Quick Queries For Third-Party Analytics: Our Partnership With StarTree

In recent years, we at BCV have heard plenty of creative metaphors on how data is the new oil, gold, etc. And, while we take no issue with the metaphors, the description is still quite a bit different for many business users. For them, more than anything, data is the new hassle.

Even though data is increasingly important for informing business decisions and staying competitive, the right data still never seems to be available when you need it.

While myriad factors contribute to the untimeliness of business’ data, what matters more are the two reasons why businesses care in the first place.

First, the cycle time between the production of data and its use is shortening — and fast. Business users simply can’t afford to wait for drawn-out batch processing jobs to complete before they make data-driven decisions. For example, if you’re dynamically pricing seats on tomorrow’s flight from MIA to SFO in response to customer interest, you want to be able to query the last hour of webpage traffic now, not at 2 AM when an ETL cron-job completes.

Second, businesses now need to surface analytics to external stakeholders — be they customers, partners, or vendors. These third parties increasingly depend on access to real-time data to operate their businesses effectively. For example, if you’re helping franchises of your convenience store chain optimize tomorrow’s inventory, you can’t wait for all of their stock and transaction data to get beamed back to HQ and batch-processed at the end of the week.

The takeaway is simple: the ability to query and analyze real-time data is no longer a luxury, but instead a necessity for business’ and external affiliates’ topline. Often, such data needs to be joined with or enriched by other real-time data or batch data to be useful.

Unfortunately, the tools in market today don’t offer what businesses need: a solution that can easily ingest real-time and batch data from disparate sources and operate at scale to surface that data to internal and external end users. After hearing this over and over from our friends in enterprise data teams, we built our conviction that there was a significant gap in the market.

That’s why we were so excited to meet Kishore Gopalakrishna, the CEO and founder of StarTree, in summer of 2020. And today, we couldn’t be happier to announce that we’ve led StarTree’s $24 million Series A round alongside our friends at GGV and CRV.

StarTree addresses the multi-billion dollar global market of real-time analytical use cases, solving precisely the issues defined earlier. Using StarTree, businesses can start querying (and enriching) their data as fast as it arrives, whether from batch sources (e.g. data lakes, warehouses) or real-time sources (e.g. streams like Kafka, Flink, Spark).

StarTree’s capabilities deliver not only internal business value, but help inform business decisions for any external stakeholders that depend on that real-time data.

The technical backbone of StarTree is just as compelling as its business implications. Kishore and his team are building StarTree’s core offering atop the Apache Pinot open-source project, which is a popular real-time online analytical processing (OLAP) datastore built and battle-tested at LinkedIn. Over the course of a decade at LinkedIn, Kishore led the development of Pinot, where it was used in tandem with Kafka and used to create features that demanded real-time data to function. For example, if you’ve ever looked at LinkedIn’s “Who viewed my profile?” page, you’ve been served data via Pinot.

StarTree is the only solution in the market that’s been hardened against the needs of a company with over 750M data-generating users. For StarTree’s customers, it means that they’re getting the best real-time OLAP datastore in the market. Customers told us that Pinot differentiated from other offerings with its:

• Superior latency benchmarks
• Simpler data ingestion, removing the need for costly pre-computation before data can be queried — especially important for time-sensitive use cases
• Support for staggered data arrival — a godsend for data teams who are otherwise required to run manual batch update jobs for delayed data

At BCV, we’ve spent hundreds of hours mapping the data infrastructure landscape — from data streaming technologies and query engines to data visualization layers.

That’s why, from that very first meeting, we knew that Kishore was a special founder and that we had to do whatever we could to partner with him.

We couldn’t be more excited for the years ahead — for the chance to work with Kishore, his team, and our co-investors to build StarTree into a category-defining company.