TOPIC: Case Studies
Enablement of Data Sharing
VP - Data Management
A key capability targeted for 2022 is the enablement of data sharing through Snowflake. Our Enterprise Lake House (ELH) will leverage Cloud Omni channels to ingest data from Data Marketplace and enrich it by joining internal and external mortgage and financial services datasets.
CSS’s core business is to receive mortgage data from customers and help them in converting mortgage loans into securities by following a common framework. Securitization is a complex business process that involves exchange of data between a number of players. Traditional methods of data sharing create multiple challenges via delays, high data latency, expensive process – Extract, Transform, Load (ETL) development and maintenance cost, risk of attacks, limited network bandwidth, and more. The challenges can keep growing as data size grows, concern of sensitive information increases, or schema and format changes are needed. The need for technical assistance keeps growing to meet these challenges.
With modern data sharing using a cloud data platform, customers can be given live access to data in a governed and secure manner without moving data. This data can be query ready to avoid any ETL for loading data into customer’s data platform. Within CSS, ELH is powered by Snowflake, a leader in providing modern data sharing capability. As an example, currently CSS creates multiple outbound datasets for the Government-Sponsored Enterprises (GSEs). These are sent as files, and the GSEs load the data into their data platforms after parsing and validation. This is a Service Level driven process which has multiple data hops, data reconciliation process for CSS as well as the GSEs. With Snowflake’s unique data sharing feature, GSEs can get personalized access to outbound files without data moving from CSS to the GSEs. This eliminates any delay, minimizes surface area for risk exposure, and eliminates the need to develop and maintain ETLs.
Data Sharing offers a lot of Analytics benefit as well. The data team at CSS has built rich analytics capability around issuance, bond administration, servicer performance and delinquency related data. Data Sharing allows our customers to take advantage of these value-added services in real time.
ELH is transforming CSS to become a data-driven company. By joining internal and external data – enriched datasets, data products or Machine Learning (ML) driven models can be created.
CSS can realize the business benefits of data sharing in three different ways:
- 1:1 Inbound/Outbound Data Sharing – This is an example of external data sharing with a specific organization. CSS is in discussion with Freddie Mac to start a Proof of Concept (POC) for demonstrating the value of data sharing. In this example, Freddie Mac is interested in consuming analytics-ready data stored in ELH Data Mart. This way, Freddie Mac can leverage analytics work done by CSS and while further enriching the value by adding their own datasets to CSS shared data.
- Data Exchange between selected companies – This involves creating a consortium of organizations that can share common data without moving. Access can be governed so each organization can have personalized access to data relevant to them. Creating a Data exchange between Freddie Mac, Fannie Mae, and CSS can be very powerful. As an example, we can create a master information of suppliers and servicers using a Master Data Management (MDM) process that can be used by all the three companies.
- Data Marketplace – This is where Cloud omni channels can be used to acquire public data from vendors, suppliers, and partners. Snowflake and AWS Marketplace have data from 300+ providers in the financial, retail, travel, environment, and other sectors. If any of the current CSS data providers are in Marketplace, data acquisition from these providers can be a lot simpler by using a Marketplace. Further, CSS can enhance its data analytics capability by augmenting current data with some useful datasets available in Marketplace. Snowflake marketplace already has data from companies like Bloomberg, LoanBeam, 1010data, First American, CoreLogic, etc., which have a lot of relevant data for CSS business.