Global Automotive Brand × Media Impact Modeling

Global Automotive Brand × Media Impact Modeling

Global Automotive Brand × Media Impact Modeling

Year

2020-2021

Year

2020-2021

Year

2020-2021

Type

Marketing Mix Modeling · Analytics · Executive Decision Support

Type

Marketing Mix Modeling · Analytics · Executive Decision Support

Type

Marketing Mix Modeling · Analytics · Executive Decision Support

Timeframe

52 weeks

Timeframe

52 weeks

Timeframe

52 weeks

Overview

Overview

Overview

A global automotive brand needed a clearer understanding of how media investment translated into real business outcomes — not just impressions or clicks.

The goal was to connect media activity to build-and-price behavior, credit applications, and ultimately vehicle sales, and to do so with enough confidence to support executive decision-making on a weekly basis.

I was engaged to design a modeling framework that could quantify media impact despite significant data blind spots downstream.

A global automotive brand needed a clearer understanding of how media investment translated into real business outcomes — not just impressions or clicks.

The goal was to connect media activity to build-and-price behavior, credit applications, and ultimately vehicle sales, and to do so with enough confidence to support executive decision-making on a weekly basis.

I was engaged to design a modeling framework that could quantify media impact despite significant data blind spots downstream.

A global automotive brand needed a clearer understanding of how media investment translated into real business outcomes — not just impressions or clicks.

The goal was to connect media activity to build-and-price behavior, credit applications, and ultimately vehicle sales, and to do so with enough confidence to support executive decision-making on a weekly basis.

I was engaged to design a modeling framework that could quantify media impact despite significant data blind spots downstream.

/01

/01

/01

Challenge

Challenge

Challenge

The brand faced three major challenges:

  • Disconnected performance signals — media data lived separately from build-and-price tools, finance applications, and sales

  • Tier 3 dealership opacity — dealership-level sales data operated as a black box, limiting direct attribution

  • Executive pressure for clarity — leadership needed defensible insight, not retrospective reporting

Traditional attribution approaches were insufficient given the lack of clean, end-to-end data.

The brand faced three major challenges:

  • Disconnected performance signals — media data lived separately from build-and-price tools, finance applications, and sales

  • Tier 3 dealership opacity — dealership-level sales data operated as a black box, limiting direct attribution

  • Executive pressure for clarity — leadership needed defensible insight, not retrospective reporting

Traditional attribution approaches were insufficient given the lack of clean, end-to-end data.

The brand faced three major challenges:

  • Disconnected performance signals — media data lived separately from build-and-price tools, finance applications, and sales

  • Tier 3 dealership opacity — dealership-level sales data operated as a black box, limiting direct attribution

  • Executive pressure for clarity — leadership needed defensible insight, not retrospective reporting

Traditional attribution approaches were insufficient given the lack of clean, end-to-end data.

/02

/02

/02

Approach

Approach

Approach

The solution was to shift from attribution to predictive, proxy-based modeling.

1. Proxy-Based Outcome Mapping
We mapped media activity to measurable downstream signals including:

  • Build-and-price interactions

  • Credit application volume

  • Regional sales indicators

These served as leading indicators where direct dealership data was unavailable.

2. Solving the Tier 3 Data Black Box
To work around opaque dealership data, we identified the closest behavioral and geographic proxies and used them as inputs into the model — effectively reconstructing impact where direct measurement wasn’t possible.

3. Fast, Multi-Regression MMM
I built a live, multi-regression marketing mix model that:

  • Ingested media activity in near real time

  • Modeled lagged and cumulative effects

  • Produced directional guidance rather than static post-campaign analysis

4. Executive-Ready Output
The model was designed specifically to support the CMO’s Monday executive meeting, providing a consistent, trusted view of:

  • Media contribution to demand

  • Expected impact on sales

  • Scenarios for reallocation and optimization

The solution was to shift from attribution to predictive, proxy-based modeling.

1. Proxy-Based Outcome Mapping
We mapped media activity to measurable downstream signals including:

  • Build-and-price interactions

  • Credit application volume

  • Regional sales indicators

These served as leading indicators where direct dealership data was unavailable.

2. Solving the Tier 3 Data Black Box
To work around opaque dealership data, we identified the closest behavioral and geographic proxies and used them as inputs into the model — effectively reconstructing impact where direct measurement wasn’t possible.

3. Fast, Multi-Regression MMM
I built a live, multi-regression marketing mix model that:

  • Ingested media activity in near real time

  • Modeled lagged and cumulative effects

  • Produced directional guidance rather than static post-campaign analysis

4. Executive-Ready Output
The model was designed specifically to support the CMO’s Monday executive meeting, providing a consistent, trusted view of:

  • Media contribution to demand

  • Expected impact on sales

  • Scenarios for reallocation and optimization

The solution was to shift from attribution to predictive, proxy-based modeling.

1. Proxy-Based Outcome Mapping
We mapped media activity to measurable downstream signals including:

  • Build-and-price interactions

  • Credit application volume

  • Regional sales indicators

These served as leading indicators where direct dealership data was unavailable.

2. Solving the Tier 3 Data Black Box
To work around opaque dealership data, we identified the closest behavioral and geographic proxies and used them as inputs into the model — effectively reconstructing impact where direct measurement wasn’t possible.

3. Fast, Multi-Regression MMM
I built a live, multi-regression marketing mix model that:

  • Ingested media activity in near real time

  • Modeled lagged and cumulative effects

  • Produced directional guidance rather than static post-campaign analysis

4. Executive-Ready Output
The model was designed specifically to support the CMO’s Monday executive meeting, providing a consistent, trusted view of:

  • Media contribution to demand

  • Expected impact on sales

  • Scenarios for reallocation and optimization

/03

/03

/03

Results

Results

Results

  • 85% accuracy in predicting vehicle sales based on media activity

  • Clear linkage established between media investment and build-and-price behavior, credit applications, and sales

  • Executive confidence increased through a repeatable, weekly decision framework

  • Reduced reliance on incomplete dealership reporting

  • 85% accuracy in predicting vehicle sales based on media activity

  • Clear linkage established between media investment and build-and-price behavior, credit applications, and sales

  • Executive confidence increased through a repeatable, weekly decision framework

  • Reduced reliance on incomplete dealership reporting

  • 85% accuracy in predicting vehicle sales based on media activity

  • Clear linkage established between media investment and build-and-price behavior, credit applications, and sales

  • Executive confidence increased through a repeatable, weekly decision framework

  • Reduced reliance on incomplete dealership reporting

Increase in online orders

+62%

+62%

Return customer rate

+40%

+40%

/04

/04

/04

This project was about translating purpose into design — creating a clear, thoughtful brand and digital experience that reflects the client’s values and supports their growth across every touchpoint.

This project was about translating purpose into design — creating a clear, thoughtful brand and digital experience that reflects the client’s values and supports their growth across every touchpoint.

This project was about translating purpose into design — creating a clear, thoughtful brand and digital experience that reflects the client’s values and supports their growth across every touchpoint.

Branding / Campaign

Golden Frame Agency

When Golden Frame was ready to launch a new service, they asked us to make a statement. We created a campaign identity that’s bold and expressive, using oversized typography, bright color accents, and social-ready visuals.