From Fragmented Data to a Single Source of Truth in Just a Few Months

An interview with Mike Zhao, Head of Decision Science, Product & Ops at Opendoor

‍

‍

Opendoor Overview

Opendoor (NASDAQ OPEN)  is a leading residential real estate digital platform that provides people across the US with a simple way to buy and sell a home. As their service grew in popularity, so did their need for insights from their data. They wanted to extract those insights to inform business decisions, but their data sources were siloed and hard to access. To unlock their data, Opendoor focused on finding a solution to improve the data access and the data consumer experience across their organization.

Mike Zhao joined Opendoor as Head of Decision Science, Product & Operations to lead their data science and analytics efforts in 2021. Based on his previous experiences at Lyft, Mike knew how much potential value Opendoor could extract from its data if it had the right data discovery platform in place. He and his team began working on a foundational framework to decide on the necessary tools and actions to support an efficient and effective data discovery journey for Opendoor.  

‍

Building a Centralized Framework

The organization defined several goals to inform its data discovery journey strategy. They wanted to:

• Organize the fragmented data sources into a single source of data truth
• Improve data access and experience for all users across their organization
• Reduce the volume of resources needed to provide that access and experience
• Identify valuable insights from their data

To reach those goals, the team needed to solve organizational pain points that were blocking data access and participation.

‍

The organization had not yet taken a strong stance on which platform to use for data documentation. Data sources across the organization were fragmented and unscalable. Many team members weren’t aware of available data tools, and competing priorities made it hard to participate, reducing the tool's effectiveness.

Data consumers found it difficult, manual, and time-consuming to find the data that they needed, when they needed it. Additionally, data team members spent too much time away from their core responsibilities connecting consumers to data. Extracting the valuable insights that lived in their data was often chaotic and resource intensive.

“Sometimes it took a day to find the data producers or owners across the company.  We had to figure out who to ask, wait for a response, and arrange a meeting to get the information we needed“ Mike said.

Mike knew that his team couldn’t solve these challenges with strategy and processes alone. They needed to motivate data owners and consumers across the organization to participate in a new approach to how Opendoor organizes, prepares, and shares its data.

‍

Setting up Data Discovery for Success

The data team widely communicated their long-term vision of success, and Opendoor created a space to experiment and focus on setting up a scalable foundation to grow their data discovery efforts.

“Data discovery is tremendously valuable, but it takes a couple of months to drive adoption and change behaviors,” Mike explained. “Having the space to make the long‑term investment for the health and the benefit of the team enabled us to prioritize the data discovery journey.”

The data team broke apart the first phase of data discovery optimization into four parts focused on driving awareness of the data discovery platform and its value to everyone.

‍

Step 1: Make Participation Easy

Data organization is a community effort. The team chose Select Star as the data discovery platform to act as Opendoor’s single source of truth for data. The easier the tool is to use, the easier it is to drive adoption.

“The lightweight nature of editing and the ability to automatically ingest documentation from third-party data sources directly into the Select Star platform is extremely valuable,” said Mike.

‍

Mike also aligned the performance of the data discovery program with the data team’s OKRs, further prioritizing participation.

‍

Step 2: Set Priorities and Scope

Mike and his team set out to provide immediate value that established trust and awareness with new users. They added documentation for Opendoor’s top 100 tables based on each table’s popularity score in Select Star to jumpstart the organization’s transition to the platform. This practice established champions across the team who helped optimize the platform and prepare to share it with the greater organization.

‍

Step 3: Spread Cross-functional Awareness

When the initial documentation was filled in, they were now ready to market the value of using Select Star to data consumers horizontally across the organization. They created guides to help ease first-time users into their new workflows.  They took time to share strategic direction and answer questions to help heavy data users and teammates ramp up successfully.

“In just four months, we were able to double the Select Star monthly active users,” Mike reported.

‍

Step 4: Continue to Optimize Over Time

As they gained traction, they concentrated on motivating Opendoor’s data users to use Select Star, which was now established as the main source of data truth at Opendoor. The more active users they had, the more documentation was added organically. As users began to benefit from accessing Select Star as a reliable reference for up-to-date and accurate data context, it became easier to gain their trust for self-servicing their data questions and contributing back to the platform.

‍

Automation Makes a Difference for Data Producers and Data Consumers

In the past, building a foundation for data discovery of this caliber would have taken years to establish. Opendoor used Select Star to establish a solid foundation that supported its data discovery journey quickly without the need to build a complex structure. The increase in cross-functional engagement created a positive cycle of results.

“People began making incremental edits, maintaining the platform, and adding more collective knowledge and information. It turned into a wonderful flywheel effect,” Mike explained. In less than a year, Mike and his team established a trusted source that empowered data consumers to self-serve.

The initial phase of the journey has already provided positive momentum toward reaching Opendoor’s initial goals, including reducing the cost of being a data consumer at Opendoor. Some of the results the Opendoor team experienced were…

• Data consumers have a fast and easy way to find the information they need from a reliable source of data truth.
• Team members can access higher-quality insights in less time. Key data sources and queries are now searchable and easy to find.
• Any user has a reasonable path to identifying which tables they should use as a starting point.
• The amount of time data analysts spend helping facilitate data questions has been reduced from hours to minutes, giving them back valuable time to work on other responsibilities.

“We’re able to move faster,” Mike said. “Finding the data information we need takes less than 5 minutes now.”

‍

Data discovery is a Continuous Journey

The results from the first phase of Opendoor’s journey are encouraging, but they’re looking forward to further optimizing their approach. Building on top of what they’ve already achieved to grow and optimize their data strategy with the organization's needs. Mike plans to focus the team's future energy on optimizing the data model to be more useful, flexible, and reliable.

“Select Star will be pivotal to help us reach our next set of goals,” Michael stated. “When we’re ready to solve lineage, ownership, SLAs, or higher level documentation, it should be easy to drive the adoption because the features already exist.”

Mike found that engagement, awareness, trustability, consistency, and a long-term view of success are essential to the data discovery journey. Using a data discovery platform supports all of those initiatives.

“Data discovery platforms are a key part of the modern data stack for a good reason,” Mike said. “Any fast-moving data-centric company would benefit from Select Star.”

‍

At Select Star, we’re focused on solving data discovery challenges. To learn more about how Select Star can help your organization set up an automated data catalog in just 15 minutes, schedule a demo today.

‍

‍

‍