“Coalesce Catalog saves the team significant time when iterating on data sets. I’ve received messages from many people across the company saying how much it helps them do their job.”
Wrong prescription for data governance
Challenges
Original data catalog was built for technical users and difficult to scale usage internally
Integrations with critical systems like Salesforce and data visualization tools did not work as expected
Absence of end-to-end data lineage made it impossible to understand downstream impacts
Diana Carrondo is the Data Governance Lead for Doctolib, and her team helps enable all internal data consumers to confidently leverage the company’s data while maintaining strict data privacy protocols. “Our data warehouse is chiefly Amazon Redshift, but we will be using Google Cloud Platform by the end of the year,” she says. “The main data sources we have are from Doctolib. They are protected by design at the source, and are never available in our data platforms thanks to a rigorous sensitivity classification process.” Doctolib also uses Airflow for orchestration, Metabase for data exploration, and Tableau for more official dashboards.
The Data Governance team is also the functional owner of the company’s data catalog. Doctolib originally had a different data catalog in place, but the team felt it was failing to meet their needs. “First of all, it didn’t address all the use cases we had for our different personas,” says Diana. “And it was more oriented to highly technical users—not great if you’re looking to scale your data catalog. We wanted something as user-friendly as possible that everyone could use.”
According to Diana, “We were searching for a solution that would allow for data self-service so that Doctolibers would stop coming to the data team for simple questions, and so we needed a data catalog to serve as a single source of truth for data discovery.” And improving data discovery and self-service wasn’t just about convenience: “It was a necessary step to make sure the whole company could move faster with AI, given our strong focus on innovation.”
She adds that despite promised integrations with other platforms, many of those they needed most, such as the integration with Salesforce, did not function properly. But the biggest issue was that the Data Governance team had no insight into data lineage, and was unable to determine what the possible downstream impacts might be of something they were integrating. “The lack of data lineage was one of the biggest reasons why we decided we needed to ‘escape’ from this platform and start looking for a better solution,” she says.
Getting a second opinion
Solution
Assessed several data catalogs in the market with help of beta testers
Ran successful proof of concept with Coalesce Catalog
Evangelized new catalog across entire organization to promote widespread adoption
As the Data Governance team assessed and tested several alternatives to their existing catalog solution, they included the larger organization to be sure they chose something that would be widely adopted. “I gathered over thirty beta testers from different teams to represent multiple personas, each with distinct use cases,” Diana explains. “I ensured the use cases were relevant and asked testers to evaluate each tool against them.”
After running a proof of concept with Coalesce Catalog and analyzing their beta testers’ feedback, the Data Governance team was convinced that it met all their requirements: “We made our case to the executives by describing our previous tool as an expensive burden, and that Catalog was a better value for investment as it was superior in every aspect: functionality, technology, security, and price. Catalog unlocked way more value.”
The team was also impressed by Catalog’s AI features, which mark an important first step toward developing data self-service for all Doctolibers—especially feature teams. These AI features are so user-friendly that they are regularly used by other teams in an organic way, with no additional effort required from the Data Governance team. Catalog’s AI assistant also supports data users with smooth integration into their favorite analytics tools.
When the new catalog was rolled out at the beginning of 2025, the team worked hard to evangelize it to all Doctolibers across the organization and encourage them to start adopting it. They wanted to reach the general users who may have considered a data catalog something too technical to be a part of their daily toolset. “I even worked with the design team to create a video campaign and design flyers inviting the whole company to a Catalog kickoff,” says Diana.
A healthier data outlook
Results
Explosive growth in adoption, going from 100 to 600 MAU on rollout
Broader engagement across 60+ product and tech teams
Stronger governance with restored KPIs, secure code-based data classification, and reinforced risk management
Major efficiency gains for governance team, freeing time from onboarding and bug triage
The Data Governance team set out to promote data self-service across the company, and the sharp increase in monthly active users (MAU) of the new catalog is their clearest measure of success. “With our previous catalog, we struggled to get more than 100 active users,” says Diana. “With Coalesce Catalog, adoption has taken off—at rollout we hit 600 active users. It delivers more value without more investment, and people are finally contributing—commenting, documenting, and keeping the catalog alive, instead of our team carrying it alone.”
And as she hoped, this adoption goes well beyond the data teams: “Product managers, engineering managers, and even security teams now use Catalog. With more than 900 people across 60 feature teams using Catalog at some point, it has become fully embedded in Doctolib’s broader product and tech culture.”
“Now our catalog provides the information we need to track ownership, classification, and documentation, allowing us to rebuild our KPIs with even stronger visibility,” Diana says. “This not only strengthens governance but also shows our team is delivering on its mission to help leverage value from data with privacy and confidence.”
Diana explains that in the previous catalog, data owners had to manually tag each column to indicate whether it contained personally identifiable information (PII) or protected health information (PHI). “That approach made the catalog less of a ‘store window’ for data and more of a place where rules were enforced,” she notes. “With our new catalog, the classification rules live in the code that manages the data, not in the catalog interface itself. This shift means the catalog can focus on presenting information in a structured way, while the underlying code—complete with version control and change logs—ensures that updates are both scalable and auditable.”
As for their future plans, the Data Governance team sees the next challenge to make sure Catalog thrives in a world where AI agents are becoming everyday tools. “At Doctolib, agentic tools such as Dust, a platform to quickly create AI agents, are used by the large majority of Doctolibers and becoming part of daily life, so one of my priorities is to build bridges that allow Catalog’s data to flow right into that ecosystem,” says Diana. She notes that Coalesce’s future plans to build an MCP server will unlock the ability for users to interact with Catalog through any of the agents they rely on. “That way, whether people use Catalog directly or through the agents they rely on, its value will always be there.”
Want to read more of Doctolib’s data catalog journey? Check out Diana Carrondo’s 6-part Medium series, Changing Our Data Catalog: Why Healthtech Company Doctolib Decided It Was Time for an Overhaul.
“With Coalesce Catalog, adoption has taken off—at rollout we hit 600 active users. It delivers more value without more investment, and people are finally contributing—commenting, documenting, and keeping the catalog alive, instead of our team carrying it alone.”