BTYD - Pareto NBD
The Pareto/NBD model does not require detailed transaction level data for each customer. It does require some aggregation of the data though, so probably easier to just start with the raw data in the first place.
The Pareto/NBD requires a “customer-by-sufficient-statistic” matrix (CBS), which consists of each customer’s frequency, recency (the time of their last transactions) and total time observed - but the timing of each and every transaction (other than the last) is not needed by the model. https://rdrr.io/cran/BTYD/man/BTYD-package.html
Graph:
- 113.020.050.30 BTYD - Pareto NBD to 113.020.050 Statistics - BTYD models
- 113.020.050 Statistics - BTYD models to 113.020.050.30 BTYD - Pareto NBD
- 113.020.050.10 BTYD - BGBB to 113.020.050.30 BTYD - Pareto NBD
- 113.020.050.20 BTYD - BG NBD to 113.020.050.30 BTYD - Pareto NBD