Что происходит после декомпозиции задачи?

После декомпозиции: от задач к выполнению

Краткий ответ

Декомпозиция (разбиение большой задачи на маленькие) это 10% работы.

Оставшиеся 90% это:

1. Prioritize (какие first?)
2. Assign (кто делает?)
3. Execute (что мешает?)
4. Monitor (прогресс ok?)
5. Adjust (план меняется или нет?)

Сценарий

Допустим я разложил большую инициативу на подзадачи.

Инициатива: "Improve retention от 60% к 75%"

Decomposition результат:

Epic 1: Improve onboarding
  - Task 1.1: Identify churn points (research)
  - Task 1.2: Design new onboarding (design)
  - Task 1.3: Implement tutorial (engineering)
  - Task 1.4: Test & optimize (PM + engineering)

Epic 2: Build engagement features
  - Task 2.1: Customer interviews (PM)
  - Task 2.2: Design notification system (design)
  - Task 2.3: Implement notifications (engineering)
  - Task 2.4: A/B test (PM + data)

Epic 3: Improve product support
  - Task 3.1: In-app help system (design + eng)
  - Task 3.2: FAQ documentation (PM)
  - Task 3.3: Chat support (ops)

Теперь у меня есть понимание scope. Далее:

Этап 1: Prioritization

Я НЕ делаю все одновременно.

Я спрашиваю:

1. Какой epic даст наибольший impact?
2. Какой quickest to execute?
3. Какой разблокирует others?

Анализ:

Решение:

• Week 1-3: Epic 1 (onboarding) + Epic 3 (support) parallel
• Week 4-7: Epic 2 (engagement) + optimize Epic 1

Этап 2: Assignment

Я assign tasks с учетом:

1. Skillset (who is best for this?)
2. Capacity (do they have time?)
3. Growth (is this stretch for them?)
4. Load-balance (don't overload one person)

Пример assignments:

Task 1.1 (Identify churn points):

• Owner: Junior PM (stretch assignment, she learns research)
• Support: Me (weekly syncs, feedback)
• Effort: 1 week
• Outcome: Report with 3 main churn reasons

Task 1.2 (Design new onboarding):

• Owner: Design lead (her strength)
• Support: Junior PM (research findings feed design)
• Effort: 1.5 weeks
• Outcome: 3 design variants

Task 1.3 (Implement tutorial):

• Owner: Senior engineer (complex, needs expertise)
• Support: Design (UX feedback), PM (requirements)
• Effort: 2 weeks
• Outcome: Working tutorial in production

Task 3.1 (In-app help system):

• Owner: Junior engineer (good stretch project)
• Support: Design (UI), PM (what goes in help)
• Effort: 2 weeks

Этап 3: Execution tracking

Я don't micromanage, но I monitor.

Daily standup (15 min):

Each person answers:
1. What did I finish yesterday?
2. What am I doing today?
3. Any blockers?

Example:
Junior PM: "Finished 5 customer interviews. Found 3 churn patterns. No blockers."
Design: "Working on variant 2 of onboarding. Need requirements from PM on [X]."
Engineer: "Implemented 60% of tutorial. Blocked on design decision: [specific question]."

My response: "Design, can you decide by EOD? That unblocks engineering."

Этап 4: Risk management

Я identify risks EARLY.

Common risks:

Risk 1: Design finishes late
  → Impact: Blocks engineering
  → Mitigation: Design starts week 1, share draft by day 3

Risk 2: Engineering underestimates effort
  → Impact: Slips deadline
  → Mitigation: 1.5x buffer in estimates. If done early, great

Risk 3: Customer research finds unexpected thing
  → Impact: Changes prioritization
  → Mitigation: Plan for pivot (have plan B)

Risk 4: One person unavailable (illness, leave)
  → Impact: Task blocks
  → Mitigation: Cross-train backups. Junior PM knows design basics

Этап 5: Weekly sync (deep dive)

Every Friday, 1 hour with team:

Agenda:
1. Metrics (are we on track toward 75% retention?)
   - Current: 60%
   - Target: 75%
   - We should see early movement from onboarding improvements

2. Progress (roadmap tracking)
   - Onboarding: 40% complete (on track)
   - Support: 100% complete (ahead!)
   - Engagement: Waiting for onboarding to finish

3. Issues (what's slow, what's stuck)
   - Design decision for Epic 2 pending (I resolve today)
   - Engineer needs clarification on requirements (PM explains now)

4. Learnings (what we learned this week)
   - Customers love new onboarding (NPS +5 in early feedback!)
   - Support feature less used than expected (adjust strategy)
   - Tutorial too long, people skip it (need to shorten)

5. Adjustments (plan changes?)
   - Onboarding working, but tutorial needs revision
   - Decision: Polish tutorial before full rollout
   - Delay full onboarding launch 1 week (worth it)

Этап 6: Course correction

Things change. Plans change.

Scenario: New competitor launches similar feature

CEO: "We need to respond. Pause retention work?"

My analysis:
- Retention work 50% done, blocks engagement feature
- Pausing now = waste what we've done
- My recommendation: Finish retention (2 more weeks) + quick competitive feature

Decision:
- Continue onboarding improvements (2 weeks)
- Start competitive response (1 week, parallel)
- Delay engagement feature 1 week

New plan: Still hit 75% retention target, but also have response to competitor.

Этап 7: Launch & measurement

После execution: Launch

Phase 1: Soft launch (10% of users)
- New onboarding for 10% of signups
- Monitor: Activation rate, time to activation, NPS
- Early result: 58% activation (vs 40% before)
- Issue: Tutorial too long (people skip)
- Fix: Shorten tutorial by 50%

Phase 2: Wider rollout (50% of users)
- Applied fix from Phase 1
- Monitor: Same metrics
- Result: 70% activation (great!)

Phase 3: Full rollout (100% of users)
- Roll out to all
- Continue monitoring
- Week 4 result: 75% retention target HIT

Этап 8: Retrospective & learning

После launch: What we learned

What went well:
- Onboarding improvement exceeded expectations (75% vs 60% target)
- Support feature adopted by 40% (nice surprise)
- Team execution smooth (good assignment strategy)
- Risk management helped (anticipated design delay, prepared)

What was harder:
- Tutorial too complex initially (wasted time on revision)
- Engagement feature delayed (worth it? Could have done parallel)
- Engineering estimates off (usually underestimate)

Learnings for next project:
- User test tutorial before launch (save iteration time)
- More parallel work possible (we were too conservative)
- Engineer estimates: 1.5x not enough, need 1.75x

Metrics success:
- Retention: 75% (goal achieved!)
- Churn reduction: -15% (biggest win)
- User satisfaction: NPS +12 (people love onboarding)
- Revenue impact: +10% (better retention = higher LTV)

Распространенные ошибки

❌ Mistake 1: Decompose but don't prioritize

• 20 tasks, no clear priority
• Team doesn't know what first
• Result: Chaos, no progress

✓ Fix: Explicit prioritization. What's most impactful? Do that first.

❌ Mistake 2: Set tasks and forget them

• Assign tasks, assume they'll be done
• Not tracking progress
• Surprised by delays in last week

✓ Fix: Weekly sync, daily standup. Early visibility = early fixing.

❌ Mistake 3: No risk planning

• Assume everything goes to plan
• One blocker halts entire project
• Miss deadline

✓ Fix: Identify risks upfront. Plan mitigation.

❌ Mistake 4: Ignore learnings

• Project ends, move to next
• Never change process
• Make same mistakes

✓ Fix: Retrospective every project. Capture learnings. Apply next time.

❌ Mistake 5: Too rigid

• Plan says do X
• Reality shows Y is better
• Stick to plan anyway

✓ Fix: Be adaptive. Plans are hypotheses, not dogma.

Framework: Decompose → Prioritize → Assign → Execute → Measure → Learn

Это не linear. It's cyclic:

Large task
    ↓
Decompose (break into subtasks)
    ↓
Prioritize (which first?)
    ↓
Assign (who does what?)
    ↓
Execute & Monitor (daily standup, weekly sync)
    ↓
Adjust (plans change, be adaptive)
    ↓
Measure (did we achieve goal?)
    ↓
Learn (what worked, what didn't?)
    ↓
Apply (improve process for next task)
    ↓
Repeat for next task

Итоговый совет

Декомпозиция это техничеческая: как разбить задачу?

Но реальная сложность это управление после разбиения:

• Keeping people aligned (not just assigned)
• Unblocking obstacles (your job, not their job)
• Adjusting to reality (not sticking to broken plan)
• Learning from each cycle (continuous improvement)

Лучшие PMs не те, кто разбивают задачи идеально. Лучшие PMs это те, кто deliver despite uncertainty, obstacles, and change.

Тот же навык: Adaptability.