Insights

Data Engineering

One of the most powerful and underestimated aspects of Snowflake is the fact that everything is represented in ANSI-SQL and is relational. We can click into our first-party database, view table structures, and even take a sampling of our first-party data. For data engineers, implementing streaming ingest, continuous data pipelines, and data governance using native Snowflake features like Snowpipe, Streams, Tasks, and secure views is simple, lightweight, and requires zero ongoing maintenance.

Data Marketplace

With the tools to ingest, transform, and conform first-party data, we can take a look at integrating third-party data. Accessing data sets via the Snowflake Media Data Cloud can be configured with a few clicks and does not require any infrastructure or code. 

Media Data Cloud partners publish a considerable amount of live, ready-to-query data on Snowflake’s Data Marketplace. This data is delivered instantly to customers through Snowflake secure Data Sharing, meaning that it is all live without the need for physical data movement or SFTPs. This is what creates Snowflake’s unique Media Data Cloud workflow.

Telling the Story Through Tableau

But data isn’t valuable unless it tells a story. Leveraging Tableau makes the enriched data set more easily explorable and understandable. Using pre-built dashboards, we can start to understand, in near-real time, how customer lifetime value is performing across all subscribers and then break it out by various campaigns and cohorts to see which ones have outperformed or underperformed against the average. We can also start to overlay third-party demographic data to gain deeper insight with virtually no effort.

ML Model

Snowflake can provide holistic subscriber insights with unlimited data and instant access to the Media Data Cloud. But Snowflake really starts to drive unparalleled ROI by bringing machine learning models directly to the data as opposed to the old approach of shipping data out to an ML environment and waiting for results.

As an example, let’s take a look at a marketing leader responsible for maximizing customer lifetime value and delivering marketing ROI. Our data scientists and data engineers have leveraged historical content, release, campaign, and acquisition data to train a model that intelligently predicts the expected CLV for any new release broken out by segment. This model has been packaged up and deployed for production use in Snowflake as a User-Defined Function. This function can be used by anyone in Snowflake using basic, standard SQL. In the case of this marketing leader, he or she can leverage the model to understand the expected CLV for all targeted audience segments and use that to optimize spend. In executing related queries, Snowflake runs Python, Scala, or Java code directly in the platform with cloud scale, quickly generating the predicted CLV. The marketing team can immediately make informed decisions about where their spend might drive better results.

Subscriber analytics is just one of many use cases where the Snowflake Media Data Cloud enables organizations to have a more holistic and accurate view of their customers, optimize spend and increase revenue. Snowflake’s unique ability to analyze all your data, seamlessly integrate third party identity and audience data, and leverage machine learning models directly in the platform are empowering media and advertising organizations to achieve results that were not possible before.