Pedro de Pauladados · infra · ia← Back to work

Interactive case

Power BI · Looker · Web

Multi-platform BI and a web analytics product

From media and CRM extraction to a custom experience for acquisition, funnel, sales, attribution and data-health analysis.

Live product preview

A decision interface, not a gallery of charts.

The public demo uses a deterministic, fully synthetic 90-day dataset. No client, account or credential data is present.

Marketing and Growth BI overview with KPIs, daily revenue, funnel and channel attributionOpen interactive demo ↗

Challenge

Start with the decision, not the chart.

Marketing operations depend on sources with different granularities and definitions. Without a shared layer, media, CRM and checkout tell incompatible stories.

Architecture

A traceable path from source to experience.

Google Ads, Meta Ads and CRM connectors feed a normalized contract. Analytical models calculate traceable metrics and an API serves only aggregated data to the dashboard.

  1. 01Scheduled extraction through APIs and n8n
  2. 02Event normalization and deduplication
  3. 03Acquisition, conversion and revenue models
  4. 04Freshness and reconciliation checks
  5. 05Responsive web analytics product

What becomes visible

Analysis tied to operational questions.

  • 01Spend, CPL, CPA and ROAS
  • 02Leads, CRM and funnel progression
  • 03Checkout, sales and net revenue
  • 04Campaigns, channels and attribution
  • 05Freshness, discrepancies and incidents

What you can explore

Follow the result, explain the cause, choose the action.

  • 01Compare the prior period or target
  • 02Filter projects, dates, channels and campaigns
  • 03Trace losses from lead to approved sale
  • 04Evaluate spend × ROAS at campaign grain
  • 05Inspect freshness, reconciliation and incidents
  • 06Open campaign diagnostics and export the filtered table
Sources that can be integratedGoogle AdsMeta AdsCRMCheckoutLead surveysn8n and dbt

Outcome

A reusable analytics-product model: rules stay documented, operations can see data quality and the interface evolves without depending on a single BI platform.