Start with the people it’s for: who they are, what hurts, what changes for them, and where they are. Then the smallest thing you can launch to them, and when. Give the first answer, not the best one.
28 blanks. Start anywhere.
A kind of person, in a specific moment. Not “small businesses” — “a solo bookkeeper in March, drowning in client receipts.” Specific enough that you’d recognise one.
List what hurts in that moment, in their words. Then pick ONE: the one that happens most often, hurts most, and that they’re already trying to fix with a workaround. The rest are later, not wrong.
Before → after, in their terms, not your product’s. If you wrote a feature, ask what the feature gets them. This is your promise.
Where they already gather, or go when this pain hits: your own network, a group you’re in, a subreddit, a Slack, a LinkedIn group, a meetup. Pick ONE place to start. If you can’t name three people yet, finding them is your first move.
The smallest thing that gets ONE person from before to after. Then a real date, the hours you’ll actually have before it, and whether those agree. If it doesn’t fit, say so — don’t move the date and don’t quietly shrink it until it does.
Three moves, each with what makes it done. Finding and messaging the people from question 4 is usually the first. Then plan what you’ll learn.
Fill in your moves, the launch date and a check-in date, and your timeline appears here.
Next: four short prompts for what comes after the plan — talking to the first five, the launch message, and the check-in.
Your answers stay in this browser. Nothing is sent anywhere — there's no endpoint for it.
/unstuck launch walks you through the same five as a conversation, in Claude, and keeps going after the plan is written.