What is Guerrilla Testing and how do we do it?

By Jeremy King

Guerrilla testing. If this term has you packing up your lab coat and clipboard for a trip to the zoo, think again. Guerrilla testing is a type of UX testing that brings quick, affordable and actionable results without a primate or complex testing environment in sight. Interested? Read on to discover what Guerrilla testing is, why it’s so great and how you can conduct your own.

Back to basics — what is Guerrilla testing

Let’s start at the beginning. Usability testing is necessary but can be time-consuming, resource-consuming and profit-consuming. Guerrilla testing is a technique that anyone can easily implement and on a budget, that changes that. Guerrilla testing is a UX testing method that involves heading out into the jungle (the public), asking people to test your product, site or app, and then recording their feedback.

The benefits of Guerrilla testing

Beyond its simple definition, Guerrilla testing brings many other benefits, including:

Real people

Involving people sat alone in coffee shops, on park benches, or in co-working spaces, means that you involve real people in the design and development of your product. And, real people means real feedback that helps you to improve what you’re testing for your audience.


Guerrilla testing is also super quick. Got a new idea? Made a tweak to your design? Having a departmental disagreement? With Guerrilla testing, you can head straight out there to get immediate feedback and, even better, ongoing feedback throughout the project.

Qualitative results

The feedback you get is qualitative and anecdotal, which allows you to quickly validate the effectiveness of your design on your intended audience. No number crunching in sight.


Guerrilla testing is also an incredibly cheap way of usability testing. With no travel, time or revenue costs you can save money, enabling you to test throughout the life of your project.

Competitor insights

Finally, Guerrilla testing lets you take a sneak peek on the other animals in the zoo — your competition. Public testers will naturally compare your product with what’s already out there — giving you valuable insights into what your competition does well, and not so well.

When to use Guerrilla testing

Sound good? You bet. Guerrilla testing is great, especially for:

  • Testing usability issues early on in a project;
  • Testing assumptions you’ve made about your audience;
  • Finding out how your design compares against your competitors’; and
  • Testing the ease of any tasks that (you assume) require no skill.

Of course, as with anything in life, there are some potential disadvantages to using Guerrilla testing, including:

The speed of the testing limiting the thoroughness; Not being able to test designs that require specific skills or knowledge; Public places being noisy, messy and distracting; Someone wanting to enjoy their tall skinny caramel latte in peace; and The tall skinny caramel latte drinker being so far off of your intended audience that their insight is about as useful as Hugo’s.

However, it’s important to remember that Guerrilla testing should always be combined with other UX testing methods, overcoming all of these disadvantages. Plus, even if Hugo isn’t your intended audience, he may have some pretty good feedback nonetheless.

How to conduct Guerrilla testing

On the whole, Guerrilla testing is hugely effective. When done right. The key to getting it right?

1. Preparation

Identify what you want to achieve, and then plan how you are going to achieve it. Specifically, you’ll need:

  • A prototype— it doesn’t need to be the finished product, but it does need to be a functioning prototype;
  • Tools — everything you need to conduct the test in a live setting (e.g. a laptop, phone, tablet, charger, video camera, etc.);
  • Test scenarios (discussed below); and
  • An idea of how long each test lasts — ideally around ten minutes.

2. Creating a scenario

Break down the tasks people need to do to use your design, narrow these tasks down to the three most important ones, and then create a scenario based on that task. Want an example?

“It’s Friday, and you deserve a takeaway. Order enough pizza to feed four, and add drinks for everyone. Go on, give it a whirl.”

3. Finding your victim

Pick a location where your target audience spends their downtime — somewhere they’re not going to be rushing away from. Do a recce beforehand to ensure that it’s not too noisy, busy or distracting and that the staff are happy for you to test there, use their WiFi and borrow their power.

4. Approaching your victim

Next, you need to approach someone, say hello, ask for a minute of their time and then explain:

  • Who you are and what you’re doing;
  • How they can help; and
  • How long it will take.

If they’re happy to take part, ask and record some more information about them and whether they’re your target audience. For example, you might want to know how comfortable they are with technology or if they use your competitor’s products.

5. Running the test

When giving your participant the scenario to complete, it’s important to tell them that they’re not being tested themselves, that they should verbalise what they are doing and thinking, and that they’re not to be shy about their feedback.

6. Observation

During the test, observe the participant visually and, if they’re ok with it, by audio, video or screen recording. Following the task, ask them about their experience, what they liked and what they didn’t. Oh, and don’t forget to take notes.

7. Working on the feedback

The most important stage of them all. Turn any feedback into actionable improvements, and don’t forget to delve into any unsaid feedback; if everyone said how great the app looked but no one managed to order their takeout — you’ve got a problem that needs fixing.

Don’t forget

And that’s Guerrilla testing in a nutshell (if gorillas eat nutshells?) But, before you grab your coffee card and head out the door, don’t forget the following:

  • Run your scenario past someone in the office (or at home) first, ensuring it’s tip-top;
  • Don’t have more than one person with you — it really isn’t a zoo;
  • Have a backup plan for poor WiFi, sound or timekeeping;
  • Vary your locations, to vary your victims;
  • Play with the amount of users but a good place to start is with five. Complex products may require more, which is a subject for another post, but you will find this out when you begin testing; and
  • A thank you is in order. Buy them a coffee or cake. Or both.

Happy testing!

There's more to read...