from the other piles, making sure I didn’t double up on topics.

This worked well – the participants were working with real scenarios so I was confident that their reactions were close to a realistic situation. Choosing more from the ‘frequent’ pile ensured that the usability test covered core questions.

Normally I wouldn’t use a large variety of scenarios with a small number of participants as I would not get enough coverage to identify repeated usability issues. However it worked in this situation because it was on a small, reasonably homogeneous set of information. The test showed that the participants’ actions and usability issues were similar for different scenarios.

3 Responses to “Selecting scenarios for a usability test”

  1. joe Says:

    Brilliant idea! I love the idea of allowing users to, in effect, write their own test.

  2. Sankarshan Says:

    Nice idea and perhaps actually the best way to deal with the case which you faced. But for scenarios where the heuristic knowledge frontend from the client side is not readily available – what exactly do you think needs to be done ?

  3. donna maurer Says:

    I think you are asking what to do if you don’t know what people use the site for (if not, tell me and I’ll think about it again).

    If you don’t know what people want to do with the site there is no point testing it – everything that you test will be artificial and the results will not be realistic. Put time into user research instead. Find out what people need out of the site, design it and test as you go.