Attribution · Data · AI
IT STEP Academy
Connect advertising spend, website activity and CRM requests into a consistent attribution and performance-monitoring system.
Attribution, Data Platforms & AI / IT STEP Academy
The decisions behind the product.
01Problem
Advertising, traffic and CRM records need to form one traceable view of acquisition, performance and follow-up tasks.
02Engineering challenge
Connect advertising spend, website activity and CRM requests into a consistent attribution and performance-monitoring system.
03System architecture
Campaignsource contextCRM requestSource recordtransforms intoWarehouse factMetric definitionfeedsDashboard04Implementation
- 01
Separate ingestion, warehouse and semantic layers align advertising, traffic and CRM records before they reach executive dashboards.
- 02
Application identity and tasks resolve from Core; IT STEP analytics resolves from its analytics database. Backend guards and workspace-aware navigation enforce this boundary.
- 03
Attribution views, metric monitoring and AI-agent workflows connect data preparation with analysis and follow-up work. Source freshness remains visible as a separate operational concern.
Design decision & trade-off
Separate raw ingestion, warehouse and semantic layers so source records and business interpretation can evolve independently. Core supplies identity/tasks; a separate analytics database supplies IT STEP reporting.
Quality & operational controls
Workspace-aware backend guards and navigation preserve the analytics boundary. Source freshness is tracked separately from metric values; a connected provider does not imply complete attribution coverage.
- PostgreSQL
- Python
- FastAPI
- Next.js
- TypeScript
- n8n
One complete workflow
- 01Collect sources
- 02Normalize records
- 03Apply attribution
- 04Review performance
- 05Assign follow-up
05Evidence
Owner-confirmed frontend/backend, attribution, databases, monitoring, tasks and AI-agent scope; archived tenant audit covering Meta, Google Ads, GA4, ITCRM and the raw/DWH/semantic model. Client figures remain under NDA.
Some project data is protected by NDA. Restricted internal screens and figures are omitted; the diagrams show the scope that can be shared publicly.
06Outcome & ownership
Delivered
Delivered the data and application layers for end-to-end analytics: source ingestion, databases, attribution, frontend/backend, monitoring and AI-agent workflows.
Role & scope
Engineering Project Lead · attribution, databases, frontend/backend, metric monitoring and AI agents
Scope of the result
Personal responsibility is engineering project leadership across the listed data and application layers. The architecture evidence is historical; current tenant uptime, attribution accuracy and ROI are not claimed.
Discuss my project
What would you like to improve?
Tell me what your business does and what you want to change. We will work out the first step.
- DiscoveryA map of current processes, problems and data sources.
- StrategyThe target operating model and change priorities.
- ArchitectureRoles, modules, integrations and acceptance criteria.