Data & Analytics· Marketing Mix Modeling · 2025 to 2026

Ipsos MMA

At Ipsos, I supported marketing mix modeling engagements that guided over $3B in annual marketing spend for Fortune 500 clients across energy, automotive, financial services, CPG, beverage, and electronics.

The Role

I worked as part of a team, sitting between the modeling group and the client to make sure models were built correctly and delivered on time. I was the connective tissue between the analytics being produced and the decisions it needed to inform.

Quality & Clarity

My core job was quality and clarity. I owned data validation and QA on the model outputs, using Excel, SQL, and Python to check the numbers and catch inconsistencies before anything reached a client. I then translated those technical results into clear, client-facing briefs and recommendations that senior stakeholders could act on directly.

Cross-Functional Delivery

The work was cross-functional by nature. I coordinated across data, modeling, and client teams, including international offices, to keep inputs consistent and modeling cycles running smoothly. I managed this workflow across six concurrent Fortune 500 engagements at once, balancing client-facing communication with the technical QA work happening behind the scenes.

Key Projects

48-Model Delivery

The largest engagement the team had handled: 48 nameplates for an automotive client, each with its own marketing mix model, delivered on four cadences at once: weekly, monthly, quarterly, and annually. Engineering sat offshore and client-facing work with a small US team, and I was in the middle, building the QA processes (structured validation checks, clear handoff points, consistent standards) that made delivery trackable across all 48 instead of reactive.

Model Version Control

The models were large, shared Excel files with no version control, hardcoded formulas, and links to files that had moved, so they crashed constantly, and fixes took days across time zones. I rebuilt them as a system: a structured template separating inputs, calculations, and outputs, self-contained with no external links, plus dedicated per-model channels and lightweight change logs. Crashes were eliminated, and cross-timezone handoffs became traceable.

Tools & Methods
Excel
Model-output QA, structured validation, and consistency checks across engagements.
SQL
Querying and cross-checking model input and output data to trace inconsistencies to their source.
Python
Automating validation and flagging inconsistencies at scale, before anything reached a client.
PowerPoint
Client-facing decks that turn model results into the narrative and recommendations senior stakeholders act on.
Data Validation & QA
Owning output accuracy, the last check between the modeling group and the client.
Client Briefs
Translating technical model results into clear, actionable recommendations senior stakeholders could act on.
Cross-Functional Coordination
Aligning data, modeling, and client teams across international offices and six concurrent engagements.
← All work