They were taping printouts to the wall.
Sales teams at a life sciences platform had every record they needed and still could not answer a basic question: who actually influences this decision?
Every project turns on one question: is the obvious explanation the right one?
Users need better search and filtering. The data is all there, so the problem must be findability.
The data was there but the relationships were not. A flat export cannot show that a small clinic drives referrals into a large hospital. Users were rebuilding that structure by hand on a conference room wall because the software refused to hold it.
I asked to see the printouts. Three teams had each built a different annotated version of the same export, and none of them trusted the others. That was the real finding: they did not need one view, they needed three views of one truth.
One map, every team’s viewthe same records, held as relationships
Model relationships, not records
Organisations render as a network sized by referral volume, not as rows in a table.The question was always relational. A list can only answer it if the user does the joining in their head.
Give each team its own saved view
Sales, marketing and account teams each save their own lens on the same underlying data.Forcing one shared view is what created the competing printouts. Three views with one source removes the version conflict without removing the difference.
Keep the export
Excel export stayed, prominently.Removing the escape hatch would have killed adoption. People need to take the answer into a meeting where the tool is not open.
Who owns the accountorganisations and ownership lines
- Show All
- Dialysis & Infusion Centers3
- Organization, CMS Teaching Hospital3
- Organization, Dept at Hospital2
- X-Ray2
- Organization, Clinic1
- Organization, Group at Hospital1
- Organization, Health System1
3 versionssame export, three sets of pen marks
Three teams had each marked up their own copy of the same export, and none trusted the others. One map underneath replaced all three, with each team's view saved on top of it.
The export never changed size. What changed is that referral volume decides how big a node is, not which row it sits on.
Every row weighs the samealphabetical, one line each
Who to call next referral paths lit up
Same nodes, same positions, every time. Sales, marketing and account teams each get their own saved lens on one underlying map.
The argument redrawn: a scan checks every row of the export for who influences the hospital and comes back empty, because a referral is not a field. Ink the links between the rows, lift the rows off the table, and the answer is a small clinic.
Who influences the hospital?918 rows to check
The answer goes to the meetingthe export stayed, prominently
When users are rebuilding your data by hand outside your product, that manual artefact is the spec. Go find the spreadsheet, the whiteboard or the printout before you run another round of interviews.
Have a problem like this one?
Send a link and a sentence. If I can help, I will record a teardown. If I cannot, I will say so.
Start a conversation