Build a composable CDP with predictive ML modeling
Overview
A true single customer view is the key to powering personalization at scale and driving higher return on ad spend.
With our composable CDP Data Product Accelerator, you can build a single customer view within Databricks and leverage Snowplow’s rich behavioral data to determine a customer’s likelihood of conversion with MLFlow.
Use your powerful audience insights to optimize your marketing strategies and deliver new business value, much faster.
What to expect
This accelerator guides you through how to build an advanced Composable CDP using the following tools:
- Snowplow – to create user behavioral data from your product
- Databricks DeltaLake – to store the data
- Databricks and MLFlow – to train and execute sophisticated ML predictions to determine a likelihood of conversion
- Hightouch – to synchronize the audience segment with marketing tools (like Braze, Salesforce and Facebook Ads) and accelerate conversion into qualified leads
The accelerator should take about 6 hours to complete, and includes three key steps.
Step 1. Build a predictive model
Using Databricks and MLflow to build a machine learning model that can accurately predict conversion events using features collected from Snowplow’s out-of-the-box modeled data.
Step 2. Data activation
With Hightouch connected to your rich user data in Databricks, you can enable your marketing teams to effortlessly build new audiences and sync to their needed destinations.
Step 3. Putting your model into production
Productionalize your ML model and visualize ad campaign performance synced from your Hightouch audiences.
Once finished, you will be able to use predictive models to achieve a competitive advantage from customer behavior data on your website, driving higher return on ad spend.