Every design decision eventually leans on research, and all research comes in two flavours. Quantitative methods count things: how many users clicked, how long tasks took, what percentage gave up. They scale to millions and produce hard numbers. Qualitative methods ask things: interviews, observation, watching one person struggle and letting them talk. They do not scale at all, and they are the only way to learn why anything happens.
Numbers are objective but context-blind: they tell you 40 per cent of users quit at step three, and nothing about what confused them. Conversations are rich but small and easily biased: people are polite, people misremember, and five interviews are not the world. Professionals use both and know exactly which question each can answer.
Now the trap, from the same car marketplace as chapter 2. That team did do research. Their A/B tests showed promising click numbers. They invited users into the office, and the feedback was positive. Both methods said go, and the product still died. What went wrong is the most instructive research failure there is: both methods were pointed at the wrong question. They tested how well the feature worked. The unanswered question was who actually has this problem and will they pay, and no amount of well-executed testing on the wrong question can answer the right one. Bonus trap inside the trap: people invited into a company's office to review its product tend to be polite, a bias with its own name in the textbooks.
The question comes first
So the discipline's real skill is not running methods, it is matching method to question. Discovering who your users are is qualitative territory. Measuring whether a change helped is quantitative territory. Ethics runs through all of it, because modern products experiment on their users constantly, mostly without meaningful consent, and a decent research culture is the difference between learning from users and farming them.
Run both methods on one question. Post a poll in a group chat: "which app do you all waste the most time in?" That is your quantitative data. Then ask one person from the chat a single why question about their answer and just listen. Compare what each kind of answer is actually worth, and notice which one surprised you.
Rogers, Sharp and Preece, Interaction Design, is the standard university textbook, and its data-gathering chapters cover every method here. Mike Kuniavsky, Observing the User Experience, is the practitioner's version.