It starts with any number of scenarios: Design and development have taken too long to produce a prototype, you need to release in three weeks, and you suspect there may be design flaws. You are trying to incorporate usability testing into an Agile development process. Or maybe you simply want to pare down your process to make it shorter and less expensive.
Completing usability testing quickly is a challenge anywhere but especially in consultancies, which have to overcome additional challenges, such as learning a new application. To assure success on these projects, I’ve developed a quick turnaround usability testing methodology (QTUT) that minimizes the time needed to complete testing. In Part I of this article, I discuss how to make the first three steps of QTUT—Sales & Kickoff, Recruitment, and Preparation—as short and efficient as possible. In Part II, I will discuss the final two steps: Testing and Analysis & Reporting.
Steps in the QTUT Process
Step 1: Sales & Kickoff
Step 2: Recruitment
Step 3: Preparation
Step 4: Testing
Step 5: Analysis & Reporting
Sales & Kickoff
A new client or a group within your company has approached you about doing usability testing. They need the results next week, which works out to six business days from today. What should you do?
Before a project kicks off, we typically have a number of discussions with the client to understand their goals and deadlines for our engagement. However, in QTUT, each day spent in the sales and kickoff process takes a day away from testing and analysis. To help our sales team conduct initial discussions efficiently, we’ve prepared a one-page guide outlining the do’s and don’ts of QTUT.
One critical “do” for our sales team is that they should discuss the method with the client and immediately set deadlines for our testing results. Starting with the final due date, an experienced tester can work backwards to determine the dates for testing, beginning recruitment, and finishing the test plan.
It’s also important to review expectations with the client. For example, current clients typically expect a certain level of quality in our deliverables. When we compress a week of work into one day, delivering a perfect document or presentation is impossible, so simply reviewing the timeline and discussing how we plan to shorten the process is extremely helpful.
Since the project must start very quickly, our sales staff and project team use part of the kickoff as a working session. During this working session, we develop specific goals, learn what types of results will be helpful, develop an initial list of testing tasks, and learn about the users that we need to test the application. We also compile a list of what we need from the client. Depending on the project, we may need:
- Information about users for both recruiting and for writing the test script. For example, do people regularly use the application or do they only use it occasionally?
- Training materials for the application and a subject matter expert who can answer questions.
- Background information, feedback, or previous testing materials that give context to the current design or may help us to write screeners and test scripts.
- Access to a stable application, especially when it is under development.
Sales & Kickoff Tips
- Create a “Do’s & Don’ts” guide to help team members through the process.
- Avoid clients who are rigid, who prefer not to participate in the process, or who want long reports.
- Use the kickoff as a working session to learn about the participants and potential tasks.
- Have a facilitator available with the domain knowledge needed to quickly learn the application.
- Refine your methodology before you have a project.
Before starting the project, you called several of your favorite recruiters to see if they could meet the deadline. You’ve gotten information about participants during the kickoff meeting, and now you simply need to write a screener in less than two hours!
Recruiting quality participants is a big challenge. There are two key components to recruiting: writing the screener and scheduling the participants. The screener is a list of questions used to filter out participants who are not appropriate for your study. Scheduling involves calling potential recruits, asking the questions on the screener, and scheduling participants who qualify and are available.
Because scheduling is time-consuming, it’s critical to write the screener immediately; in fact, we often start and finish the screener immediately after the kickoff meeting. That’s why it’s important to get information about participants during this meeting.
The level of effort required to recruit depends upon whether you’ve recruited similar participants before and whether you handle the scheduling internally or externally. If you’ve recruited similar participants before, you can reuse the same screeners almost verbatim. To facilitate recruitment for new participant types, we’ve built a list of standard questions so that, at worst, we need to write only a few new questions. In addition, we use a screener template that has our facility procedures, location, and contact numbers; the facilitator need only fill in the specific test questions.
External schedulers. We typically use external schedulers, which is particularly helpful for QTUT, because it takes burden off the facilitator. I highly recommend contacting your external recruiters before agreeing to do a study to see if they can meet the timeline that you need. If at least one of your external recruiters cannot meet your timeline, do not hesitate to ask your client for assistance before agreeing to the study.
Internal schedulers. If you do your own scheduling for QTUT, you may have a list of participant leads in house, so writing a screener isn’t always necessary. However, depending on your timeline and your success rate when calling, it’s often necessary to ask your client or a co-worker to make the calls for you.
You might be tempted to schedule testing sessions very closely together in order to complete more sessions per day. However, we find it more efficient to leave time between sessions in order to quickly debrief and begin analyzing what we observed during the previous session (I’ll discuss this analysis method in more detail in Part II). You may also need the extra time to modify the testing script as needed before the next session.
- Start recruiting as soon as possible.
- Create a screener template so that you only need to write the questions.
- Reuse questions from past screeners.
- If you do your own recruiting, ask a coworker to make the calls for you.
- Budget more money than usual when using third party recruiting firms.
- Screen carefully but avoid particularly restrictive screeners.
- Schedule backup participants (“floaters”) to cover multiple time-slots in case a participant fails to arrive.
- Leave time between participants to summarize results and to change the tasks as needed.
You’ve started recruiting. You are familiar with the application type but you need to learn this particular application. You need to write the tasks and questions for your test script and you need to clear them with your client.
With recruiting underway, you can turn your attention to learning the application and writing the test script.
One critical note in our sales guide is to avoid aggressive timelines for applications in unfamiliar domains or those which require extensive training. Even if you’ve followed this advice, you will still likely need to learn certain aspects of the application.
It may seem obvious, but you cannot learn an application if you do not have access to it. Often the clients who come to us are still developing the application, so be sure to schedule adequate access to a stable version.
After you’ve had a chance to learn the application, writing tasks and questions for QTUT is similar to a normal study. We rely on client stakeholders more than usual because they help to prioritize testing needs and to identify key tasks. In QTUT, most clients are releasing soon so they’re more interested in finding easy-fix problems than in statistical data or qualitative feedback. Consequently, we tend to focus our effort on creating tasks vs. writing a lot of questions. You also need to consider what you can realistically summarize quickly. For example, if it takes two hours to compile error rates, then eliminate error tracking from your test plan.
We run pilot tests as early as possible because they help us more rapidly iterate and improve the script. Keep in mind that you probably will not have perfected your script when your first participant arrives, so you may need to modify slightly the tasks and questions between sessions.
- Avoid gathering data that takes a lot of time to compile and analyze.
- Use open-ended questions such as “Tell me about what you did yesterday at work” and allow time for discussion with the participant in order to learn about them during their session.
- For tasks that require more information that you currently know about participants, use a series of questions to build more realistic tasks.
- Use Likert-style scales to get data if you need it, and rely less on task completion data (as your tasks may often change).
Next time: Part II
In Part II, I will discuss how to facilitate the sessions, and I’ll describe a novel approach to analyzing and reporting the results.