Case: the brief for the physics launch

Read any draft brief twice: once as a family would (does anything here say what changes for me?) and once as the person who must refuse extra work in December (what can I point to?). Leo's draft fails both readings in three places.

1. Outputs as the goal. Goal and success list the curriculum, tutors, test, page and booking. Every line could be true on 1 February with zero bookings: the trap from Goal or outputs.
2. No out list. “Everything Vira and Danylo listed” invites the mobile app or chemistry to creep in later.
3. No single owner. “The physics team” means nobody in particular answers for the launch.

What the draft gets right is worth keeping: the fixed date set by the exam season, Hanna's 600,000 UAH ceiling and the one-page length. Fix the gaps; don't throw away the limits.

A brief reads top to bottom because each part serves the ones above it:

1. Goal: the change the project exists for (families preparing for the spring exam book physics with Klasna).
2. Success criteria: how you'll know it worked, delivery and result.
3. Scope: what work gets you there (in) and what doesn't (out).
4. Limits: what bounds that work: the fixed date and the budget ceiling.
5. Assumptions: what the plan bets on inside those limits, each with a check and a date.
6. Owner and sponsor's sign-off: who answers for it and who agreed.

Writing in this order is also a thinking aid. You can't judge whether the diagnostic test belongs in scope until you know the goal; you can't tell whether a limit hurts until you know the scope. A brief that starts with the task list has skipped the reason for every task.

A budget ceiling is a limit: something the project can't change by itself, like the fixed date. But a limit is only useful if everyone reads it the same way, so the brief says what it covers and what it doesn't.

Hanna's 600,000 UAH pays for building the launch: the curriculum, practice sets, tutor selection, the diagnostic test, the landing page and booking by subject. It doesn't pay tutors for delivered lessons after opening: those 430 UAH per lesson come from operations, because teaching physics after 1 February is operational work (Project, operations or product work?).

Without that sentence, February payouts could be counted against the ceiling and the launch would look over budget the moment it succeeds, or someone might spend the build budget on lessons. Hanna asked for it directly: “say what isn't paid from it.”

Lines from a plan often arrive mixed together. Sort each by asking what kind of statement it is.

• Limit: the project can't change it by itself. Classes open on 1 February 2027, set by the exam season.
• Assumption: the plan believes it, but nobody has checked. “Each of the 8 tutors has room for about 15 new learners in February” (120 ÷ 8 = 15) is hidden inside two other numbers, and no candidate has been asked.
• Risk: an event that may happen and would cost days or money. Vira's review may run late and push back the demo lessons.
• Success criterion (result): what must change after opening. 120 learners booked between 1 and 28 February.

The difference between an assumption and a risk is subtle but useful: an assumption is something you believe is true now; a risk is something that may happen later. Name only the main one or two risks in the brief; the full register comes later in the course.

Some assumptions are never written as sentences: they sit between two numbers that each look reasonable on their own. Divide, multiply or compare the plan's numbers and check whether the result is something anyone has confirmed.

The physics plan wants 120 learners booked in February and at least 8 tutors. Together they mean 120 ÷ 8 = 15 new learners per tutor in one month. Strong candidates teach at school and are free only evenings and weekends, so 15 new learners each is a real question. Nobody has asked them.

Once spotted, the hidden number becomes an ordinary testable assumption: “each tutor has room for about 15 new learners in February; Roman asks candidates for their free weekly slots when booking demo lessons, in November.” If the answer is “eight at most”, the plan needs more tutors or a lower target, and October or November is the cheap time to learn that.

A risk is an event that may happen and would cost the project days or money. The brief should show the main ones, because the sponsor signs knowing them, but it is not a risk register.

For physics, one line is enough: Vira's curriculum review may run late, and tutors can't prepare demo lessons on unfinished material. Maybe a second: tutors found too late. That tells Leo where the plan is fragile without turning the one-page agreement into a list of forty worries.

Why hold back? Because a long list on the brief crowds out what the signers must agree on (goal, success, scope, limits) and gives every risk the same weight. The full register, with probability, impact, an owner and a trigger for each risk, comes later in the course, when the plan exists and there's something concrete to protect.

Anything not on the in list is out until someone decides otherwise: the rule behind the out list from The one-page brief. In the physics brief, chemistry, a mobile app and group lessons are named as out. Chemistry looks like “the same work, twice”, but it means a second curriculum, a second tutor pool and a second test against the same date and the same 600,000 UAH.

A signed brief doesn't forbid change; it makes change visible. When marketing asks three weeks later for a free trial lesson for every visitor, you neither add it quietly nor refuse because “the brief is signed”. You treat it as a change: show its cost in days, money and tutor hours, and let Leo decide what comes out or which limit moves.

The quiet “yes” is how scope creep happens: the date and budget stay, the work grows, and the delay shows up in January.

“Families want physics” is a hope until it has a check. Using the three parts from The one-page brief (what you'd see, how you'll check, by when), turn the plan's biggest bet into something that can fail before most of the budget is spent.

Danylo's idea does it: put the landing page up in December with a waiting list, and on 20 December compare sign-ups with the 120-learner target. If sign-ups are thin, there's still time to change the marketing plan or the target. Counting bookings after 1 February is the result check, too late to steer anything; asking sales for their opinion gives no number.

How it fits together: the brief names the change, checks it in two parts, fences scope with an out list, states its limits and tests its bets early, with one owner and the sponsor's signature. Module 2 breaks the in list into deliverables someone can accept.