Release Features                            Date: December 16, 2020


We have created Databricks notebooks to achieve some requirements that cannot be achieved on Insights+. 

Following are the details:

Event Communication Billing

Previously CS team and finance team members were not able to compute the used credits during event communications such as loyalty program messages, loyalty promotion messages, and DVS (Dynamic Voucher System) messages on Insights+ or through Insights+ backend using Databricks. 

To resolve this barrier, we have introduced a notebook-driven solution to address this requirement. A self-explanatory Databricks notebook with sample codes is mentioned here (Databricks - APAC2 cluster notebook). 

You can clone or download the notebook to make changes (some caveats are mentioned in the notebook for your reference), and then run or schedule.
Relevant Tickets: CAP-55243, CAP-42641, CAP-48732

PMA Source Data Access on Databricks

A few CSM team members have requested access to source tables (previously accessed via PMA) for working with data that is not available in Insights+ backend tables (accessed via Databricks). 

To cater to such requests, we have now made source data tables available on-demand on Databricks for reporting purposes. The steps to be followed for getting access to source tables on Databricks are here.

Relevant Tickets: CAP-45946, CAP-55805

Points Liability Export

CS team members have mentioned that some clients review Points Liability data on a recurring basis. Consequently, we have built a notebook where you can export Points Liability at a user_id-level and also on a particulate date (in the past or as of today). Similar to the Event Communication Billing notebook mentioned above, clone/download the notebook and change it/create schedules according to your requirements. You can find the notebook is here.

Relevant ticket: AI-8391

Historical Activity Segment Updation

In Insights+ using the User Segments feature, it is possible to classify customers into various activity segments (active, lapsed, dormant, and so on) based on their days since the last transaction. It is also possible to track customers’ activity segments on a recurring basis in the future using the SCD option. However, we cannot utilize User Segments to go back to a specific point in time and tag a customer to an activity segment. 

To enable classification of customers based on their activity segment of a specific date from the past, we have built a notebook-driven solution that allows you to

  1. Define the names of various activity sub-segments based on days since the last transaction of customers.
  2. Mention the date in the past on which each customer’s days since the last transaction will be computed. Accordingly, their activity segment will be assigned.
  3. To update an existing segment (or create a new one) with a historical activity segment entry, create a CSV file with user_id and activity segment value that can be directly uploaded using Insights+.

You can find the notebook here. For additional information on User Segments usage along with this notebook, see here.

Relevant ticket: CAP-13976

Recency-Frequency-Monetary (RFM) Segmentation

Several brands use RFM segmentation to understand their customer base and also target specific sub-segments of customers based on where they rank in terms of RFM attributes. We have recently come across requests where analysts mentioned that RFM segments could be made available as a part of Insights+ User Segments.

To facilitate this, the Data Science team has created a notebook that will help you understand your brand’s customers using various measures and then classify them into High, Mid, and Base segments based on their RFM attributes. Also, the notebook allows you to export user_id and corresponding segments for utilization in the User Segments module  (similar to the activity segment notebook mentioned above).

The notebook can be found here

Relevant ticket: CAP-13976