The Greenlight Funnel
A greenlight board where scripts move through stamped statuses — and the unlinked pile becomes part of the evidence.
Companies with a higher share of late-stage development projects show higher released conversion (lower bound) and stronger released medians.
Question: Do companies with different development-stage portfolios produce measurably different released outcomes?
Method: Normalize messy status strings into canonical stages; compute stage shares per company; summarize released-linked outcomes and unlinked volume.
Prediction: Higher late-stage share associates with higher released-linked medians, but unlinked volume varies by company.
Test: Compare companies stratified by late-stage share and inspect the released-linked medians with uncertainty from missing linkage.
A company’s slate appears as stacks by status — a snapshot of what’s on the table.
The funnel narrows: late-stage stamps crowd the end while early scripts pile high.
Released-linked survivors reveal outcomes, while the unlinked lane reminds us what data can’t see.
- imdb.company_dev_projects
- imdb.companies
- tmdb.movies
- 17_greenlight_funnel.json
- No timestamps: this is a snapshot, not a true funnel over time.
- Many dev projects can’t be joined to released outcomes.
- Status strings are messy; normalization is approximate.