Eleven weeks. $1.4M of leaked revenue, recovered.
+10% on $14M baseline
One source of truth
Vs prior 6-month RFP cycle
Ten retired, one survived
The problem.
Five paid channels. Six tools telling them what was working. Meta said CAC was $42. Triple Whale said $61. GA4 said $58. Klaviyo flow attribution said three different numbers depending on the lookback window. The bank statement said something else. The team was approving weekly budget shifts on a CAC figure no one in the room could defend. Two previous consultants had delivered more dashboards. They had eleven. They still didn't have an answer.
The approach.
Two weeks of stack archaeology — every join, every gap, every place a UTM died in transit. Three weeks rebuilding a probabilistic attribution model on first-party data (orders, sessions, email engagement, paid touchpoints), validated against a holdout of three product launches. Four weeks of board-and-team alignment: collapsing eleven dashboards into one weekly executive screen, archiving the rest. Two weeks of handoff — every dbt model, every Looker tile, every query documented so the in-house analytics lead could own it without me.
The outcome.
The new model put $1.4M of annual revenue back onto channels that last-click had been quietly handing to Meta and Klaviyo. The team paused two underperforming campaigns and reallocated $180K/month into a partner channel the old view had been calling breakeven (it wasn't). Weekly revenue meetings dropped from 90 minutes to 20. The CEO ran the meeting the week I left.
“He found the revenue we'd been leaving on the table for two years — in eleven weeks. Then he taught us how to keep finding it.”
If your dashboards disagree and your team approves budgets on whichever number tells the best story this week — let's talk. One slot open for Q3 2026.