Here at Facet, we’ve been heads down building our core product that we will be building upon for the future. Facet is inspired by what we learned while running the Metamarkets business, and more importantly from the thousands of users who love using the Metamarkets product and the power it gives them over their data. Metamarkets was an amazing product that customers loved but it had a limitation.
What we set out to do, is simple on paper, but harder in application. We aspire to keep everything people love about Metamarkets (an easy to use, intuitive data exploration tool) and remove everything which wasn’t liked (unpredictable and large costs, no control over data modeling, duplicative processing). This always was a difficult task, because the application was built on Druid. At the time, nothing else existed where data that size could live and return fast reliable data response times. In addition, the only way to transform and load the into Druid was through complex ETL processes. Users were constantly saying “I love the Metamarkets dashboard, can’t you just connect to my own internal database?” or “I already perform this processing internally, can’t we just use that?”
With Facet, this is no longer the case, and the functionality you loved with Metamarkets can be obtained without the items you hated. Facet brings modern exploratory analytics and to your CDW. With minimal setup time and no SQL knowledge required in order to explore your data.
This has become possible with the rise of the modern Cloud Data Warehouse (CDW), most notably, Snowflake and BigQuery. These data warehouses, have really solved two major problems, queryable data storage, and data transformation. Queryable data storage means that you can easily dump the data into a data warehouse and begin to query it as you need it. Data transformation has evolved from a complex ETL process into a more simplified ELT process using the data warehouse, transformed with dbt jobs, into a clean table ready for queries in the data warehouse. Every day, you see new posts about how the use of Snowflake has optimized work processes, reduced costs and reduced maintenance for engineers and data teams everywhere.
Because Snowflake is so versatile and fast, it is actually our recommended backend to be used with the Facet application. As avid Druid users, we had assumed that Snowflake or BigQuery would not be fast enough to provide an optimal user experience for an exploratory analytics tool, but this is simply not the case. Leveraging and optimizing snowflake, we have seen sub-second response times in datasets with billions of rows of data inside of them. Facet on top of Snowflake was a breakthrough for Facet.
Anybody who is currently on Snowflake or BigQuery today, can become a user of Facet in minutes. Anybody who is storing raw data in S3 or GCS buckets in the hopes of making it actionable, can simply begin to load that data into Snowflake or BigQuery in no time at all, and begin to benefit from a no code, data exploration tool. While giving a demo of the Facet product, a data scientist at an AdTech company told me “I’ve always hated Metamarkets because it’s restrictive and too expensive, but having this front end on top of my data would be amazing.”