Are You Prepared for Your Usability Study? Three Steps for Success
by Christine Perfetti
At DrupalCon Austin, I'm very excited to be teaching a full-day workshop, UX Bootcamp: Evaluating and Improving Your Web Sites. In the workshop, I'll be sharing all of the essential steps for designers and developers to evaluate and improve their web sites. Read on to find out how to get started.
The first thing I tell people is that usability testing isn't an arduous process. It's a technique that anyone can learn with training and lots of practice. At its core, a usability test involves putting a person in front of the product and watching what they do. The goal is to observe how well users accomplish their goals with a product.
If you're thinking about conducting your first usability test, you'll want to take the time to prepare appropriately for the study with three important steps:
- Writing a test plan
- Recruiting the right participants
- Creating tasks
After completing these steps, you can begin your usability study with confidence.
Step 1: Writing a usability test plan
At the beginning of any usability testing project, you'll want to meet with your design team, engineers, and organization's stakeholders to identify what they hope to learn from the usability test. It's essential to have everyone who has an impact on design decisions attend the planning meeting. This gets everyone on the same page and ensures you won't encounter any surprises when the stakeholders observe the test sessions.
We typically allot two hours to the planning meeting. We start by giving an overview of the testing process to ensure everyone is on the same page. This also gives the team an opportunity to ask any questions they have about the usability study.
We then work to define the focus of the test and outline the research questions we'd like to answer. With a large and complex product, it would be impossible to evaluate all concerns and issues in one usability test project.
Because of this, we ask the stakeholders and design team what their biggest issues, risks, and concerns are with the product. We also ask the team what product features they know the least about and how users will interact with them. After getting the team's feedback, we have a much better sense of where to focus the usability study.
Based on what we learn from the planning meeting, we write a test plan to outline the goal of the tests, the research questions, and the methodology. The usability test plan ends up being the blueprint for the test we'll run. It can be a formal or informal document, but typically includes the following sections:
- Goals of the study
- Research questions and issues
- Description of the target audience and key user behaviors we'll recruit for
- Test method
- Project schedule and timeline
- Tasks we'll ask users to complete
- Data we'll collect
Step 2: Recruiting Users
When preparing your first usability study, you'll want to find people who are representative of your target audience. You have a couple of options for doing this. With in-house recruitment, you can assign someone internally to find your users. Or, you can outsource to a reputable recruitment agency.
When thinking about the right participants to recruit for a study, many teams start by focusing on demographics, such as age, gender, or ethnicity. Unfortunately, in most cases, recruiting for demographics will be one of the least effective ways to find the most appropriate users for your tests.
For example, if you were recruiting for a usability study of a video game, what demographic would you usually think of? People who play video games are typically thought of as boys in the 13-24 age range. But if we only recruited for that demographic, we may mistakenly recruit boys who don't actually play video games. Additionally, many women also play video games. We would risk missing their feedback if we only recruited the boys.
In our work, we only focus on demographics when it's a critical component of the target audience, such as when we're recruiting for a university portal devoted to college students or a social security benefits web site devoted to elderly users. Rather than focusing on demographics, we focus on the specific behaviors the target audience exhibits and their levels of tool knowledge and domain knowledge.
For any recruitment project, we work with the client to answer the following questions:
- What are all of the specific behaviors we’re looking for in our users?
- What level of tool knowledge do users need?
- What level of domain knowledge do users need?
The first question we ask our clients is, What are the key behaviors of your audience? For example, if we were recruiting for a video game system, we'd want to find people who play video games. If we were recruiting for an e-commerce site, we'd want to find people who use the internet and have bought products online before.
Users' behaviors are also informed by their amount of tool and domain knowledge. The users' tool knowledge, the level of familiarity users have working with particular tools or products, can have a huge impact on how they behave. We recruit for tool knowledge when the test includes intermediate or advanced features and techniques. For example, if we are recruiting for a video game, we may need to find people who already know how to play with a specific game console, such as the Wii.
We recruit for domain knowledge when users need to know and understand specific information to use the product. For example, in our recruitment for a video game, if we were testing the adventure game, Myst, we may want to find people who are already familiar with the game and how it's played.
When finding users, we start by asking clients if they have a database with existing users and prospects. This is always a great place to start. We've also had a lot of success with recruitment firms, Craigslist, user groups, temp agencies, real estate agencies, and even friends and family.
Once you've found some potential recruits, it's necessary to screen the people by phone before bringing them into the lab. To help with this process, we develop a screener to qualify or disqualify potential users. The screener is a script that helps the recruiter apply the requirements we're seeking in users. The parts of a screener script include:
- Greeting and the purpose of the study
- Qualifying questions to assess whether we've found a user who exhibits the right behaviors
- Scheduling dates and times for the test session
- Compensation information
Step 3: Creating Tasks
The tasks you create for a usability study are essential for gathering the right data. The tasks determine what you'll test and impacts what parts of the design your team fixes. If you give users the wrong tasks, you risk focusing on the wrong parts of the design. Even worse, you may provide your design team with misleading recommendations. Yet teams often spend little time preparing for this crucial step.
When creating tasks for a usability study, you'll want to start by asking yourself the following questions:
- What are your users' goals with the product? List out the specific actions users most commonly complete with the product.
- What are your business goals? Think about all of the different ways your product or web site will increase revenue or reduce organizational costs. The best tasks focus on areas of the design crucial to your organization's business goals.
- What are the greatest risks with the design? If there are certain areas of the design where you have little knowledge as to how users interact with it, this is another area to address in your tasks.
There are three types of tasks for a usability test: verb-based tasks, scavenger hunt tasks, and interview-based tasks. Jared M. Spool and his team of researchers at User Interface Engineering first introduced this framework for designing tasks back in 1999.
Verb-based tasks ask users to accomplish a specific action with the product. Verb-based tasks are most commonly used to test software, hardware, and web applications. For example, for an email system, we might ask users to:
Respond to the email you just received from Kate Austin Write a note to your mother Copy the text of this page to another document Send the message from Kate to your friend, Lisa.
All of the tasks begin with a verb and ask users to complete a specific action. Verb-based tasks effectively evaluate the product's functionality and give teams the capability to test multiple users on the same tasks. Before the advent of the Web, almost all tasks for evaluating products were verb-based tasks.
Scavenger hunt tasks
Unlike verb-based tasks, we don't use scavenger hunt tasks to evaluate software or web application functionality. Instead, scavenger hunt tasks help us to assess content-rich systems such as CRMs, rich data displays, and information-rich web sites.
With scavenger hunt tasks, we ask users to find a specific piece of information. These tasks help design teams evaluate whether users find and understand the product's content. The tasks almost always begin with the verb, find. Some examples of scavenger hunt tasks:
You were at a party last week. The discussion turned to recipes for authentic Italian pasta dishes. Go to the Food Network site and find an Italian recipe for pasta.
The doctor stops by on morning rounds and wants to know how much Mr. LaFleur's blood pressure has been out of the normal range during the night. Find a record of Mr. LaFleur's blood pressure.
The downside of traditional tasks such as verb-based and scavenger hunt tasks, is that it's challenging for teams to predict whether they've chosen realistic tasks for users to accomplish. Because of this, teams risk giving users tasks to complete with the product that aren't related to what they would actually do in a real-life situation. Interview-based Tasks
To address the limitations of verb-based and scavenger hunt tasks, we use interview-based tasks, a task methodology developed by User Interface Engineering. With interview-based tasks, we interview users before and during the test to uncover users' real goals with a product. During the recruitment phase, we screen candidates to ensure they have the appropriate interests before they come to the lab.
With interview-based tasks, when users first arrive for the test session, we don't actually know what we'll specifically be asking them to do during the session. Instead, at the beginning of the test, we interview users to get a better idea of how they use a product. For example, when evaluating an investment web site, we would start by asking the users specific questions, such as:
Based on the users' specific responses to the questions, we'll work with them during the session to create tasks that are relevant to their specific needs. While we won't ask all users to complete the same tasks, we get a very good sense of how the product works for users in the real world.
Interview-based tasks work best with sites and products that are almost ready to ship and populated with real information and data. Without real content or data for users to manipulate, it's impossible to mirror the true experience for the user.
After writing your test plan, recruiting your users, and creating your tasks, you'll be ready to run your sessions. You'll also be confident you're fully prepared to gather rigorous data from the sessions.
Want to learn more?
These are just a few of the topics I will share at DrupalCon Austin. Come join me this June to learn more!