November 03, 2025
You canโt hire your way out of a data stewardship problem. ๐๏ธ
I keep seeing the same pattern at small biotech companies:
โ Team doubles after funding
โ Data requests pile up
โ Leadership posts a Data Scientist role
โ New hire gets buried in โquick reportsโ
โ Six months later, same bottleneck exists
Hereโs whatโs actually happening: Theyโre hiring a Ferrari to fix their roads. ๐๏ธ
๐๐ฎ๐๐ฎ ๐ฆ๐ฐ๐ถ๐ฒ๐ป๐๐ถ๐๐๐ ๐ด๐ฒ๐ป๐ฒ๐ฟ๐ฎ๐๐ฒ ๐ป๐ฒ๐ ๐ถ๐ป๐๐ถ๐ด๐ต๐๐
โ Build ML models
โ Design novel analyses
โ Answer questions nobodyโs asked yet
โ Create competitive advantage through discovery
๐๐ฎ๐๐ฎ ๐ฆ๐๐ฒ๐๐ฎ๐ฟ๐ฑ๐ ๐ฏ๐๐ถ๐น๐ฑ ๐๐๐๐๐ฒ๐บ๐ ๐ณ๐ผ๐ฟ ๐ถ๐ป๐๐ถ๐ด๐ต๐๐
โ Create self-service tools
โ Build data pipelines
โ Answer the same questions 100 times (so nobody has to again)
โ Create competitive advantage through efficiency
Both roles are critical. But timing matters.
When scientists spend 8+ hours weekly on routine reporting, thatโs not a data science problem. Thatโs a data stewardship problem.
A real example: One client was hiring their third data scientist while their team waited days for basic QC reports. We built a self-service dashboard instead. Reporting time dropped from 8 hours to 45 minutes. The data scientists? They finally got to do actual data science.
๐ง๐ต๐ฒ ๐๐ฒ๐๐ ๐ณ๐ผ๐ฟ ๐๐ต๐ฎ๐ ๐๐ผ๐ ๐ป๐ฒ๐ฒ๐ฑ:
If someone asks โCan you pull this report?โ for the tenth time this month โ You need a steward
If someone asks โCan you discover something we donโt know?โ โ You need a scientist
Most early-stage biotechs need stewardship infrastructure before advanced analytics. Build the roads before you buy the Ferrari.
The good news? Data stewardship directly enables better data science. Once scientists can self-serve routine analyses, your data science team can focus on the questions that actually require their expertise.
Whatโs been your experience? Have you seen teams struggle with this distinction?
