6.6%
Qualified → engaged
The largest modeled funnel bottleneck, pointing to engagement quality rather than more top-of-funnel volume.
Case study / Revenue analytics
GTM Revenue Analytics
From disconnected activity to one validated, decision-ready reporting layer.
I built an end-to-end portfolio system that consolidates lead, CRM, email, SMS and calls, LinkedIn, Loom, website, bookings, sales, verification, and paid ads activity from twelve source files.
A from-scratch rebuild of the reporting architecture I designed at KP Media Group — reconstructed here with placeholder data to protect client confidentiality. The architecture, logic, and modeling are unchanged.
Problem / Approach
Make the full funnel comparable.
Activity lived at different grains and used different identifiers. The build preserves source detail, aligns shared business entities, and exposes curated views so leaders can evaluate funnel, channel, campaign, and paid performance without manually reconciling exports.
Architecture / Data flow
01
12 source CSVs
Lead, CRM, engagement, website, bookings, sales, verification, and paid ads
02
PostgreSQL staging
Raw source tables preserved for validation and controlled transformation
03
Analytics model
Five shared dimensions and eleven activity-focused fact tables
04
Reporting layer
Executive funnel, channel performance, and campaign performance views
05
Power BI
Semantic model, measures, campaign/date filters, and five dashboard pages
Decisions that protect trust
- Qualified meeting requires both showed_up and is_qualified to be true.
- Channel booking signals can overlap, so they are not added together as official bookings.
- Sales campaign attribution follows the associated booking rather than a separate guess.
- Source row counts, duplicates, transformations, reporting totals, and filtered Power BI measures were checked.
What the model enables
Executive funnel health
Channel-level engagement
Campaign efficiency
Paid post-click performance
Modeled insights / Not client outcomes
Three signals that guide the next operating question.
151
Email booking signals
Email supplied the strongest scalable booking-intent volume in the modeled channel view.
8,885 / 23
Sessions / booking signals
Paid traffic reached the site, but post-click conversion warrants closer attention in this modeled dataset.
Five decision views
The static pages below show the final Power BI reporting experience. Use the controls to review each view.
Dashboard / 01
Executive Overview
How is the modeled GTM engine performing, and where should attention go next?
Build the operating layer