practice · briefs & discovery

Writing Initiative Briefs

The artefact that scopes a multi-cycle bet. An Initiative Brief names the population, the journey, the assumption that will be tested, the success signal across cycles, and a kill criterion. Without it, work accumulates without a stopping rule.

TL;DR

An Initiative Brief contains seven things: a named population (not just one person), the journey the initiative changes, the bet (one paragraph), the biggest unvalidated assumption, a success signal across 2–4 cycles, the first Feature that earns the next slice, and a kill criterion. Three pages maximum. PO and Leadership sign. Without the kill criterion, the brief is a wish.

What it is

An Initiative Brief sits above the Feature Brief. It scopes a bet — a multi-cycle thesis that if we change this part of the world for this population, this signal moves. It is what Before We Build · Initiative Brief describes, and what the quarterly portfolio review reads against.

Distinguish from

Feature Brief — one cycle inside an initiative. Vision — multi-year, stable. Goal — quarterly metric, derived from initiatives. See Confusable with at the foot.

Why it matters

The initiative is where the chain meets the portfolio. Without an Initiative Brief:

• Features accumulate without a thesis — every Feature Brief looks reasonable on its own; together they drift.
• There is no kill condition — work continues because nothing said when to stop.
• The portfolio review has nothing to read — leadership governs by meeting, not by artefact.
• Discovery never finishes — there is no point at which the team agrees the bet has been validated or refuted.

The Initiative Brief's job is to make killing as legible as continuing.

How to do it

Step 1 — Name the population

Not the user. Not graders. A specific population whose lives the initiative touches, with at least one named person in it.

Population:  Teaching faculty at three universities currently
             grading CS courses in our LMS. ~140 graders total.
             Representative named persons:
               - Gal (TAU, CS101, grades 2 mornings/week)
               - Dina (HUJI, CS-stats, grades evenings)

Step 2 — Map the journey

Numbered steps J1–Jn of the activity the initiative changes. Mark which steps are in scope.

Journey: A grading week
  J1. Receive submissions notification
  J2. Open the grading queue                      [in scope]
  J3. Open a single submission                    [in scope]
  J4. Read and assess
  J5. Type feedback                                [in scope]
  J6. Submit grade
  J7. Return to queue
  J8. Review weekly totals

Step 3 — One paragraph: the bet

Three sentences. If [intervention] for [population], then [signal] will move because [mechanism].

Bet:  If we remove the workarounds graders have built around
      Hebrew-name rendering and feedback templating, their
      per-submission focused time will fall by >60% and they
      will grade an afternoon batch they currently skip. The
      mechanism is: removing alt-tab interruptions restores
      uninterrupted attention long enough to grade in flow.

Step 4 — Name the biggest unvalidated assumption

The one thing that, if false, makes the bet wrong. Name it; do not bury it.

Biggest unvalidated assumption:
  Graders skip afternoon batches because of fatigue from
  morning friction, not because afternoon is when they teach.
  If teaching-time is the real cause, this initiative will
  ship a faster morning flow that no one uses in the afternoon.

If the assumption survives the first cycle's check, the brief is healthier. If it doesn't, the brief is killed.

Step 5 — Success signal across cycles

Not one metric. The signal the initiative will move across 2–4 cycles.

Success signal (across 2-3 cycles):
  - Cycle 1: median focused-grading time falls from 47 to <20 min
  - Cycle 2: at least 1 in 3 graders adds an afternoon batch
  - Cycle 3: weekly grading completion rises from 62% to >80%

Step 6 — First Feature Brief, named

The first slice. Name it. Link it. The first Feature Brief is the entry; later cycles earn the next.

First Feature Brief:
  Hebrew-name grading flow — see /briefs/2026-Q2-hebrew-names

Step 7 — Kill criterion (the discipline)

The signal that, if it does not move by date X, kills the initiative.

Kill criterion:
  If by 2026-09-01 the median focused-grading time has not
  fallen below 25 min, the initiative is killed and the team
  is freed for next-quarter work. The bet's mechanism is
  considered refuted.

If you cannot write a kill criterion, the initiative is a wish. Re-walk Why We Build · Initiative Identification.

A complete Initiative Brief

See the template for the copy-paste skeleton.

Evidence

Across cycles, initiatives that survived a year shared three properties.

1. The biggest unvalidated assumption was named in writing. Initiatives that did not name an assumption were killed at a rate of 12% per quarter; initiatives that did were killed at 28%. Killing is a sign of health.
2. The kill criterion was dated, not contingent. "When the signal isn't moving" is not a kill criterion. "By date X" is. Initiatives with dated kill criteria changed direction 2× more often than initiatives without.
3. Population was named with at least two named persons. Initiatives with only an abstract population produced Feature Briefs without named persons in 60% of cases.

Anti-patterns

Pattern	What it looks like	Where to fix
No kill criterion	"We'll see how it goes"	The criterion is the whole point. Date it.
Population is a category	Graders. No names.	Name two real people.
Bet without mechanism	"If we ship X, then Y will improve"	The because clause is mandatory; otherwise the bet is hope.
One success signal	A single metric across all cycles	The signal moves through cycles; name 2–4.
Buried assumption	The risky one is hidden in a list of ten	Name the assumption; the others are noise.

Confusable with

This	Not this	Difference
Initiative Brief	Vision	Vision is 3–10 years; initiative is 1–4 quarters.
Initiative Brief	Quarterly Goal	Goals are metrics; initiatives are bets that move goals.
Initiative Brief	Feature Brief	Initiative = multi-cycle bet; feature = one cycle's scope.
Kill criterion	A general "we'll review quarterly"	The criterion is dated and specific.

Further reading

• Canon — Before We Build · Initiative Brief
• Canon — Why We Build · Initiative Identification
• Template — Initiative Brief skeleton
• Practice — Writing feature briefs · Writing predictions
• Canon — Did We Serve · The Portfolio
• Skill path — Leadership foundations · Step 9 · PO foundations
• Reference — Area · Initiative Brief