Party testing
5 tips on how to engage users and get valuable insights while getting tipsy.
Disclaimer; This isn’t your ordinary process article on how to recruit for a user test or how to get the most valid results from it. No, this is how we turned a party into a test lab to quickly gather user insights.
The proverb says Children, fools and drunken men tell the truth. What happens if you decide to test that “hypothesis” and apply it to a design project? Well, that is what we decided to find out while working with the redesign of the Danish Metrology Institute’s (DMI) website and mobile app.
So how do you go about testing on children, fools and drunken men? Children were the easy part since we easily could bring in the resident office child to test on. Testing the full concept and getting him engaged in card sorting and a user test on our wireframes proved to be a bit hard, (well duh, he is six years old) but getting feedback on the understanding of icons and weather symbols was a breeze.
Next, drunken men. We had an afterwork cocktail party planned and therefore the prefect opportunity to do some live experimenting on our drunken friends. Lets break down what and how we went about setting it all up.
The same three tests we had set to the office child were prepared during the day leading up to the party and one facilitator was responsible for each user test. The guests, or users, started arriving at the party and we were casually asking them if they wanted to attend one of the tests. We made sure that the test still took place in the party setting. We wanted the users to feel relaxed and to be able to enjoy a good beer during the test.
Card sorting
The facilitator started out by explaining the exercise, encouraging the users to think out loud and getting the guest to prioritise the cards with weather information.
This part of the test took relatively short time to execute and gave us some useful insights. The card sorting was done with one guest at a time, as it was important to ask follow-up questions and to document the exercise. Even though the test was short, it removed the guests from the party setting. Therefore, this exercise was only done with four guests right at the start of the party.
User test
Similar to the card sorting exercise, the user test was also done with one user at a time. The test started with general questions about the guest and how/why they check the weather. After that the facilitator presented wireframes as she wanted to see if the guest understood the concepts, and to see which one they preferred. Usually, we record video of our user tests but because of the background noise and the casual setting we were limited to note taking. This meant that the facilitator spent a lot of time documenting both during and after the tests. Since this sort of test requires a lot of time and focus, we stopped testing after just three guests.
Icon test
For the third exercise, the icon test, we used a wall in our office. On one side of the wall we hung the icons we had designed for interpretation and on the other side, names of different weather phenomena. Next to the wall, a written explanation of what we wanted our guests to do. This meant that the exercises did not have to be explained and the guests could get started without anyone supervising them.
The exercise started attracting other guests, as our guests wrote and drew their explanation of the icons and the weather phenomena on post-its and pinned them to the wall. The more friends arrived at the party, the more the wall was filling up with feedback. In total, around 90% of our guests participated in this user test. It surprised us since we weren’t actively encouraging the guests to participate.
Sure, some of the guests were discussing their answers which might have influenced their responses but at the same time it also contributed to a relaxed and casual atmosphere.
The two first tests took place at the beginning of the party, so the guests had only drunk a couple of beers or cocktails. The icon test however was an ongoing exercise throughout the whole evening, where more and more beers and cocktails were consumed… Thus, some of the results were the product of tipsy guests who came up with more creative interpretations like ‘acid rain’ or ‘raining with spears’. That kind of creativity has to be kept in mind when analysing the results.
Key insights
All in all it was a great experience hosting user tests in a more casual atmosphere in contrast to a classic lab or interview setting. It became clear to us that the easier the test setup was, the more interaction and feedback did it yield. In our limited test it was obvious that the visual and easy to understand icon test was the best performer since it had a social aspect to it, was easy to commence without any supervision, and it dealt with a subject matter that everyone could relate to.
Want to conduct your own test?
Here are our five tips on party user testing:
- No context needed: Choose exercises that can be tested by users outside the bigger context of the project, like the weather icons.
- Self-explanatory: Find exercises that users can easily identify and that are self-explanatory.
- Self-documenting: Don’t run around taking notes but rather find a method where the users document themselves or their results/process while doing the test, such as writing on post-its. In our test the colourful post-it wall also helped to attract more test subjects.
- Using the qualities of the party (or setting): Don’t force a regular user test into the party setting, rather allow the test to support party interactions, discussions, and creativity.
- Casual and fun/booze: Don’t forget to have enough beer, great music, and of course awesome guests who are willing to party test with you. But keep the guests’ “extra creativity” in mind when analysing the results ;).
You can of course apply all these findings and our five tips to any casual test setting, be it testing a concept in a client’s reception, cafeteria or restrooms. But you might want to replace the alcoholic beverages with something else…
Oh, and what about testing on fools you ask? Well, we haven’t found a way of doing that yet but we are open to suggestions if you have any!