DESIGN OF ANDROID TV BOX CUSTOMER SEGMENTATION MODEL USING DATA MINING TECHNIQUES
Android TV box is a product of PT X that turns ordinary TVs into Android-based smart TVs. Its users will get interesting content from PT X and Google, one of which is watching movies. This service is offered to its users through a subscription-based rental system on a monthly basis for a fixed fee....
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Subjects: | |
| Online Access: | https://digilib.itb.ac.id/gdl/view/79300 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Android TV box is a product of PT X that turns ordinary TVs into Android-based smart TVs. Its users will get interesting content from PT X and Google, one of which is watching movies. This service is offered to its users through a subscription-based rental system on a monthly basis for a fixed fee. The number of users who revoke the subscriptions continues to increase every month. The rental system makes this problem need to be overcome to prevent losses. Low customer retention rates can be caused by a less personalized customer retention strategy. The current retention strategy is only given to users who will unsubscribe. PT X already has its user segmentation, but still uses an intuitive approach and is not based on historical data.
Recency, frequency, and duration (RFD) variables were used as the basis for segmentation because they are able to reflect user engagement. The models evaluated were agglomerative hierarchical clustering (AHC), k-means, combination, and DBSCAN. The selected segmentation model is a combination of k-means and AHC algorithm. This segmentation model produces three segments, which were used as the basis for designing user retention strategies. |
|---|