Usability testing is not what Google Analytics is designed for.
What if you could get your customers to tell you whether they would use what you are developing before you spend resources developing it? What if your customers could tell you what to change or not change before you start making changes? Henry Ford’s ‘faster horse’ quote might be more of a failure of not asking the right questions than it is of not getting the right answers. What if Mr. Ford had asked users what they liked and didn’t like about using a horse for transportation? You might not get ‘a faster horse’ from those questions. All of us have had the sad realization that we wish we would have known something sooner.
Usability testing should be performed during the entire development lifecycle to minimize our regrets about what we didn’t know at any particular time. Usability testing is about giving us confidence in our decisions. Usability testing is designed to tell us what Google doesn’t know about our website.
Usability testing is designed to help us understand what is valuable to our users. It’s a methodology where we share our ideas with a small group of representative users to give us insight about all of our users. It’s where we vet our ideas before committing hundreds of hours of development time towards something that at best will deliver no value to our users and at worst reduce the value. In a single word, it’s where we build the ‘confidence’ that if we build it, they will come.
Usability testing is qualitative research. We usually recruit around 5-7 participants for the testing sessions. Sessions usually consist of a list of tasks for the participants to complete on a computer in under an hour. If the session is unmoderated, no researcher is in the room with the participant or if it’s a remote session, no researcher is on the phone with the participant during the session. Typically, software is used to record the participant’s cursor activity on the screen along with any verbal comments while they are going through the tasks. It’s a cost-effective process if speed of capturing data is more important than allowing researchers to ask follow up questions in a moderated session.
Moderated sessions (where a researcher plays an active role during the session) can be either face-to-face or remote on the phone. While these sessions typically take longer to set-up due to scheduling, the advantage is that the researcher can prompt the participant for more information based on their answers. Both moderated and un-moderated sessions have their place depending on the goals of the testing and the available resources.
In moderated and un-moderated testing sessions participants are told ‘they’ are not being tested and that what is being tested is ‘the usability of our idea and how we are proposing to solution that idea’. Participants are encouraged to ‘think aloud’, meaning, in the most positive sense of the word to ‘ramble out loud’. Let us know your fleeting thoughts as you are accomplishing various tasks with our product. We want to understand things such as:
- What would you expect to happen if you were to click on a particular item?
- Did what happened after you clicked on the item match your expectations?
- Which items on the page are important to you?
- What items are useless to you?
- What is confusing on the page?
- And my favorite ‘if this wasn’t a testing session, would you have already left this website for a competitor’s website’?
Frequently, what participants think is more insightful than what they actually do. Because of this, moderated sessions can provide rich insights into motivation or the all-important ‘why’ participants did what they did. It is critical that whether sessions are moderated or un-moderated that task and interview questions are structured to be neutral. Neutral questions do not give participants the sense that there is a right and a wrong way of doing things. Participants should be reminded throughout the entire process that they cannot do anything wrong and that the development team expects to have made some errors. Usability testing should be structured to discover what the development team was not aware of.
So, to recap. Usability testing measures what your users think about your website which relates to the bigger picture of your brand. Depending on your goals and resources, you can choose between moderated or un-moderated testing. Along those lines, one option could be to perform a limited amount of un-moderated testing early on much like a pilot test. Followed up with moderated testing to delve deeper into the ‘why’ and to confirm or validate what you believe you saw in the un-moderated testing.
Usability testing is a methodology that is useful in the entire lifecycle of any development process.
- Wouldn’t you want to know as early as possible if your users find what you are proposing is valuable to them and easy to use?
- Wouldn’t you want to know if a small tweak in week 4 of the development cycle could increase customer satisfaction by 30% and reduce costs by 15%?
If you only perform usability testing at the end of a project, much like Usability Acceptance Testing (UAT), you lose this opportunity. Plus, the goal of usability testing is to guide development rather than to confirm what you built is doing what you set out to do. It does little good to make the wrong usable things.
Usability testing is designed to tell you what Google wasn’t designed to tell you. Usability testing is your chance to discover the ‘why’ so you have a clear understanding of what you need to do next.