Interviewing someone well

An interview in which the interviewer works through their list is a questionnaire delivered aloud, and it produces answers rather than information.

Preparation

What matters before.

Research the person properly, which means going beyond the obvious sources so your questions are not ones they have answered a hundred times.

Decide what you actually want to find out, which is different from a list of questions.

Write more questions than you will use.

Order them roughly, easiest first.

And then be prepared to abandon the order entirely, which is the point of having prepared it.

Opening

Which sets whether you get anything useful.

Explain the format and the length, which reduces uncertainty.

Start with something easy and concrete.

Do not start with the hardest question, since rapport has not been established and the answer will be guarded.

And show you have done the research early, which changes how seriously the interviewee treats the conversation.

The question types

Which produce different things.

Open questions produce narrative and are the default.

Closed questions confirm facts and should be used sparingly.

"Tell me about a time when" produces specifics rather than generalities, which is why it dominates structured job interviews.

"What happened next" keeps a narrative moving.

"Why" produces motivation and can feel accusatory, so it is often better as "what led to that".

And "what am I missing" near the end, which frequently produces the best material in the whole interview.

The follow-up

Which is where the actual information is.

The prepared question produces the prepared answer; the follow-up produces something new.

Which means listening to the answer rather than reading the next question.

Ask about the specific thing they mentioned in passing.

Ask for an example when they generalise.

Ask for the detail when they summarise.

And notice what they avoided, which is information.

Silence

The most underused tool.

After an answer ends, wait.

People frequently continue, and the continuation is generally more candid than the first part.

Three or four seconds is enough and feels considerably longer.

Which requires resisting the reflex to fill it, and which is the single change that most improves an interviewer.

Difficult subjects

Handling them.

Place them in the middle rather than the start or the end.

Signal them: "I want to ask about something difficult."

Ask directly rather than circling, which is more respectful than it sounds.

Accept a refusal without pressing, and move on cleanly.

And do not fill an emotional silence with reassurance, which cuts off what they were about to say.

Job interviews specifically

Where the stakes and constraints differ.

Structured interviews — the same questions, in the same order, scored against defined criteria — predict job performance substantially better than unstructured conversations, which is one of the more robust findings in the field.

Which means resisting the temptation to have a free-flowing chat, however pleasant.

Ask about past behaviour rather than hypotheticals.

Take notes and score afterwards.

And be aware that liking someone is not evidence of competence, which is what unstructured interviews mostly measure.

Recording and notes

Practical.

Ask permission to record, always.

Record even if taking notes, since notes miss the exact words.

Take sparse notes — timestamps and prompts — rather than transcribing, since writing prevents listening.

Note your own observations separately from what was said.

And write up the impressions within the hour, since they degrade fast.

Closing

The last few minutes.

"Is there anything I should have asked?"

"Is there anything you want to add?"

Confirm anything you will quote or use.

Explain what happens next and when.

And thank them properly, since they gave you their time and generally something more than that.

Warming someone up

The part that determines how candid the rest is.

People give guarded answers to interviewers they have not decided about, which means the first few minutes are spent establishing what kind of conversation this is.

What helps: a specific reference to their work that shows you read it properly; an admission of what you do not understand; and a first question that is genuinely easy to answer.

What does not: a difficult question early, a question they have answered a hundred times, or a preamble about yourself.

When the answers are too short

A common problem with reticent interviewees.

Do not respond by asking more questions faster, which produces an interrogation and shorter answers still.

Slow down. Wait after each answer.

Ask for a specific instance rather than a general view, since concrete questions are easier to answer than abstract ones.

Offer something yourself, which changes the exchange from extraction to conversation.

And where someone is genuinely reluctant, ask what they would prefer to talk about, which occasionally produces the whole interview.