Data Architecture

Unified Lead Source Framework

Inconsistent lead‑source values across Hubspot and Salesforce caused misclassification and murky attribution. I created a governed taxonomy, engineered deterministic assignment logic and embedded controls across all ingestion paths. This normalized more than 200 legacy values, cut misclassified leads by 30 % and ensured 100 % of inbound channels were standardized.

The Problem

Marketing attribution is only as trustworthy as the data architecture underneath it. At an enterprise organization supporting multiple products and portfolios, teams were working with inconsistent source values, misclassified inbound leads, and incomplete attribution inputs. Reporting was noisy, funnel visibility was blurred, and marketers lacked confidence that performance data reflected reality.

I rebuilt the architecture behind how lead source data is defined, captured, governed, and consumed across systems. This meant normalizing 200+ legacy values into a single structured taxonomy, designing deterministic logic for source assignment, embedding governance into intake forms and UTMs, and aligning Salesforce and Hubspot around a shared source-of-truth.

The result was a clean, scalable attribution architecture that reduced misclassifications by 30%, restored accuracy in funnel and ROI reporting, and gave the business a reliable foundation for understanding channel performance and pipeline contribution. This work was recognized with Lenovo’s Marketing Catalyst Award for elevating revenue attribution and decision-making.

Lead source was unreliable because the underlying data architecture had

no unified rules, no hierarchy, and no governance to keep it stable.

The Solution

Built a Unified Taxonomy

I consolidated more than 200 inconsistent source values into a clean, hierarchical model that defined exactly how every channel, campaign, and interaction should be classified.

Engineered Deterministic Assignment Logic

I redesigned how every lead is assigned a source by creating rules that prioritize UTMs, form inputs, API signals, and controlled mappings across systems.

Governance with Detection, Not Just Standards

I embedded validation, logging, and alerting into all data ingestion paths so non-compliant values were automatically flagged, traceable, and correctable at the source.