# Start Here

This system is intentionally larger than an interview cheat sheet. Start with a
decision, not with directory browsing.

## Choose your objective

| Objective | Start | First proof of work |
| --- | --- | --- |
| Learn PM from the beginning | [30-day foundations](../01-roadmaps/01-foundations-30-day.md) | One evidence-backed product decision case |
| Prepare for interviews in 2–8 weeks | [Eight-week sprint](../01-roadmaps/02-interview-sprint-8-week.md) | Baseline mock, error log, and a role-specific drill queue |
| Build durable on-the-job mastery | [26-week mastery](../01-roadmaps/03-mastery-26-week.md) | Weekly artifact plus feedback and revision |
| Diagnose one weak domain | [Competency map](COMPETENCY-MAP.md) | Two fresh timed questions and one domain artifact |
| Practice immediately | [Question-bank guide](../04-question-banks/README.md) | A recorded answer scored before reading the worked answer |
| Compare with a solved example | [Worked answer index](../04-question-banks/WORKED-ANSWER-INDEX.md) | One scored comparison and one corrected section |
| Research a PM topic | [Resource guide](../02-resources/README.md) | Claim/limitation note applied to a decision |

## Fifteen-minute placement check

Without opening an answer, speak for two minutes on each:

1. Design a better way for a family caregiver to coordinate recurring
   medication.
2. Estimate daily digital-payment transactions in a large city.
3. A checkout conversion dashboard drops 15% overnight while revenue is stable.
   What do you do?
4. Define success for a new team-collaboration feature.
5. Decide whether to run an A/B test when treatment changes a shared
   marketplace.
6. Explain a product strategy as choices rather than goals.
7. Prioritize a reliability fix, a customer request, and a growth experiment.
8. Tell a story about changing direction after evidence contradicted you.

For each, score:

- **0:** no usable structure;
- **1:** vocabulary and generic framework;
- **2:** coherent bounded approach;
- **3:** evidence-aware decision with trade-offs and next test;
- **4:** handles second-order effects, operating mechanism, and challenge.

Start with foundations if five or more answers are 0–1. Use the interview sprint
if an interview is scheduled and the median is at least 2. Use a specialization
only after the relevant prerequisite domains reach 2.

## Your first study session

1. Read the [delivery contract](DELIVERY-CONTRACT.md) so “complete” and
   “verified” are unambiguous.
2. Copy the [learning ledger](../08-progress/LEARNING-LEDGER.md).
3. Choose one bank and read only its `BLUEPRINT.md`.
4. Attempt one L1 or L2 question before reading the answer.
5. Score the first attempt with the [universal guide](../03-frameworks/universal-scoring-and-red-lines.md).
6. Read the worked answer and identify the single largest reasoning gap.
7. Re-answer only that section.
8. Schedule fresh retrieval in seven and twenty-one days.

## What not to do

- Do not read all 325 resources before practicing.
- Do not memorize framework acronyms as an answer.
- Do not treat a prompt register as 400 completed answers.
- Do not claim current market facts from memory.
- Do not imitate a case's recommendation without its context.
- Do not score yourself after seeing the canonical answer and call that the
  first attempt.
- Do not optimize interview fluency while ignoring accessibility, safety,
  reliability, ethics, or economics.

## Definition of progress

Progress is observable:

- a fresh answer gets shorter and more decisive without losing reasoning;
- assumptions become explicit and sensitive assumptions are tested;
- metrics gain populations, denominators, guardrails, and decisions;
- a recommendation survives credible alternatives;
- feedback causes a material rewrite;
- the same judgment transfers to a different product context.

The live corpus build state is in [STATUS.md](../08-progress/STATUS.md). It is
not the same as your learning state.