Supernova Award Category
The Problem
Given the importance of the endowment fund and research funding, Notre Dame relies on analytics to help it gain an edge in giving. Alumni and potential donors interact with the university in many different ways: mobile, web, and physical presence. The key to understanding their desire to contribute and identifying the next big donors lies in using the data generated from these interactions.
Prior to implementing Snowflake Computing, Notre Dame relied on a legacy on-premises data warehousing solution that was not able to keep up with the growing needs of the analytics team at Notre Dame. The previous system couldn’t handle a large amount of data, and as a result, queries were taking far too long to complete and analysts working on the data set were starting to get frustrated. In addition, the legacy data warehouse could not easily ingest and query JSON data coming from mobile and web. Plus the environment was not flexible and scalable to accommodate the seasonality in donor behavior.
The Solution
Notre Dame identified multiple potential partners that claimed to have solved the business problems faced. These included Notre Dame’s legacy on-premises data warehouse provider, and cloud based data warehouse providers like Snowflake and others. Notre Dame identified three criteria to evaluate the data warehouse technologies: Features and Performance, Cost and Vendor history. The judgment of these criteria was based on each partner performing a Proof of Concept. Snowflake beat the competition hands down in features and performance, and cost categories, with a competitive score in Vendor history.
Moreover, using Snowflake, Notre Dame was able to bring the environment into production in just 3 weeks with just one developer. In fact, the end user analysts were in such disbelief that one of them commented: “I don't actually believe that this is working [because it’s so fast].” Today, analysts can query data quicker, enabling them to ask more questions.
The results
The university receives large amounts of data like social media data, which is essentially JSON data. Using key features of Snowflake, Notre Dame is able to feed this data directly into the platform easily, and in one location. In addition, analysts can can make this data available to the data scientists quickly, with ability to query the JSON data easily and quickly through their standard tools like Tableau and SQL. This data is essential to identifying who is the next million dollar donor.
Additionally, there is significant seasonality of the demand for funding and giving data. With Snowflake, analysts can enhance performance by scaling the environment up or down with just the click of a button. This allows analysts to concentrate on deriving actual insight, rather than spending time tuning their queries.
In addition, Notre Dame had to analyze 4 billion rows of data within a minute to understand classroom utilization. Today, using the smallest instance of Snowflake, analyze can query in 5 - 10 seconds.
Metrics
Using Snowflake’s data warehousing architecture, Notre Dame was able to shift over its entire existing data warehouse – no refactoring required. Snowflake provides Notre Dame with the scalability and concurrency they require. Notre Dame can now scale up its warehouses and storage with the click of a button. In terms of concurrency, Snowflake allows for 300+ users to run queries at the same time, thus proliferating the use of data to make decisions. Analysts are seeing performance gains from 30 minutes to under a minute. Now analysts can make quicker decisions to gain insight, rather than waiting for queries.
With Snowflake also came a lower total cost of ownership. With their previous data warehousing system, they required one full-time database analyst, two data warehouse developers and over 200 indexes. With Snowflake, all Notre Dame needs is a single data warehouse developer to manage the data and the source system connections, while the system manages itself.
The Technology
Snowflake Elastic Data Warehouse as a Service
Disruptive Factor
Using Snowflake’s Elastic Data Warehouse has provided Notre Dame’s business analysts with quicker answers and ability to ask many more questions. Analysts don’t have to wait a long time for the answers any more. Snowflake is equipped to handle the scalability analysts require and accommodate the fast growth in the alumni database.
Plus today, Snowflake lets Notre Dame analysts focus on adding more data to support better models for data scientists solving problems like who is the next million dollar donor. Plus analysts can ask better and different questions, and find the answers quickly, to continue supporting the college’s dedication to top notch education, research and facilities and maintain their status as a leading university.
Shining Moment
Before Snowflake, one of the analysts had a shirt hanging in her cubicle that read, “I’m only here because my query is still running.” This is a far cry from the statement of another analyst, who after implementing Snowflake said in disbelief: “I don’t actually believe that this is working [because it’s so fast].”