Data Dictionary
KPI definitions, sources, calculation methodology, attribution logic, and synthetic data notes.
KPI Definitions
| Term | Definition | Source |
|---|---|---|
| Sell-Through Rate | Tickets sold divided by total capacity, expressed as a percentage. | Ticketing system |
| ROAS | Revenue attributed to marketing divided by marketing spend. | Attribution engine |
| CPA | Marketing spend divided by number of ticket purchases. | Attribution engine |
| CPM | Cost per 1,000 ad impressions. | Ad platforms |
| CTR | Clicks divided by impressions. | Ad platforms |
| Conversion Rate | Purchases divided by sessions. | Web analytics |
| Revenue per Available Seat | Total revenue divided by capacity for a performance. | Ticketing system |
| Assisted Conversion | A touchpoint that contributed to a purchase but was not the last click. | Attribution engine |
| Forecasted Occupancy | Model-projected final sell-through based on current sales pace. | Forecast model |
| Incremental Ticket Sales | Model-estimated tickets attributable to a marketing change. | Forecast model |
| Sales Velocity | Daily tickets sold, typically compared against a rolling 7-day average. | Ticketing system |
| Probability of Sellout | Model-estimated probability that an event will sell 100% of capacity. | Forecast model |
| Attribution Window | Lookback period for crediting a conversion to a marketing touchpoint. | Attribution engine |
Attribution Logic
- Default attribution: last non-direct click with a 30-day lookback
- Comparison models available: first-touch, last-touch, linear, assisted-conversion
- Assisted conversions are counted when a channel touch precedes purchase within window
- Cross-device stitching: probabilistic, using logged-in customer identifiers
Forecast Assumptions
- Baseline: pace-curve regression fit to prior comparable events
- Weekly seasonality, campaign event flags, and inventory tier are model inputs
- Confidence intervals: 80% by default, widened when pace data is sparse
- Refresh frequency: hourly ticketing, daily marketing, weekly attribution
Synthetic Data Methodology
All data on this platform is fictional. Event names, venues, customer counts, transactions, and channel metrics are generated to demonstrate realistic patterns: seasonality, mobile vs desktop conversion gaps, retargeting saturation, budget-constrained search, and weekend production strength. No real customer, company, or transaction information is used.