Research for research's sake is an expensive mistake. The method is chosen for the question you need answered — not the other way around. Here are the five core methods and their boundaries.
Depth interviews: the “why”
Interviews answer why people behave the way they do. 8–12 respondents per segment reach saturation: new interviews stop producing new insights. The classic mistake is asking about future behavior (“would you buy…”) instead of past experience (“tell me about the last time you…”).
Usability tests: the “where it breaks”
A test with five users finds about 80% of critical interface problems. Moderated tests give depth; unmoderated give scale and speed. In the checkout relaunch, six test iterations with 48 participants let us ship the redesign without a single rollback.
Diary studies: the “how the scenario lives”
When a scenario stretches over time — choosing a mortgage, planning purchases — a one-off interview fails. A 2–4 week diary shows the real context. In the fashion marketplace search study, the diary stage produced insights impossible to get in a lab.
Surveys: the “how many”
A survey quantifies what interviews discovered, but cannot replace them: without the qualitative stage you do not know which answer options to offer. Rule: first “why” on a small sample, then “how many” on a large one.
Analytics: the “what is happening”
Metrics show what happens but are silent about causes. The funnel registers a payment-step drop — only the user can name the reason. Strong teams close the loop: analytics finds the anomaly, interviews explain it, a test validates the fix.
Not sure which method fits your question? Describe the task — we will suggest a study design before any work starts.