Hydration Room runs IV hydration and wellness clinics around the country. I started working with them as a consultant through PRIORA in 2025 and joined in October 2026 as Director of Data & Technology, so this is as much a description of my day job as it is a case study.
The problem
Like most multi-location businesses that grew quickly, Hydration Room had its data wherever each system happened to put it. Bookings lived in one place and the CRM in another, and call tracking, payroll and patient surveys each had their own logins and their own version of the truth. Answering a simple question about a clinic meant exporting from three tools and reconciling them by hand, and two people asking the same question would often walk away with two different answers.
What I built
A single warehouse in BigQuery, modeled in dbt (about 200 models at last count), that pulls from all of those systems and holds one definition for each metric. On top of it sit dashboards built for the people who actually run the clinics: district managers, nurse managers, the front desk, the membership team and marketing. Each of them sees the handful of numbers their job turns on rather than one sprawling report meant for everybody. Clean data also flows back into the CRM, so the team can work from the warehouse’s numbers without ever opening the warehouse.
AI agents now do much of the development. They work from plans I’ve approved and open pull requests, and I review every one before it ships. It’s the arrangement I describe on the Work page, a very fast B squad with a senior checking the work.
What changed
Tying call tracking back to bookings attributed six figures of revenue to paid search. Reconciling the tip pool turned up a five-figure correction in a seven-figure pool. Neither took a clever model, only getting the systems into one place and asking the obvious question.
Built with BigQuery, dbt, Metabase, Reverse ETL to the CRM, AI agents with pull request review