# North America User Growth Operating Playbook

> This is a reusable operating template distilled from historical growth analysis and planning work across creator and commerce products. It does not claim that FYPRO, StoreClaw, or any other named product implemented every process below or achieved any implied result.

## How to Use This Playbook

- Run one market, language, product surface, and cohort definition at a time.
- Replace every bracketed field before an experiment starts.
- Freeze event definitions and exclusion rules before reading results.
- Separate observed facts, proposed actions, shipped changes, and measured outcomes.

## 1. Operating Chain

`ICP + Job-to-be-Done -> Primary Source + Assisted Sources -> Qualified Visit -> Identity + Authorization -> First Core Job -> Usable or Publishable Outcome -> D7 Second Core Action -> Subscription -> First Renewal or Refund -> Attributable Revenue`

GEO is a separate evidence and discovery system. If it sends traffic, record it as a source cohort at `Qualified Visit`; do not merge citation or indexing metrics with product activation metrics.

## 2. ICP Brief

| Field | Fill-in |
|---|---|
| Market and language | [Example: United States / English] |
| Primary user | [Specific role, behavior, and stage] |
| Job-to-be-done | [The concrete task they need to finish today] |
| Trigger | [What happened immediately before they looked for help] |
| Current workaround | [Tools, people, or manual steps used now] |
| Cost of the problem | [Time, money, risk, delay, or lost output] |
| Required first outcome | [A result the user can inspect, save, apply, export, or publish] |
| Proof of value | [Observable behavior, not stated interest] |
| Repeat reason | [Why the same user should return within seven days] |
| Payment reason | [What becomes valuable enough to subscribe to] |
| Disqualifiers | [Who should not be acquired yet] |
| Research evidence | [Interview, session, ticket, query, or cohort source] |
| Evidence date and owner | [Date / owner] |

## 3. First Value Contract

`For [ICP], first value occurs when [required identity or authorization] and [core job] reaches [success state], producing [inspectable outcome] within [fixed window].`

Default activation definition for this playbook:

> A legitimate new account completes required authorization, one successful core job, and a usable proof action such as save, apply, export, or verified publish within 7 x 24 hours of `signup_completed`.

The activation denominator is all legitimate `signup_completed` accounts in the same registration cohort. Exclude internal, test, duplicate, and confirmed fraudulent accounts only through rules frozen before analysis.

## 4. Event Dictionary

| Event | Fire condition | Identity and deduplication | Required properties |
|---|---|---|---|
| `qualified_visit` | Non-internal visit meets the declared engagement rule | Once per `session_id` | `anonymous_id`, source fields, market, language, device |
| `signup_completed` | A real account is created successfully | Once per canonical `account_id` | `user_id`, `account_id`, `signup_at_utc` |
| `authorization_started` | User begins required account or data connection | Once per attempt | `provider`, `attempt_id`, entry point |
| `authorization_succeeded` | Provider confirms usable authorization | Once per successful grant version | `provider`, scopes, `authorized_at_utc` |
| `authorization_failed` | Attempt reaches a terminal failure | Once per failed `attempt_id` | error code, recoverability, device |
| `first_core_job_started` | User submits the first valid core task | Once per `job_id` | job type, input source, experiment ID |
| `core_job_succeeded` | Job reaches the defined successful terminal state | Once per `job_id` | output ID, duration, cost, quality status |
| `core_job_failed` | Job reaches a terminal failure or user-visible timeout | Once per `job_id` | failure reason, cost charged, recovery offered |
| `outcome_saved_or_applied` | User saves, accepts, or applies the result | Once per output and action type | output ID, action type |
| `outcome_exported` | Export completes and a file or destination is available | Once per export ID | output ID, format, destination |
| `outcome_published` | Publish succeeds or a verified content URL is recorded | Once per platform content ID | platform, content ID, publish URL |
| `second_core_action_completed` | A later, distinct core job succeeds | Once per later `job_id` | prior job ID, job type, result status |
| `subscription_started` | First paid charge succeeds | Once per subscription | plan, currency, net amount, term |
| `renewal_due` | The first paid term reaches its scheduled renewal time | Once per subscription | due time, plan, cancellation state |
| `renewal_succeeded` | The first renewal charge succeeds | Once per charge ID | net amount, currency, charge time |
| `subscription_cancelled` | Cancellation becomes effective or scheduled | Once per state change | reason, effective time, source |
| `refund_completed` | Refund is completed, not merely requested | Once per refund ID | amount, reason, related charge |

### Required Common Properties

`event_id`, `event_time_utc`, `user_id`, `account_id`, `anonymous_id`, `market`, `language`, `device`, `product_version`, `primary_source`, `assisted_sources`, `campaign_id`, `creative_id`, `click_id`, `icp_segment`, `first_job_type`, `experiment_id`, `consent_status`.

### Fixed Windows and Derived Metrics

- **Activation within 7 days:** first-value contract completed in `[signup_at, signup_at + 168h)`.
- **D7 second core action:** a distinct successful core job in `[signup_at + 168h, signup_at + 192h)`. Login, visit, open, or unfinished generation does not count.
- **30-day first payment:** first successful paid charge in `[signup_at, signup_at + 720h)`.
- **First renewal rate numerator:** subscriptions with a successful first renewal charge.
- **First renewal rate denominator:** all non-test, non-free, non-lifetime subscriptions whose scheduled first `renewal_due` time is at or before the reporting cutoff, including those canceled, failed, or refunded before that due date. Never divide by all new subscriptions or current active subscribers.

## 5. Identity, Attribution, and Cohort Rules

- Use one canonical `account_id`; map anonymous and device identities into it without rewriting raw events.
- Count an account once in each registration cohort, even if it changes devices or signs in through several providers.
- Assign one `primary_source` using a frozen attribution rule and keep earlier meaningful touches in `assisted_sources`.
- Keep `unknown` as a valid source; do not force unattributed users into Direct or Organic.
- Store raw UTM values, click IDs, referral codes, landing page, market, language, and device.
- Report account cohorts separately from sessions, users, organizations, and subscriptions.
- Record the timezone and currency at the top of every scoreboard.

## 6. Mature Cohort Scoreboard

| Metric | Numerator | Denominator | Cohort maturity |
|---|---|---|---|
| Qualified visit rate | Qualified sessions | Eligible sessions | Session window closed plus ingestion delay |
| Signup rate | Legitimate signups | Qualified visits | Cohort end plus ingestion delay |
| 7-day activation | Accounts completing first value before 168h | Legitimate signup accounts | Last signup in cohort plus 168h and ingestion delay |
| D7 second action | Accounts with a distinct success in the fixed D7 window | Legitimate signup accounts | Last signup plus 192h and ingestion delay |
| 30-day first payment | Accounts with first paid charge before 720h | Legitimate signup accounts | Last signup plus 720h and billing delay |
| First renewal | Successful first renewals | Subscriptions whose first renewal due date has passed | Latest due date plus billing settlement delay |
| Refund rate | Distinct settled charges with one or more completed refunds; report refunded amount and net revenue separately | Settled paid charges | Refund observation window closed |
| Activated CAC | Eligible channel spend | Activated accounts | Activation cohort mature |
| Paid CAC | Eligible channel spend | First-paid accounts | 30-day payment cohort mature |

Scoreboard header:

```text
Cohort: [start UTC] to [end UTC]
Market / language / device: [value]
Primary source / campaign / creative: [value]
Data delay assumption: [value]
Exclusion rule version: [value]
Maturity: MATURE / PENDING
Owner: [name]
```

## 7. Channel Experiment Card

```text
Experiment ID:
Business problem and funnel gate:
ICP and market:
Primary source and assisted-source rule:
Baseline cohort and maturity:
Evidence behind the hypothesis:
Single variable being changed:
Control and treatment:
Primary metric:
Guardrails: quality / cost / complaints / refunds / privacy
Sample and fixed observation window:
Instrumentation check:
Owner and launch date:
Rollback condition:
Decision date after cohort maturity:
```

Do not compare channels on signup cost alone. The same experiment must preserve source, landing page, first job, activation, D7, payment, renewal, and refund visibility.

## 8. Behavior-Triggered Lifecycle

| User state | Trigger | Next action | Success metric | Guardrail |
|---|---|---|---|---|
| Authorization incomplete | Failure or no success after declared interval | Show value preview, permission reason, and recovery path | Authorization success | Complaints, repeated prompts |
| Core job failed | Terminal failure | Resume, retry, credit protection, or human help | Successful first job | Cost and abuse |
| Outcome ready but unused | Success without save, apply, export, or publish | Explain one proof action and remove the next obstacle | First-value completion | Message fatigue |
| Activated but no second action | Activation followed by D7 inactivity | Return with context, next idea, or reusable template | D7 second action | Unsubscribe rate |
| Repeated use but unpaid | Repeated core outcomes without payment | Explain plan fit after demonstrated value | 30-day first payment | Refunds and trust |
| First term nearing due | Renewal due approaching | Summarize delivered value and unresolved problems | First renewal | Cancellation and complaints |
| Core contributor | Repeated outcomes and specific feedback | Invite co-creation, case permission, or referral | Qualified referred activation | Incentive fraud |

## 9. Roles and Responsibilities

| Role | Accountable for |
|---|---|
| Growth Lead | One weekly objective, priority, resources, and final Scale/Iterate/Stop decision |
| Product Growth | Authorization, first value, and D7 repeat path |
| Data / Engineering | Events, identity, attribution, experiment assignment, and data quality |
| Paid Growth | Activated CAC and mature paid CAC, not signup CPA alone |
| Creative / Content | Hooks, landing-page promise, UGC, and evidence-led cases |
| Lifecycle / CRM | Behavior triggers, D7, first payment, and first renewal |
| Customer or Creator Success | Failed-user observation, core-user support, feedback, and case permission |
| Monetization / Finance | Settled revenue, refunds, contribution margin, plan, and renewal logic |

In a small team, one person may own multiple rows. Write every row beside a named owner anyway, and keep Data, billing, and release checks independent enough to challenge a growth decision.

## 10. Monday-to-Friday Delivery Rhythm

| Day | Decision work | Required deliverable |
|---|---|---|
| Monday | Read mature cohorts, failures, revenue, and data health | Growth Health Check and one Weekly Objective |
| Tuesday | Select one bottleneck and design one-variable tests | Approved Experiment Card and instrumentation checklist |
| Wednesday | Observe failed paths and speak with affected users | Five evidence notes and updated failure taxonomy |
| Thursday | Ship the smallest product, page, creative, or lifecycle change | Release ID, QA evidence, rollback owner |
| Friday | Wait for eligible evidence and choose the next action | Decision Log: Scale, Iterate, or Stop |

Daily checks: authorization, job success, export or publish, billing, attribution, fraud, complaints, and event delivery.

## 11. Decision Log

```text
Decision: SCALE / ITERATE / STOP
Experiment ID and version:
User and scenario:
One change tested:
Mature cohort and comparison:
Primary metric result:
Guardrail results:
Data-quality status:
What the evidence supports:
What remains unknown:
Next owner / date / release / spend authority:
```

## 12. Scale, Iterate, or Stop Gates

### SCALE

- Identity, attribution, billing, and core events reconcile within the declared tolerance.
- The primary metric improves in at least two complete, comparable cohorts.
- Required activation, D7, payment, and renewal cohorts are mature; quality, cost, complaints, refunds, privacy, and fraud remain inside approved guardrails.

### ITERATE

- The hypothesis remains plausible, but the cohort is immature, the sample is weak, or one controllable friction remains.
- The next test changes one variable and keeps the same ICP, success definition, and attribution rules.
- No severe trust, safety, billing, or data-quality issue is present.

### STOP

- Data cannot be reconciled or the event definition changed during the test.
- The channel produces registrations without first value or repeat use.
- The metric improves only by worsening a guardrail, or the product cannot reliably deliver the promised first outcome.
- A cohort remains low-fit after source and landing-page correction.

## 13. Public Evidence Rules

- Publish a result only when the baseline, change, comparison, date, and post-change outcome are documented.
- Label analysis, recommendation, implementation, launch, and measured result as different states.
- Do not expose customer identities, private account data, internal budgets, or unreconciled revenue.
- Do not turn another team's product total into an individual's acquisition or revenue result.
- If implementation or outcome evidence is missing, publish the method and the evidence boundary instead.