# Relationship decision practice

Choose one chart, one caveat and one action. All numerical outputs are fictional. Keep personal responses blank until you practise.

## N01: Compare weekly account-value definitions

Decision question: Weekly distinct accounts, anchored to each milestone time. Full UTC weeks only. Ready/export alternatives do not redefine DB01 or prove audience value. Missing confirmations undercount.

Query: `metabase/relationship_lab/N01.sql`. Chart/table: `lab/relationship/N01.svg`; complete values: `N01.csv`.

PRD mapping: scope, measurement/eligibility, evaluation, rollout and open decisions.

- One takeaway:
- One caveat/counterargument:
- One action:
- Guardrail:
- Next evidence needed:
- 25-second spoken answer:

## N02: Task drivers and review coverage

Decision question: Task fractions use stated denominators. These are candidate drivers, not additive contributions to distinct-account North Star value. Missing reviews are unknown.

Query: `metabase/relationship_lab/N02.sql`. Chart/table: `lab/relationship/N02.svg`; complete values: `N02.csv`.

PRD mapping: scope, measurement/eligibility, evaluation, rollout and open decisions.

- One takeaway:
- One caveat/counterargument:
- One action:
- Guardrail:
- Next evidence needed:
- 25-second spoken answer:

## N03: Explain metric relationships and guardrails

Decision question: Links are labeled identities, descriptive associations or hypotheses. Arrows are not causal weights; selected definition is a practice alternative.

Query: `metabase/relationship_lab/N03.sql`. Chart/table: `lab/relationship/N03.svg`; complete values: `N03.csv`.

PRD mapping: scope, measurement/eligibility, evaluation, rollout and open decisions.

- One takeaway:
- One caveat/counterargument:
- One action:
- Guardrail:
- Next evidence needed:
- 25-second spoken answer:

## J01: Recorded task paths and recovery endpoints

Decision question: Widths are tasks. Events ordered by recorded time; review and export are parallel, not forced into a funnel. Missing events are skipped. Terminal state is not proven abandonment. Before-creation timestamps excluded in J02.

Query: `metabase/relationship_lab/J01.sql`. Chart/table: `lab/relationship/J01.svg`; complete values: `J01.csv`.

PRD mapping: scope, measurement/eligibility, evaluation, rollout and open decisions.

- One takeaway:
- One caveat/counterargument:
- One action:
- Guardrail:
- Next evidence needed:
- 25-second spoken answer:

## J02: Chronology, missing review and recovery audit

Decision question: Inspect audit counts before interpreting flow. Failed attempts can recover within one task. Unreviewed ready output is unknown quality, not rejected output.

Query: `metabase/relationship_lab/J02.sql`. Chart/table: `lab/relationship/J02.svg`; complete values: `J02.csv`.

PRD mapping: scope, measurement/eligibility, evaluation, rollout and open decisions.

- One takeaway:
- One caveat/counterargument:
- One action:
- Guardrail:
- Next evidence needed:
- 25-second spoken answer:

## G01: First-week exposure to mature quality return

Decision question: All mature signups including zero activity. Workflow filters first-week task exposure only; later quality return covers any workflow. Quality early exposure requires review and use before day seven. Association is not causal.

Query: `metabase/relationship_lab/G01.sql`. Chart/table: `lab/relationship/G01.svg`; complete values: `G01.csv`.

PRD mapping: scope, measurement/eligibility, evaluation, rollout and open decisions.

- One takeaway:
- One caveat/counterargument:
- One action:
- Guardrail:
- Next evidence needed:
- 25-second spoken answer:

## G02: Signup cohorts and later quality return

Decision question: Same mature signup-month denominator and fixed day28–41 outcome. Inspect G03 counts. Latest partially mature signup month is not equivalent to a full-month cohort; prediction excludes it.

Query: `metabase/relationship_lab/G02.sql`. Chart/table: `lab/relationship/G02.svg`; complete values: `G02.csv`.

PRD mapping: scope, measurement/eligibility, evaluation, rollout and open decisions.

- One takeaway:
- One caveat/counterargument:
- One action:
- Guardrail:
- Next evidence needed:
- 25-second spoken answer:

## G03: Cohort counts and uncertainty

Decision question: Counts precede rates. Standard error is a descriptive binomial approximation; small groups and clustered team accounts weaken independence. No treatment effect or company baseline.

Query: `metabase/relationship_lab/G03.sql`. Chart/table: `lab/relationship/G03.svg`; complete values: `G03.csv`.

PRD mapping: scope, measurement/eligibility, evaluation, rollout and open decisions.

- One takeaway:
- One caveat/counterargument:
- One action:
- Guardrail:
- Next evidence needed:
- 25-second spoken answer:

## G04: Early task bands versus later return

Decision question: Each point is an activity band, not a person. Aggregation can conceal user differences. Zero-task signups remain. Use counts and persona slices before proposing a test.

Query: `metabase/relationship_lab/G04.sql`. Chart/table: `lab/relationship/G04.svg`; complete values: `G04.csv`.

PRD mapping: scope, measurement/eligibility, evaluation, rollout and open decisions.

- One takeaway:
- One caveat/counterargument:
- One action:
- Guardrail:
- Next evidence needed:
- 25-second spoken answer:

## G05: Weekly early signal and later outcome

Decision question: Both lines use the same mature signup-week users. Shared trends can reflect acquisition mix or seasonality; they do not prove leading-signal causality. Edge signup weeks can be partial.

Query: `metabase/relationship_lab/G05.sql`. Chart/table: `lab/relationship/G05.svg`; complete values: `G05.csv`.

PRD mapping: scope, measurement/eligibility, evaluation, rollout and open decisions.

- One takeaway:
- One caveat/counterargument:
- One action:
- Guardrail:
- Next evidence needed:
- 25-second spoken answer:

## G06: Held-out signal calibration

Decision question: Earlier complete mature signup months train group probabilities; latest complete mature month is held out. Missing train group produces null, not zero. Compare actual held-out outcomes; predictions alone prove no improvement.

Query: `metabase/relationship_lab/G06.sql`. Chart/table: `lab/relationship/G06.svg`; complete values: `G06.csv`.

PRD mapping: scope, measurement/eligibility, evaluation, rollout and open decisions.

- One takeaway:
- One caveat/counterargument:
- One action:
- Guardrail:
- Next evidence needed:
- 25-second spoken answer:

## G07: Held-out prediction versus training base rate

Decision question: Brier error uses actual held-out outcomes. Lower is better. Baseline and signal score exactly the same covered users; positive difference favors the signal. Synthetic exercise, no deployment or causal improvement claim.

Query: `metabase/relationship_lab/G07.sql`. Chart/table: `lab/relationship/G07.svg`; complete values: `G07.csv`.

PRD mapping: scope, measurement/eligibility, evaluation, rollout and open decisions.

- One takeaway:
- One caveat/counterargument:
- One action:
- Guardrail:
- Next evidence needed:
- 25-second spoken answer:

## G08: Inspect mature user evidence rows

Decision question: One row per mature signup; joins retain zero early activity. Workflow filters early exposure, not user eligibility or later any-workflow return. Inspect row evidence before claiming activation predicts renewal.

Query: `metabase/relationship_lab/G08.sql`. Chart/table: `lab/relationship/G08.svg`; complete values: `G08.csv`.

PRD mapping: scope, measurement/eligibility, evaluation, rollout and open decisions.

- One takeaway:
- One caveat/counterargument:
- One action:
- Guardrail:
- Next evidence needed:
- 25-second spoken answer: