HOMER unifies student data on Databricks to personalize early childhood learning

Overview

HOMER, an early childhood education company, built a unified data lake on Databricks in six months (versus up to two years on other platforms) to consolidate hundreds of millions of monthly event transactions, using Delta Lake for ETL pipelines and MLflow to build churn and engagement models in about six weeks, driving a 1%-5% increase in conversions with an estimated revenue impact of up to $20 million.

The challenge

HOMER had vast amounts of digitally native product data that was disorganized, disparate and inaccessible, with no unified foundation to power a personalized learning experience; teams also lacked a way to collaborate, leading to duplicated work on the same problems.

The solution

HOMER built a unified data lake on the Databricks Data + AI Platform in just six months, using Delta Lake to centralize access to hundreds of millions of monthly event transactions and power ETL pipelines, then used MLflow to accelerate delivery of churn and conversion models to production.

Reported business value

HOMER can now see student lifetime value and key conversion and churn moments, informing targeting strategies; a 1%-5% increase in conversions is estimated to have an impact on revenue of upward of $20 million and growing, with $2-4 million in revenue per point of increased engagement on paid offerings.