Medical Device: Commercial-Stage Manufacturer

What the Business Looked Like

A commercial-stage medical device manufacturer sold into hospitals through health authorities and distributors, a layered relationship its CRM did not model. Sales, pilots, implementations, and research all ran through one undifferentiated pipeline, and roughly $2.05M of surgical closed-won revenue carried a reorder pattern nobody was tracking. The historical record had been accumulating without maintenance for years.

What Broke and What It Cost

Health authorities were duplicated between two and five times over, a domain auto-association rule was linking contacts to the health authority instead of the hospital they work at, and five active deals including an $81K opportunity had no contact attached at all.

Nobody could tell a live deal from a dead one, so pipeline reviews turned into archaeology. Duplicated health authorities split the history of the same buyer across several records, deals with no contacts attached could not be worked by anyone other than the person who opened them, and a reorder pattern worth tracking was invisible inside closed-won noise.

What Was Built, in Dependency Order

A data assessment delivered the morning of the first working session, then a distributor-aware parent and child entity model, four purpose-built pipelines with full stage sets, a gated 12-stage sales pipeline with probability weighting where unqualified deals sit at zero percent, and a 100-point hospital tier-scoring model held back from launch until it was QA'd against known-answer records

1. 01Data assessment delivered before any build was quoted
2. 02Distributor-aware parent and child entity model
3. 03Four purpose-built pipelines and a gated 12-stage sales pipeline
4. 04Hospital tier scoring, adoption, and client-side ownership

What Changed

Leadership reads a demand matrix instead of guessing, and the forecast cannot overstate itself because the pipeline documentation states which stages count as procurement-grade until two quarters of real conversion data exist.

A triage workbook dispositioned roughly 175 deals into 11 keep, 79 review, and 71 close-lost, presented to the client as honesty rather than indecision. The entity model was rebuilt the same day distributor complexity surfaced on a call, with the reporting distinction written in that a new usage site under an existing distributor is not a new revenue deal. Six team members were onboarded one to one with live CRM entry tests against a naming convention, and a client-side business development analyst took ownership with a playbook, written guardrails naming what he can and cannot change, two dashboard specifications, and a recorded walkthrough. The next phase of automation and revenue rollups is specified and deliberately unbuilt until the engagement structure is agreed.

"The original specification called for a custom object. The client's plan tier does not include custom objects. We found that out before building it, and shipped the version their tier supports instead of an upgrade invoice."