Ten questions, each with its own model.
Each type names the question it answers and how it learns: fixed rules, a statistical fit, or machine learning that checks itself on data it did not see.
- Demand ForecastHow much will we sell next period?Machine learning
- Price ElasticityHow much does volume react to price?Statistical fit
- Price OptimizationWhich price maximises profit?Statistical fit
- Discount DriversWho gets discounted, and where is it too deep?Machine learning
- Cross-Price ElasticityWhich products take sales from each other?Statistical fit
- Promo UpliftDid the promotion move volume?Statistical fit
- RFM SegmentationWho are our best and lapsing customers?No training
- Customer ClusteringWhich customers behave alike?Machine learning
- Transaction ClusteringWhich kinds of deals do we make?Machine learning
- Transaction DriversWhat explains a value on each transaction?Machine learning