Opt in to get the latest updates.
Propensity modelling helped our client use customer segmentation in cross selling
Client is a US-based leading publisher with print and digital media properties.
- Customer level segmentation and analytics performed on ad hoc basis with smaller segments of customer data for critical campaigns
- Client wants to understand which section of customers are likely to move up/down in the conversion funnel and how to improve customer loyalty using promo codes and targeted campaigns
- Lacking intelligence to identify customer segments to roll out promotions
- Lacked a robust data management platform for streaming data, harmonization and enabling consumption of insights
- Narrow focus on data-driven decisions along with balkanized data strategies stand as obstacles for analytical maturity
- Multi-channel customer touchpoints were analyzed using website traffic data streamed into Snowflake and enabled integrated with secondary datasets at SKU level
- Creating Customer 360 view consisting of website behavior, past purchase data, and current subscriptions and/or newsletter sign-ups
- Snowflake was also used for hosting dev/test environments for sampling
- Using python, data connection was established with Snowflake to access the customer journey data
- XGBoost is used to train models to generate propensity scores for customers
- Segments are created based on scores for customers
- Client able to generate insights on user behavior for upsell/cross-sell opportunities using easily accessible data platform
- Deep dive analysis based on integrated data helped them optimize the marketing budget and reduce cost per conversion
- Campaign performance analysis helped in getting insights on the effectiveness of campaigns based on propensity and conversions
- Success of new platform influenced leadership to migrate legacy databases to snowflake thereby breaking data silos and drive decision-based data strategy and analytics maturity
Want to learn more about our expertise, let’s connect!