From spreadsheets to a living graph: Meridian's first 90 Days

Meridian’s data team was maintaining 14 spreadsheets to track pipeline health, customer health, and operational metrics. The spreadsheets were not a legacy problem. They were the current solution, actively updated by three people across the ops and finance teams. Every Monday, someone spent close to three hours pulling numbers from four different tools to produce a report that leadership reviewed that afternoon. The team had talked about fixing it for over a year. The problem was that fixing it meant rebuilding something that was, technically, still working.

The first 30 days

In their first 30 days with Strand, the focus was connection and validation. They brought in Salesforce, Stripe, Mixpanel, and their internal PostgreSQL database. Cortex mapped the relationships automatically, and the team spent time verifying that the numbers coming out of Ask matched what they had been calculating manually. In most cases they did. In a few cases they did not, and tracing those discrepancies revealed two metric definitions that had drifted between tools without anyone noticing.

Day 90 and beyond

By day 90, ten of the 14 spreadsheets had been retired. The Monday report was replaced by a shared Ask query that any team member could run themselves. The three hours of weekly prep time had dropped to under 20 minutes of occasional spot-checking. The bigger shift was cultural: the ops team stopped being the bottleneck for data questions and started being the team that built the queries other people used.