Guerrilla testing

This informal method of user research involves approaching people in public spaces and asking them to share their opinions or experiences of a website, app or product.

Strengths

  • Quick, simple and easy to carry out, especially at short notice
  • No need for a complex technical set up, room bookings, multiple researchers or extensive data sets and analysis
  • Can be done simply with one person, using a device like an iPad, or even a paper prototype

Weaknesses

  • Does not give you in-depth data
  • Depending on timing, can be tricky to find enough participants

Instructions

You will need

  • At least one full day to prepare, carry out and analyse the research
  • Portable version of the thing you want to test: for example a webpage or survey loaded on a tablet, a print out of a webpage, or a paper prototype of an app with each screen on a different sheet
  • Script including:
    • What you're going to say to encourage people to take part
    • The instructions you'll give them when they do, eg "Please think aloud as you work through each task"
    • A list of your test scenarios or questions
  • Notepad and pen to take down your observations
  • Location where the sorts of people you want to test with will be plentiful and available: for students try the University Library or Fraser Building

Instructions

Observations

UX Research Intern Holly Pringle shares her experiences of carrying out guerrilla testing for the UofG UX Project

To streamline the design of the Library homepage, we followed the Top Tasks Management method. To get responses to our Top Task Survey, I put it on an iPad and took it to the Library. I think students find it much quicker and easier to do surveys on an iPad than on paper, because we got far more participants this way. 

I asked random students to choose their top five tasks, and they also had the option to add comments in a free-text box. Out of 62 participants, only 1 added a comment: perhaps because it’s awkward to type on an iPad when you’re standing and holding it, and perhaps because it required more thought than the quick and informal atmosphere allowed.

To test the prototype of the new Student Printing page, I approached students around the printers in the Library, and on the hill outside. I asked them to complete some typical tasks on the prototype page, again on an iPad. I made notes on how they worked through each scenario and on their general behavior. I also rated how quickly and smoothly they were able to complete each task, to give us an indication of how efficient the new page design was. This worked well since we only needed a couple of minutes with each participant. This method could definitely be employed effectively during busier times of the year.

For future website tests it may help to find a screen recording method for the iPad, so we can record each user’s journey through the pages. We may also have benefited from testing the Student Printing page before and after the redesign, so we could compare its performance.

I found that:

  • If you approach people confidently and in a friendly manner they are very receptive and helpful.
  • People are happier to take part if you assure them it won’t take too long.
  • It seems to work best when people are outside or on a study break as they are more willing to give you their time.
  • Those whose native language is not English can find it a bit more difficult to understand, or to find things on a webpage, but they are still valid participants and we need to consider them in our designs.
  • For the Student Printing pages, it worked well to approach students at the printer because they weren’t expecting it, so being caught off guard helped get their attention, and they had to wait for their printing to finish anyway so they had the time.
  • It helped that I have experience working in a customer facing role: it made it easy to approach people.

While the data we can get from guerrilla testing is useful, it shouldn’t be the only user interaction you rely on when looking into user behavior or validating designs. You should also try to incorporate at least one more in-depth form of data collection, such as interviews or behavioral data from session recording, heat maps or eye tracking.

Further reading