1. Pattern Discovery
- Fan segmentation models including preferences (price, seating, opposition, number of games) and overall value (similar to an RFM).
- Association analysis including:
- Market basket analysis: this model identifies what games go together for fans with a view to tailoring season ticket packages to the right offering of matches.
- Sequence analysis: This model can help enhance the knowledge of customer behavior by knowing what their likely next move is. For instance, the team using a time identifier can track how fans have gone through the system i.e. fans who start as 3 game holder move to 15 games then Class A season before going Platinum.
- Novelty detection: This model is produced using historical fan behavior. It makes it easy to identify particular behavior that indicates what has happened is out of the ordinary. For instance, it can help identify when a fan has statistically exceeded their time between attending matches.
2. Forecasting Models
- These models forecast anything and everything that might be of use to the team. Subjects like crowd number; food sales etc….these forecasting models can be further enhanced by introducing exogenous variables/series e.g. weather, opposition win pct, team win pct, and so on.
- If the data is available, these models could be enhanced by including: game spend on F&B, merchandise, as well as other purchases through the online merch store and physical merch store(s).
3. Prediction Models
- Fan churn model for likelihood of ticket holders not renewing
- Fan propensity model to predict response to marketing campaign
- Predicting season value of a fan spend within the venue as a dollar figure