# Facebook lookalike seed examples

All companies, counts, receipts and outcomes below are fictional teaching data.

## 1. Coherent paying-customer source with native state unknown

- Modeled outcome: paid annual-plan starts for one product, one geography and an 18-month window.
- Raw contacts: 1,320.
- Eligible after removing 60 duplicate contacts, 35 contacts outside the product or market scope, and 25 opt-outs: 1,200.
- Permission: Approved in the fictional fixture.
- Native state: Unknown because no dated Audiences receipt exists.
- Decision: `plan_ready_native_unknown`. Ask for the receipt; do not invent a match count.

## 2. Mixed CRM source

- Proposed source: customers, students, vendors and unqualified leads in one export.
- Problem: the records do not represent one modeled outcome.
- Decision: Hold. Rewrite the inclusion rule and split unrelated populations before upload.

## 3. Sparse high-value source

- Raw contacts: 87.
- Eligible after exclusions and deduplication: 62.
- Problem: the eligible population is below Meta's currently stated minimum of 100 people for a lookalike source.
- Decision: Hold. Broaden the documented hypothesis or use another source; do not turn raw count into eligible or matched count.

## 4. Fictional interim readout

- Spend cap: $1,500 per arm.
- Outcome maturation: 14 days.
- Acceptable cost supplied by the fictional business: $250 per qualified attended meeting.
- Challenger spend: $1,200.
- Mature qualified attended meetings: 4.
- Pending outcomes: 2.
- Observed mature cost: $300.
- Decision: outcomes pending. Even after maturation, fewer than the fixture's eight mature events per arm would remain `inconclusive`.

Use the completed JSON beside this file for a fully review-ready fictional planning fixture. It proves contract completeness, not account state or performance.