I am guilty of saying how much easier usability testing is than any other user research methods. We talk about usability testing being an essential and foundational skill—one of the first concepts you should learn as a researcher. Compared to generative research and ethnography, usability testing is a breeze.
That’s true to only a certain extent. While it’s easier to read a list of tasks and pointed questions than it is to have an open script (like for generative research), that doesn't capture the entirety of usability testing. Usability tests are more simple to run, but that doesn't mean they are easy.
Writing usability testing tasks is still one of the most complex parts of the research process for me. Getting these tasks right is imperative as they color the rest of the study and dictate how good your data will be.
I used to write usability tests quickly and not give much thought to the content. I would throw them into a prototype without much context (I thought this was the right way to do things) and give little to no instruction.
The participant would be trying to answer my questions based on little knowledge or understanding, which does not reflect real-world usage. These sessions would sometimes end with the team more confused than the participant and leave us without much insight.
I knew I had to change something, but I didn't know what to do or how to do it in a way that still produced valid and reliable data. Eventually, I got my bearings and created a process for writing solid usability testing tasks.
What is a usability testing task and when to use them
Before I foray into the world of constructing usability tasks, it is crucial to understand what usability testing tasks are and when it is helpful to use them.
We aim to understand how someone uses an interface by measuring efficiency, effectiveness, and satisfaction with usability testing. We can gauge these metrics by having people attempt realistic tasks on an interface, whether that be a prototype or a live product.
You ask participants to complete these tasks and observe what happens when they try. Usability testing in this way can help us understand if people can complete a task (effectiveness), how long it takes them (efficiency), how many errors they encounter (effectiveness, efficiency), and how they feel after using the product (satisfaction).
Not every single usability test needs to be like this, so it is important to decide if you need to go through this process. I have some sessions where I wanted feedback on concepts or wanted the participant to explore the prototype to see what they would do without direction. For these instances, I didn't use usability testing tasks.
However, I have started using these tasks more frequently as they help find problems, quantify them through metrics, and understand how big the problem is. To get these measurements and valuable insights, we need to be mindful of how we run the study and the types of questions we ask. We need to ensure we are writing tasks that don't bias the participant or skew the data.
Constructing a task
Throughout the years, I have honed this skill through a lot of practice and, if you are looking to write great usability tasks, that would be my first piece of advice—practice, practice, practice. These are the steps I go through when constructing my usability testing tasks:
- Start with a goal. Start with what you want the user's end goal to be, not the goal of the task. What does the user need (or want) to accomplish at the end of the action? What is their goal for using this product?
- Include some context. Instead of just throwing participants into action with no relevant information, give them context on why they need to use the product. You can also consider this context background information and why they would use the product in the real world.
- Give them relevant information. Since you are recording metrics, you don't want to be vague in your instructions. If users need to input dates, locations, or particular data in a form, give them that information. You don't want the user to guess what you want them to do, resulting in skewed data.
- Ensure there is an end the user can reach. If you are trying to get someone to accomplish a task, make sure they can. There should be a reachable "end," which is satisfying for the participant and helps you record if the participants could complete the task.
- Write your task scenario(s)*. Once you have brainstormed this information, write your task scenario. I have included some examples at the end of this article.
- Conduct a dry run (or two!). After writing down your task scenarios, perform a dry run with either internal employees, friends, family, or anyone that will give you their time. This dry run will allow you to practice the test flow, make sure the tasks make sense, and indicate whether there are too many.
*A common question I get is how many tasks should be in one usability test. It depends on the complexity of tasks and how much time you have with the participant. For a 45-60 minute session, I generally include five to seven tasks. By including a dry run in your process, you can get a good idea of how many tasks you can fit into the session.
Things to avoid
I see a few recurring mistakes that pop up during many usability tests I observe. I have to actively tell myself to avoid these mistakes (which is where practice comes in) as they are easy to slip into. Keep these in mind when writing your script and during your practice run:
- Using words in the interface. If you are trying to get users to "sign-up," "check-out," or "add to cart," and those words are in your interface, don't include them in your task. Using words in your interface makes tasks easier for participants as it leads them to the correct answer. Instead, use synonyms of the words. If you are trying to get someone to subscribe to your newsletter, ask them, "how would you get more information via email?"
- Creating elaborate scenarios. I love fiction writing and am guilty of this mistake. Sometimes I get carried away with my scenarios, and suddenly, the participant has been leading this unbelievable life that has brought them to this product. The participant has to read through the details of the scenario to complete it, so including elaborate details that aren't conducive to the task can skew your data. For example, I was asking someone to demonstrate how they would purchase train tickets in the past. I came up with this scenario that "they wanted to go on holiday to Spain to sit on the beach because work and life were stressful, but they couldn't find the perfect connection, etc." Most participants laughed at the scenario and, although it might have been relatable, it took us out of the real-world situation.
- Offending or triggering the participant. I have made this mistake several times. If you can, avoid offensive details, such as religion, weight loss, politics, health-instead of making the scenario about the participant, make it about the subject. Avoid scenarios "blaming" the participant or putting them in an awkward situation. Also, consider that you don't know the participant's position, so try to avoid holidays such as Mother's Day or Father's Day, as these could be triggering to participants.
- Biasing with content. There are usability tests I have observed where the person gives some added "content" to what they want the participant to do. For example, telling the participant to choose the option that gives them an "awesome discount." Try to keep your words and descriptions as neutral as possible.
Walking through two examples
All of this theory can seem straightforward until you go to write tasks. I seemed to hit writer's block constantly when trying to combine all these steps. If you are hitting that same wall, here are some examples to help you breakthrough.
I work at my favorite company, Dog Wishes, which helps rehome dogs from shelters by offering starter kits, training sessions, and subscriptions to help new dog parents adjust and excel.
Task goal: Find a trainer session to help them with puppy separation anxiety
- Start with a goal. I want new dog parents to a trainer to help them tackle puppy separation anxiety for this task.
- Include some context. I will give a brief overview that these new dog parents have a puppy experiencing separation anxiety (chewing, barking) when they leave for errands, and they are looking for a trainer to help them with this.
- Give them relevant information. To make this easy, I will give the participants dates (in the future!) to book. Ahead of time, I will look at where the participant is from, find the closest city trainers available, and give them that information. I will tell them they want an available trainer on the date specified specializing in separation anxiety, and charges less than $75 per hour.
- The end state. The users will get to the booking page, which would be the end. If I wanted them to BOOK the session, I would have to provide fake credit card details.
Task scenario: Your puppy is displaying separation anxiety symptoms (chewing, barking). You want to find a trainer who is available in New York City on May 7th, 2021, who specializes in separation anxiety and charges less than $75 per hour.
Task goal: Purchase a food subscription for puppy food
- Start with a goal. I want new dog parents to find and purchase a new food subscription for puppy food for this task.
- Include some context. Since this new dog parent just adopted a puppy, they want to find a subscription that delivers puppy food once a month. (Just as a note - it was hard for me not to include words or phrases such as convenience, healthier food, or supporting local shops)
- Give them relevant information. To make it easier, since there could be many results, I will give a brand they are looking for and the cadence of the subscription (monthly). Since I want them to purchase the product, I will provide them with a fake credit card and shipping information.
- The end state. Completing the purchase and arriving at the confirmation page.
Task scenario: You just purchased a puppy, and you are looking to purchase dog food. You want to find and buy a monthly subscription for the Fresh Dog Food for puppies.
Remember your participants and audience! Just because your tasks are realistic doesn't mean you are already there - you need to make sure you recruit the right participants for the study to get the best data and insights!
Nikki Anderson is a qualitative user experience researcher with about 5 years in the field. She loves solving human problems and petting all the dogs. Read more of her work on Medium.