Interview Discussion Guide Workshop

What happens

Agree the top 3–5 questions the interviews must answer.

Facilitator does

Ask the room to write each learning objective in the form "we need to understand whether/how/why..." before any question writing begins. For each candidate, ask: "What decision would we make differently if we knew the answer to this?" Objectives that wouldn't change any decision are nice-to-know questions, not interview priorities. Keep the priority list to three, maximum five.

Participants do

Write learning objectives as genuine unknowns — questions whose answer the team does not already have and that would change what gets built.

Output / signal of success

3–5 prioritised learning objectives are written on the board in the form "we need to understand..." Each one is traceable to a specific product or design decision the team needs to make.

What happens

Build the intro, warm-up, core topic, and wrap-up sequence.

Facilitator does

Structure the flow before writing individual questions. Put the four-part sequence on the board — intro (2 min), warm-up (5 min), core topics (40–50 min), wrap-up (5 min) — and agree how much time each learning objective gets before anyone writes a single question. A guide that front-loads core questions before rapport is built produces defensive, surface-level answers.

Participants do

Draft questions for the sections you have the most context on, then flag gaps for others to fill. Work from the learning objectives, not from the product or the solution you want validated.

Output / signal of success

A skeleton flow is on the board showing the four-part sequence, with time allocations per section and each learning objective mapped to a section. No individual questions yet — structure first.

What happens

Refine question wording to remove bias and add depth probes.

Facilitator does

Read each question aloud and apply two tests: does it contain a word that tells the participant what answer is expected? Does it allow a yes/no response that closes down the conversation? For every closed question, ask the room for the open version. For every leading question, ask what assumption it contains and rewrite it to expose that assumption rather than bury it.

Participants do

Add one specific follow-up probe to each core question — not "tell me more" (too generic) but a specific angle: "Can you walk me through the last time that happened?" or "What did you consider before deciding that?"

Output / signal of success

Every core question is open-ended, non-leading, and has at least one specific probe written next to it. A participant could answer every question without having any idea what answer the interviewer is hoping for.

What happens

Run a mock interview on one section and check guide length against session time.

Facilitator does

Ask one participant to play the interviewer, one to play a participant, and run the warm-up plus one core topic section at full pace. Time it. Multiply the time per section by the number of sections. If the total exceeds the scheduled session length by more than 20%, cut questions — not probes. The questions that get cut first are the ones that address nice-to-know learning objectives, not the priority unknowns.

Participants do

In the mock session, answer naturally and flag any question that feels confusing, double-barreled, or leading. Note where the conversation would naturally go versus where the guide constrains it.

Output / signal of success

The completed guide has been piloted and timed. Its total length matches the available session time. Every question is open, non-leading, and has a probe. The guide can be handed to a new interviewer and run without author explanation.