For a Large Bank
- The Client is one of the leading universal banks serving an ASEAN country.
- It is one of the largest bank in the country in terms of assets.
- The Client wants to group together all it’s retail customers that are most similar in terms of demographics, product holdings, behaviour and performance so that they can curate their product offer strategies for each segment
- Converted customer account level transaction data to customer level data .
- Shortlisted the customers based on a mix of conditions .
- Feature engineering and feature selection techniques were performed to finalize the variables .
- We performed preprocessing and cleaning of the dataset. The unsupervised ML model then clustered the customers into 10 clusters.
- The output provides customer level 10 clusters
- The detailed analytics of each cluster wrt customer product holdings, demographics, digital activity, cards related features etc
- The model output is being consumed by the campaign system for offer differentiation