From ad hoc testing to a 2-day QA SLA

I led data quality for the Fortune Brands Innovations (FBIN) account at Data Radian Technologies, managing a team of 6–7 data quality analysts over about two years.

There was no standard QA process. Every analyst tested in their own way, often with complicated, bespoke logic, so the knowledge of how to do the work lived in individual people's heads instead of in a repeatable process. That made the team hard to manage day to day, and if someone went on PTO, their tickets just stalled — no one else could pick up their work.

I designed and standardized the QA operating model from scratch — workflows, documentation standards, testing templates, and approval processes — so every analyst tested the same way instead of inventing their own approach.

I deliberately kept the testing logic simple rather than letting it stay complicated and analyst-specific, so any analyst could pick up any other analyst's ticket.

I led the migration of thousands of regression test cases from dbt-based testing to Snowflake Data Metric Functions, moving the platform onto modern cloud-native architecture.

I built a centralized Tableau data quality product giving the team and stakeholders real-time visibility into production data quality metrics and regression outcomes.

I standardized on one simple, uniform testing approach rather than letting each analyst keep refining their own more complicated logic. That gave up some of the fine-tuned, case-by-case precision individual analysts had built up — in exchange for a team where any analyst could cover any other analyst's work, which mattered more once we were 6–7 people deep into a two-year engagement.

The standardized process established a 2-day SLA for closing QA tickets, each backed by proper evidence documentation — replacing a process where turnaround depended entirely on which analyst picked it up.

It also made the team resourcing-flexible: analysts could cover for each other, including PTO, which wasn't possible when everyone tested their own way.