Driving Explosive Growth at DataRobot with a Culture of Data Democratization

Case Study

About DataRobot

DataRobot is a company that is revolutionizing data science via the world’s most advanced enterprise automated machine learning platform. With a new round of $100M in funding in October of 2018, DataRobot has grown over 60% in terms of employee count in 2018.

Daniil Bratchenko/>

What is unique about DataRobot is that it is a company that is all about data. And it is a company in the forefront of the data democratization culture by moving to enable every employee to access and analyze data. We sat down with Daniil Bratchenko, VP of Business Operations and Analytics, to discuss the impact of data democratization at DataRobot and what advantages and pitfalls companies that go down the same direction should expect.

The challenge: comparing time and resources in building vs. buying an analytics solution to use across the whole company

The company had started developing some basic analytics and reports internally in Python to send out via email. They quickly found out that they were spending a significant amount of time creating these report, so they realized that they would save a lot of time and engineering resources if they could have a data analytics tool in place.

The evaluation criteria to bring in a new data analytics tool were three-fold:

  1. Stand out in terms of user-friendliness so that everyone in the company can use. DataRobot didn’t have an internal analytics team, so it is imperative that the data analytics tool is simple to use by everyone without support from a data or engineering team.
  2. Connect to the company’s relational databases in a secure and reliable manner.
  3. Complete a set of identified use cases around creating certain visualizations with specific types of visualizations.

The team reviewed a number of tools including Chartio, Looker, Tableau, PowerBI, Qlick, and Amazon Quicksight. They shortlisted Chartio and Looker for a Proof Of Concept, and ultimately Chartio was selected based on the product’s overall ease-of-use and intuitive Visual SQL interface.

We built these dashboards in both Looker and Chartio, but in Chartio the process was way faster and the experience was far more intuitive.

The solution: connecting people and data across the organization for data-driven decision making

Today DataRobot’s data stack is comprised from the following main technologies:

  • Postgres as the main production database
  • Airflow for data processing jobs
  • Segment to pull data from different sources into Postgres
  • Chartio for data analytics and visualiations

The company pulls data from a number of cloud applications they are using including Salesforce, Marketo, Hubspot, Google Ads, Facebook Ads, LinkedIn Ads, Mandrill/Mailchimp, Zendesk, NetSuite, Greenhouse (for recruiting data), as well as product usage data.

In the true spirit of data democratization, almost every new employee gets a seat on Chartio. Some teams that are running their operations and relying their success on Chartio include:

DataRobot team Use case
Marketing team Lead prioritization, attribution & team performance
Sales development team Day-to-day operations, cadence tracking & team performance
Customer success team Account health reviews
HR team Operations tracking, dashboards for their stakeholders (hiring managers, team leads)

Daniil’s favorite Chartio feature is the Data Pipeline, which allows users to build powerful data transformations within the Chartio product. Daniil commented, “Data Pipeline is the feature that I would look for in other tools if I had to evaluate them again.”

We would need 10 more data analysts working full-time to be able to build & maintain the types of data analytics that Chartio allows us to do in every team in the company.

The outcome: record adoption rates for analytics and a culture of data democratization

Today, DataRobot is offering a Chartio seat to every employee. The company has built an onboarding course on Chartio that is available on the company’s Learning Management System. The company also has a dedicated internal Slack channel to support and enable Chartio users. But according to Daniil, most people learn how to build charts on Chartio by themselves and in the context of their day-to-day job responsibilities.

DataRobot has an 83% adoption rate over the past 3 months and over a third of their Chartio users are actively creating and editing charts and dashboards in Chartio. The company is currently using over 1,500 dashboards to display 15,500 charts using 3,400 interactive variables and 8,200 contextual elements in Chartio.

The next challenge for DataRobot is how to continue to scale this data democratization culture and continue to reap the benefits. One key element is how to explain data, data structures and data context to all new employees who work with data, especially as the company is growing at a record pace.

According to Daniil, the recipe for data democratization is “to maintain good data, have an easy-to-use tool, and encourage a data-driven culture, for example leadership could support asking questions that require data to answer.”

Democratization of access to data is super important when you see how it works, and if we didn’t have it, we would be much less effective as a company. It might be less obvious when you don’t have it.