Prana Thrive is an Australian online gut-health clinic using microbiome testing to build personalized treatment protocols. The clinical model worked. What didn't exist was an operating layer: content ran through the founder by hand, nearly all new-client growth leaned on paid search, and the numbers that should steer the business lived scattered across half a dozen tools and a spreadsheet.
Where it started
The funnel itself worked: paid search fed landing pages, email nurture warmed leads, and consultations got booked. What sat underneath it was manual effort. The weekly patient newsletter was researched and written by hand and slipped whenever the founder's week filled up, every microbiome test result was interpreted manually, and revenue, bookings, and ad performance lived in separate tools with no reconciled weekly picture.
Sequencing the work around clinical safety
The design rules went on record before anything was built: every system drafts, a human approves, and nothing clinical ships unreviewed. We sequenced the work in waves, content systems first, clinical workflow second, then the operating system tying the business together, so the first thing the clinic saw was a working system in the safest possible lane.
Settling the attribution question with evidence, not a new tool
The Company OS backend's centerpiece is a weekly leadership brief that pulls calls, program revenue, store revenue, CRM attribution, and ad spend from source APIs, reconciled against the founder's own tracking spreadsheet to the dollar before it was trusted. The same build measured ad-attribution capture at 28%, handed a concrete fix plan to the client's ads specialist, and re-tested until capture hit 92%, closing the debate about buying a third-party attribution tool.
The previous fixes failed for the same reason: they automated the typing but left the founder driving. We automated the driving instead, and kept every clinical judgment with a person.
