Tool
Conversation Starter Picker
Questions that get deeper as you go — a stratified draw, not a flat one.
Conversation Starter Picker
Depth 1 of 5 — small talk. The draw only ever picks from the current tier, then you decide whether to climb.
6 of 6 questions unused at this depth.
Twenty-five questions, sorted into five depth tiers, drawn one tier at a time. You control when to climb, and the draw only ever picks from the tier you are currently on.
This is the one prompt tool on the site that is deliberately not a uniform draw over its whole corpus, and the reason is simple: opening with the depth-five question is a well-known way to end a conversation rather than start one.
It also tracks which questions you have already used at each depth, so a long conversation does not circle back to the same three questions at tier two.
How the Conversation Starter Picker works — and why it's fair
The corpus is stratified: every question carries a depth from one to five, and the pool for any draw is the subset at your current depth, not the full list. Within the tier, the draw is uniform and unbiased — but across tiers it is emphatically not, because tiers you have not selected have a zero percent chance rather than a small one.
That is a stratified draw, and it is genuinely different from the weighted draw the raffle picker uses. Weighting changes the relative odds while leaving everything reachable; stratifying removes options from the sample space entirely until you move the boundary yourself. Neither is more fair than the other — they answer different questions.
Within the current tier the tool tracks what it has already handed you and draws from the unused remainder first, falling back to the full tier only when you have exhausted it. That is the same no-repeat idea the cold-call picker uses, applied per tier rather than across the whole list, so climbing to depth three gives you a fresh unused set rather than continuing a single global cycle.
The depth labels are ordered but not evenly spaced, and it is worth being honest that the ordering is an editorial judgement rather than a measurement. Depth one to two is a small step; depth four to five is a large one, and several depth-five questions are ones many people would not want asked at all. The tool will not climb for you.
Nothing about the ladder is enforced. You can jump straight to five, and the draw will oblige. The design merely makes the default behaviour sensible rather than making the alternative impossible.
When the Conversation Starter Picker is fair — and when it is not
What it does guarantee
- Within the selected tier, every unused question has an equal chance.
- No question repeats within a tier until that tier is exhausted, so a long session keeps producing new material.
- Changing tiers gives you a fresh unused set at the new depth rather than continuing one global cycle.
What it does not
- It is not uniform across the corpus, on purpose. Questions in tiers you have not selected have a zero chance, not a small one.
- The depth assignments are editorial judgement, not measurement. What counts as a personal question varies enormously between people.
- It cannot tell whether the person you are talking to is comfortable. Climbing is a decision for the humans involved, not for the tool.
Two worked examples
Starting a first conversation at depth 1
- Six questions sit at depth 1, so the first draw is 1-in-6 across them.
- The second draw is 1-in-5 from the unused remainder, not 1-in-6 again.
- After six draws the tier is exhausted and the cycle restarts from the full six.
Jumping straight to depth 4
- Five questions sit at depth 4; every question at depths 1 through 3 has a zero percent chance.
- The first draw is 1-in-5 — and one of the five is "What are you avoiding right now?"
- The tool will not warn you. That is the trade for letting you set the boundary yourself.
Where a depth ladder makes the difference
Long car journeys and walks are the natural fit, because both give you the two things the ladder needs: hours rather than minutes, and no obligation to make eye contact while answering. The tiers roughly track the shape those conversations take anyway.
It is also used as a deliberate structure for one-to-one meetings between people who work together but do not know each other — the first three tiers cover almost everything useful there, and the last two would be inappropriate, which the explicit boundary makes easy to respect.
Some people run it in reverse as a wind-down: start at three or four with someone you know well, and drop to one deliberately when it has been enough. Nothing in the tool prevents that, and the tier control makes it a single tap.
For a room rather than a pair, the icebreaker question picker does the time arithmetic that group formats need. For a game rather than a conversation, the would-you-rather generator produces dilemmas that a group can vote on rather than answer.
Frequently asked questions
Why not just draw from all twenty-five questions?
Because a uniform draw would open with a depth-five question about a quarter of the time, and "What do you want that you have never said out loud?" is not a conversation starter. The stratification is the tool.
Does it stop me jumping to depth five immediately?
No. It makes the sensible path the default and leaves the other one available. Nothing here is enforced.
Do questions repeat?
Not within a tier until you have used every question at that depth. Once a tier is exhausted the cycle restarts, which the tool tells you.
Is the memory shared across depth levels?
No. Each tier keeps its own unused set, so moving from two to three gives you a full fresh pool rather than continuing one long cycle.
How were the depths assigned?
By editorial judgement, and the tool says so. There is no measurement behind them, and reasonable people would move several questions up or down a tier.
Can I add my own questions?
No. Each question sits at a specific depth, and an unassigned one would break the tier boundary that the whole tool depends on.
Is this suitable for a work setting?
Tiers one to three, yes. Four and five are not appropriate between colleagues who do not know each other well, and the explicit tier control makes that boundary easy to hold.
Does the tool remember my tier between visits?
No. It starts at depth one every time, which is the right default for a conversation you have not had yet.
What if I want the questions in a fixed order rather than drawn?
Then you want a list, not a picker. The value here is that neither person chose the question, which is what makes an unexpectedly personal one easier to answer.
Related tools
Would You Rather Generator
Draw a dilemma, tally the room's split, and see where the group actually lands.
Icebreaker Question Picker
Tell it your group size and minutes — it does the time maths, then draws.
Truth or Dare Generator
Draw a truth or a dare — two decks, adjustable dare intensity, no repeats in a session.