December 2, 2020 | Investment Themes

A unified, open-core platform for data integration: Our investment in Jitsu

Avatar

Written by

Tony Liu

Data integration is not a new market. It has been and continues to be a foundational piece of the data stack. That is, the ability to ingest all relevant data at appropriate frequencies from multiple data sources. These include data from customer facing applications, SaaS applications, internal applications, and even hardware devices such as POS machines. The collected data then becomes the basis of business analyses, machine learning and data products. 

While newer entrants like Fivetran get a lot of buzz, incumbents like Informatica and Talend still loom large. There is also significant fragmentation in the market. In fact, if you ask engineers they might suggest Segment as another data integration solution even though it’s viewed by many as a customer data platform (CDP) and, for data sources not supported by these solutions, teams build data collection tools in-house that are expensive to build and difficult to maintain. 

We believe that the data integration market is starting to consolidate. 

In particular, we’ve seen that the rise of the data lake or warehouse as the single source of truth has fundamentally changed how data moves. This means that the next generation of data integration companies must fit into this new data infrastructure model. In recent times, Fivetran has undergone rapid growth piggybacking off of Snowflake’s success and Segment was recently acquired by Twilio for $3.2 billion. 

Together these two companies represent the two main models for data integration – batch and event-based data collection. Fivetran captures data from SaaS applications in batch fashion. It has a curated set of high quality connectors, which are widely adopted but often prohibitively expensive. On the other hand, Segment is widely used to push event data from custom applications into a data storage solution. While the Segment platform offers customer data collection, it is ultimately a CDP rather than a pure data integration company. As a result, there are inherent limitations with its data model and its pricing does not scale for the data integration use case since it is priced as a CDP.

The standardization of the modern data stack and the fragmentation of the data integration market, including the recent acquisition of Segment, has created a significant opportunity. We believe that Jitsu is well-positioned to be the definitive modern data integration platform. 

Jitsu: the modern data integration platform

Jitsu is an open-core data integration and event collection platform built for data teams. Peter Wysinski and Vladimir Klimontovich founded the company with the goal of making it really easy to send data to the data warehouse or data lake without compromising data privacy. The platform simplifies data integration in a single highly performant platform that combines event (e.g. Segment) and batch-based (e.g. Fivetran) data collection. Developers no longer need to use multiple tools and can also build their own custom connectors on top of the platform. Today, Jitsu has open sourced its event collection project, EventNative, which has gained significant traction within a few months of launch. The team plans to release its first set of pre-built connectors early next year. 

Beyond data integration capabilities, Jitsu is laser focused on building the platform with data privacy and interoperability in mind. Teams can: 

  • Migrate between self-hosted and managed environments with one click
  • Replace their Segment and Google Analytics backend without editing code or configurations
  • Send data to multiple destinations and multiplex to prevent data warehouse lock-in. For example, they can send data with low latency to a real-time data warehouse and send batch data to an analytics data warehouse

We are really excited to join Peter and Vladimir on their journey. They are repeat founders and were CTOs at their previous companies, where they solved the problem of ingesting large volumes of event data (upwards of one million events per second!) with low latency. They are pure technologists and have designed Jitsu with a developer first mind-set and a focus on data privacy. Congratulations to Peter, Vladimir and the entire Jitsu team!