SIRUM is revolutionizing healthcare access in the United States as the largest distributor of surplus medicine. Harnessing the capabilities of technology, SIRUM facilitates the donation process for organizations, including nursing homes, pharmacies, and manufacturers, ensuring that unused medicine reaches those who need it most.
As SIRUM prepared for its next phase of growth as a startup, it needed new data capabilities to scale efficiently. The challenge was clear: reevaluating the most effective ways to derive value from their data, and most importantly doing so in a manner that guaranteed the quality of their data. This meant consolidating and centralizing their data to modernize and improve their analytics.
Being a small company, the challenge lay in implementing their data stack. It had to be user-friendly for various stakeholders in the organization, as well as easy to maintain without external technical support — a solution that enables self-service. Steep learning curves and onboarding delays for new teams or hires had to become a thing of the past.
The new stack also had to make data dependable and trustworthy. The company was identifying errors and rectifying them in a reactive manner, having to analyze their origins and check KPIs. Without a proactive approach to data quality, uncertainty was building up around KPIs, impeding confidence in decision making.
Lastly, the solution had to be scalable. Startups grow, and they do so quickly. This data stack had to function as a robust foundation that would enable SIRUM to experiment with and incorporate machine learning and AI capabilities in the future.
In essence, the task was to assist SIRUM in becoming an increasingly data-driven non-profit organization, ultimately enhancing its ability to support the patients it serves.
Airflow served as a scalable solution for overseeing data pipelines and provided DAG templates, simplifying the creation of multiple workflows based on dbt. These workflows facilitated essential functions like re-processing and rollback. dbt was chosen to design, document, and validate views and tables.
Soda was chosen for data quality monitoring and assurance due to its capability to define tests that become aggregated SQL queries, whose results are consolidated into a readable format stored in a table. Soda provides a wide array of built-in tests, extending beyond SQL file-based tests.
Mutt Data, a professional services partner, was responsible for the design and implementation. Adam Kircher, Co-founder at SIRUM, says: "Having previously collaborated with Mutt, choosing them for our modern data stack design and implementation was a no-brainer. We found a strategic partner with end-to-end support and expertise and a process-driven approach that gets the job done."