Guerrilla testing with your prototypes! / by Gavin Lau

Why Guerrilla?

Whether you’re exploring a particular research problem or testing a design solution that you already have, guerrilla research can be a huge time-saver. Guerrilla testing in it’s simplest form is getting onto the locations where you would find users of your persona and asking them to use your prototype for a minute or so..

Most stakeholders and clients refrain from usability testing because it involves spending lots of time and money to see the results. But why would anyone object to a usability method which does not cost much and gives amazing insights before you jump to development? That’s guerrilla testing!

This article will walk you through the process of executing a guerrilla user test for rough and high fidelity prototypes. The beauty of this format of testing is that it doesn’t have a hard-and-fast approach. I’ll talk about how you can use CanvasFlip for guerrilla testing to save hours and money while getting closest to amazing experience every time.


Step 1 : Creating tasks to be tested

Create a list of task whose flows need to be tested

While designing an app, we always have a set of flows in mind that our users will be doing on the app. The first thing is to write down all these flows.

So just for example let’s check out few of the task flows for Whatsapp -

  1. Send a message to a friend
  2. Create a group of friends
  3. Send photos to your friends
  4. Broadcast a message
  5. Change your display picture and so on…..

Similarly, make a list of tasks for your app that you want to test. Think of everything that your user can possibly do your app and write it in the form of tasks.

Guerrilla user testing is easy and you’ll get valuable insights even if you’re doing it for the first time. You can do almost everything wrong — as long as you get your tasks right your user testing will always work.

Prioritize top 3 tasks — Convert to test scenarios

You can’t really test everything in a guerrilla testing, we really have to respect our users’ time. So we need to prioritize top 3 tasks. One does not really have to break his/her head on this. You just have to rank tasks in the order of frequency of usage of a task. For eg, when I’m on whatsapp, the one thing that I always do is send a message to a friend, second in the list migh be sending an image and so on… This pattern has to quick, like really quick. So just in case you miss out on something really important, you can always test it the next time.

Once 3 tasks are finalised, you need to convert tasks into scenarios which users can understand and grasp easily. Scenarios should be crisp and easy to follow. While finalising the scenarios remove any part that does not contribute and simplify the language.


Step 2 : Create a prototype

Honestly, users do not really relate to rough sketches much, but detailing it out a little before the test, might help a lot. On the contrary, testing with rough sketches sometimes gives a user a sense of ownership, that they are an integral part of the building process. The choice of the level of fidelity is completely yours.

To test your designs, you definitely need a prototype. There are multiple prototyping tools over the internet, but specifically for guerrilla testing I strongly suggest you use CanvasFlip. Why? Because it’s my baby? Well ya that too, but honestly, it helps you skip a lot of steps while testing. We’ll see how :)

 Create a prototype for user testing

Create a prototype for user testing


Step 3 : Get out of the building and get started

Guerrilla testing is straightforward — have participants use your prototype, while you draw insights from their prompt sessions. This does not necessarily have to be the traditional way, where you stand right behind the shoulder and take down notes of the sessions or use cameras to record whatever is happening. With CanvasFlip, as I said you can skip a lot of these steps, yet not compromise on the insights.

During the test, make sure you don’t continuously keep taking notes. It only leaves the user stressed about what you are writing, good or bad or have they done something wrong. All the insights you require will be recorded in CanvasFlip, so I believe one should avoid taking notes all the time.

Get as casual as possible to avoid scaring your users — That’s when CanvasFlip does the magic.

Get as casual as possible to avoid scaring your users — That’s when CanvasFlip does the magic.

Here’s how to do it with CanvasFlip -

  1. Just share the prototype link with users you meet at a cafe, or bus stand, or any public platform over whatsapp or gmail or any convenient social channel.
  2. No plugin or app installation is required to get started with the sessions.

That’s it! All these sessions and more insights about the entire test is saved in your CanvasFlip dashboard.

 Share your prototype and let users use it

Share your prototype and let users use it


Step 4 : Your guerrilla testing insights are captured. Analyse them.

Guerrilla testing is about finding and fixing the most severe problems, not agonising over every possible obstacle a user might encounter. And the best way to find the most severe usability problems is to replay the sessions of the users and analyse where users are getting stuck, or what is that grabbed the interest of your user.

Study the path taken by users and why: User videos on CanvasFlip, records and replays the sessions of the your users. I use user video to compare the path I imagined the user would take and the path user actually took. Surprisingly, most of the times the two differ a lot. Now that’s quite a valuable input for any designer.

Apart from that I also use it to find out parts of the sessions which were confusing or intimidating for the user.

Capture the task completion ratio: Conversion funnels are more of a cumulative insight for the entire test. It gives you numbers that you really really need for iterations on designs — drop-off rate and conversion rate. It gives you number of drop-offs on each screen, which is quite helpful to decide which screen needs to be taken up on priority.

I also use this report to find out the most popular flow for a particular task. I often come across situations when I end up designing a task from 3–4 sources. Later deciding which flow to crap off and which to retain is a dicey situation. Conversion funnel truly helps in clearing air in such situations.

Heat maps are quite handy when you want to decide which page area is getting enough engagement. More than often we are most concerned about our Call-to-action’s performance — Heat maps are the best to gauge that. I use heat maps to study the page areas that got enough attention from the users. It is a great input for iterating on the call-to-actions and important UI elements on the page.


Now, implement quick fixes and test again!

Final words

Guerrilla testing is the most amazing way of getting value out of a huge crowd in less time. But if you don’t have that time for testing, because an hour of your time costs 50+$, you can also resort to remote unmoderated usability testing.

Remote testing removes many of the challenges related to scheduling, travel, paperwork and setup. You can literally shave off days, or in some cases weeks, with remote testing. And with CanvasFlip you reduce those hours even further. :)