Understand what's driving revenue. Not just what ad platforms report.
I use AI-assisted analysis, Python, and ecommerce data to evaluate advertising performance across channels, investigate attribution differences, and turn complex reporting into clearer marketing decisions.
Every marketing platform tells a different story.
Google Ads reports conversions. Meta Ads attributes purchases to campaigns. Email platforms claim influenced revenue. Shopify records actual orders. The numbers often disagree—and adding them together can double-count the same sale.
Fragmented data
Performance lives across disconnected ad, analytics, email, and ecommerce systems.
Conflicting attribution
Different platforms, lookback windows, and rules can assign credit for the same transaction.
Unclear decisions
Inconsistent numbers make it difficult to tell which investments merit more attention.
AI-assisted analysis grounded in ecommerce data.
I combine hands-on performance marketing experience with Python-based analysis and AI-assisted workflows to investigate what's changing and what may be driving results.
Advertising performance
Compare spend, attributed revenue, ROAS, conversion rates, and trends across campaigns and channels.
Customer acquisition
Evaluate customer acquisition costs, first-time customer performance, average order value, and repeat behavior where data supports it.
Data reconciliation
Investigate differences between platform-attributed conversions and actual ecommerce transactions, including tracking issues and reporting definitions.
AI-assisted investigation
Use AI to accelerate data exploration, surface anomalies, formulate hypotheses, and communicate findings—without treating generated explanations as proof.
A sale can have more than one marketing touchpoint.
Consider a customer who discovers a product through a Meta ad, returns via Google Search, and finally purchases after clicking an email. Different attribution models assign credit differently, even though the total transaction revenue is unchanged.
Explore a $200 customer journey
Interactive example · fictional dataChoose an attribution model:
Demonstration only. This illustrates model mechanics, not measured incrementality or proof that each touchpoint caused the purchase. Actual attribution requires reliable, appropriately collected journey data.
What attribution can help explore
- Which channels appear early versus late in observed journeys?
- How do rankings change under different models?
- Where do platform claims diverge from order revenue?
What attribution alone can't prove
- Whether an ad caused a purchase that otherwise would not have happened.
- Every interaction across devices, browsers, or offline channels.
- A single definitive answer to budget allocation without additional testing.
From raw marketing exports to actionable findings.
The goal is not simply more dashboards; it is a repeatable process for turning data into decisions.
Collect & standardize
Bring together available ad-platform, website analytics, email, and ecommerce reports. Normalize dates, metrics, and campaign labels.
Validate & reconcile
Check data quality, reporting windows, conversion definitions, and totals against transaction records.
Analyze & compare
Evaluate trends, segments, channel performance, attribution perspectives, and potential tracking anomalies.
Translate into action
Use Python, visual dashboards, and AI-assisted summaries to communicate findings, uncertainty, and possible next steps.
Tools and methods
My work draws on ecommerce platforms, campaign reporting, structured data analysis, and automation.
- Python and pandas for repeatable analysis
- SQL and ShopifyQL for commerce reporting
- Streamlit and Looker Studio for visualization
- GA4, Google Ads, Meta Ads, and Klaviyo data
- AI-assisted analysis and documentation
AI supports exploration and communication; calculations and attribution assumptions should remain auditable.
Explore the projects behind my approach.
These repositories show the tools and analytical workflows I use and develop. They are evidence of technical work, not a claim that every attribution method shown here is already deployed in production.
Ecommerce Marketing Dashboard
Interactive ecommerce reporting and analytics built with Python and Streamlit.
View on GitHub ↗Ecommerce Python Analytics
Python scripts and analytical workflows for marketing and ecommerce datasets.
View on GitHub ↗Marketing AI & Automation
An exploration of reusable AI-assisted marketing analysis and automation workflows.
View on GitHub ↗Better measurement. Better marketing decisions.
If your ecommerce reporting is spread across ad platforms, email tools, analytics, and Shopify, I can help evaluate the data and identify where a clearer measurement framework could make a difference.
Connect through my website ↗