(Gorodenkoff/Shutterstock)
Organizations are drowning in data, yet they’re struggling to convert that data into actionable intelligence. Building that last mile of connectivity between raw data and useful information is the goal of a young firm called the Modern Data Company.
Before co-founding Modern Data Company in 2018, Srujan Akula and Animesh Kumar helped build data management systems for a number of companies, including gaming companies and adtech and martech firms, and they saw similarities among the challenges.
“We realized that everybody is trying to kind of get a handle on their data management in a similar fashion,” Akula said. “Everybody seems to be struggling on the other side with getting value out of the data.”
Data engineers, in particular, were hitting their heads trying to work with all of the different systems they needed to hammer the data into shape. They had to navigate a morass of data integration, transformation, lineage, governance, and other tools to get the data into a shape where it could be readily consumed by their line-of-business colleagues.
That’s when the collective lightbulb went off in their heads.
The co-founders of Modern Data Company
“Our vision is, like any other product management, if data gets treated like product, if you really understand what is the outcome you’re trying to drive, what exactly needs to be the shape of the data, the governance, and if that can get captured and we can provide simple ways for data engineers to facilitate that, that could be a win,” Akula tells BigDATAwire in a recent interview.
Their idea was to build a data operating system–or a DataOS, if you will–for data that could serve as the foundation for a suite of components that provide the specific capabilities needed to serve data as a product, whatever the use case. For instance, if a marketing team is building a customer 360 system, then the Modern Data Company could support that by presenting the data in a way that makes sense for that use case.
“So it starts with the discussion with the business team to capture the intent, and then the DataOS gives you all of the capabilities to ensure that the shape of the data, the semantics of the data, the way the business needs it,” Akula says. “We have a data contract construct that where we capture quality, freshness, completeness, schema validations to governance tools. All of these are captured into that singular construct.”
Instead of worrying about the nitty gritty details of the data management infrastructure, Modern Data Co. instead will provide a full soup-to-nuts offering, a la a data fabric or data mesh. Its DataOS Data Product Hub, which it launched on October 31, provides the full array of data management capabilities, such as a data catalog, statistical profiling for data quality, audit and logging for governance and compliance, observability for schema drift, and a semantic layer for understanding. If customers already have some of these tools, they can build DataOS around it.
“More often than not, customers already have an existing infrastructure,” Akula says. “All of these companies have invested in Databricks, Snowflake, or Dremio. But they still are seeing value in our capability to close that last mile gap in data. How do I make data consumption much simpler than what it is today? That’s kind of been our story.”
The DataOS Product Data Hub (Image courtesy Modern Data Co.)
But it’s not just about exposing high quality data in an optimal format to business users (the company follows the Linux Foundation’s Open Data Product [ODP] specification as well as Apache Iceberg to avoid vendor lock-in). It’s also about creating a full-formed data product with all of the capabilities customer expect–a data contract, service level objectives (SLOs), etc.
“My vision and our vision as a company has been if you treat data like a product, it can be a true business asset where you don’t need to have technology teams involved,” he says. “Black box the underlying infrastructure and give this curated, governance-approved, single-source-of-truth data set data products that are powering all of your use cases.”
Instead of pestering data engineers for weeks to get a particular data, Modern Data Co. aims to empower the line of business to simply search for the right data product. If they’re allowed to use it, provisioning the data via an API (REST, GraphQL, etc.) is simply a few clicks away, according to Akula.
By serving as a sort of lens through which users can access data products, Modern Data. Co. functions as a form of data virtualization. Akula makes no bones about his intent to replace Denodo, one of the established leaders in the data virtualization space. While the Modern Data Co. doesn’t have nearly as big a presence as Denodo, it is currently getting the attention of some pretty sizable customers. Akula says he is targeting the enterprise segment, or companies with revenues of $10 billion or higher.
According to Akula, the company has successfully deployed its technology at a $27 billion distribution company that was struggling to build and deploy a data hub to power a customer 360 solution. After implementing the data hub in just three weeks, five other business teams sought access to the curated data, he says.
“It took us a laser focus of almost five and a half years to mature the platform to provide this kind of lifecycle management,” he says. “Because it’s not just about building data products and creating a view or data store like a view sprawl that you have today. But how do you manage the lifecycle? How do you understand how much are you spending? What is the ROI? There’s a lot that we have done around the data product lifecycle to make it enterprise ready.”
After being heads-down in engineering mode for years, the company is now ready to begin really showing the world what it’s got. That has brought some market success, as according to GetLakta, the San Mateo, California company generated $18.4 million in revenue in 2023.
But Modern Data Co. clearly has bigger goals. The company wants to keep expanding the use cases on top of the DataOS, and making data products easier to build and consume.
“It’s making the life of a data engineer a lot simpler,” he says. “Data engineers are working with clear outcomes rather than the boilerplate and the minutia, because the DataOS takes care of all of that through our declarative programming constructs. I’m excited that it’s making the data engineer a lot more efficient, but business teams are able to move 90% faster on any use case. And that is the big for me, is how nicely the business teams are able to collaborate and work off of this hub.”
Related Items:
Mastering Data Modeling: Insights from a Data Product Developer
Why a Universal Semantic Layer is the Key to Unlock Value from Your Data
Data Products: The Solution to the Data Silo Trap
