Leon
Liu

Marketing Science and measurement analytics professional turning advertiser signals, experiments, and data pipelines into growth decisions, executive dashboards, and scalable revenue impact.

Ask about my experience View timeline
Range 7+ years across Meta, Horizon Media, and PHD
Mode Marketing Analytics x Ads Measurement x Data Visualization
Output $1B+ spend analysis, $5B opportunity sizing, 20% revenue lift
Experience system map 2014 2020 2024 Now Stats Marketing PHD Horizon Meta Sales AI

Case Study Lab

Day 1 prototype: a publishing system for turning technical methods into business-facing proof. Each case will show the real use case, concept, method, code artifact, output, limitations, and non-technical translation.

How to read these

Every case starts with a business decision.

The goal is to prove that I can choose the right measurement method, build the data workflow, communicate uncertainty, and translate the result into action for Marketing, Sales, Product, Engineering, or executives.

Decision

What should the business run, fix, stop, or scale?

Method

Which analytics approach fits the question and constraints?

Artifact

What code, chart, dashboard, model, or agent demonstrates the work?

Translation

How would I explain the result to a non-technical stakeholder?

Day 1 output

Reusable case anatomy

This is the page grammar I will reuse for MMT, MTA, MMM, lift studies, segmentation, dashboards, AI apps, vibe coding projects, and analytics agents.

Business question Real use case Concept Code Output Limitations Plain-English version
Case study index

Methods to publish next

All Measurement Pipeline AI Collaboration

Matched Market Test

MMT
Business question
Did the campaign create incremental sales when user-level randomization was not available?
Artifact
Synthetic market data, pre-period fit chart, lift readout, decision table.
Analyst Incrementality Day 2

Multi-Touch Attribution

MTA
Business question
Which touchpoints appear to influence conversion paths, and where can optimization start?
Artifact
Touchpoint SQL table, path model, channel contribution chart, causality caveat.
Technical Attribution Journey

Marketing Mix Modeling

MMM
Business question
How should the next budget dollar move across channels after accounting for time lag and saturation?
Artifact
Weekly synthetic data, adstock transform, response curves, allocation table.
Executive Budget Modeling

Lift Studies + Meta-Analysis

LIFT
Business question
What reliable pattern emerges after many brand or conversion lift tests?
Artifact
Lift readout, confidence interval explainer, forest plot, learning agenda.
Analyst Experiment Synthesis

Customer Segmentation

SEG
Business question
Which behavior-based customer groups deserve different messaging, media, or Sales treatment?
Artifact
Feature table, clustering workflow, segment cards, activation map.
Business Audience Activation

Pipeline to Tableau

DATA
Business question
How do scattered inputs become a repeatable dashboard trusted by executives and clients?
Artifact
SQL model, Python cleaning step, Snowflake schema, Tableau KPI spec, QA checklist.
Technical Pipeline Visualization

AI Web App Dashboard

APP
Business question
Can a dashboard become an interactive decision surface with AI-generated summaries?
Artifact
Local web app, synthetic dataset, chart interactions, recommendation panel, guardrails.
Product AI Dashboard

Vibe Coding Projects

BUILD
Business question
How can AI speed up analytics prototyping without weakening measurement rigor?
Artifact
Prompt, prototype screenshot, QA checklist, before/after workflow.
Product AI build QA

Analytics Agent

AGENT
Business question
Can an agent answer analytics questions from trusted context with citations and uncertainty?
Artifact
Data sources, prompt/rubric, good vs bad answers, evaluation method.
Technical Agent Evaluation
01 / BUSINESS

Question + context

Start with the decision, scale, stakeholders, and constraint. No method name until the reader knows why it matters.

02 / METHOD

Concept + code

Explain the method in plain English, then show a compact SQL, Python, R, dashboard, or agent artifact.

03 / DECISION

Output + translation

Close with chart, dashboard, model readout, limitation, and the non-technical version for Sales or executives.

Marketing analytics timeline

The arc starts with statistical modeling in applied meteorology, then moves into marketing intelligence, ads measurement, full-funnel analytics, data pipelines, and AI-enabled Marketing Science work at Meta.

Applied Meteorology, Sun Yat-sen University

Built a quantitative base through statistical modeling, mathematics, and atmospheric science before moving into marketing intelligence.

MS Marketing Intelligence, Fordham

Shifted the analytical foundation toward consumer behavior, marketing case work, and business-facing insight communication.

PHD Media: Google ads measurement

Built Tableau and Alteryx planning tools for Google's $100MM+ ad strategy, ran A/B tests on $5B+ campaign spend, and delivered matched-market incrementality analysis.

Horizon Media: analytics manager

Led analytics for Paramount+, AMC, NFL, Honda, and others; built Snowflake/Python pipelines and Tableau dashboards that tripled reporting efficiency.

Meta: Marketing Science Consultant

Evaluates $1B+ monthly advertiser spend, identifies 13% lower cost-per-result opportunities, and scales AI-assisted measurement workflows for 200+ partners.

Next: measurement systems at scale

The direction is deeper integration across Marketing Science, AI agents, CRM signals, Sales workflows, and productized measurement infrastructure.

Marketing Science capability map

The center of gravity is measurement analytics: experimentation, attribution, MMM/MTA, pipeline-to-dashboard systems, and cross-functional translation between advertisers, Sales, Product, Engineering, and Marketing teams.

Marketing Analytics Measurement Pipelines Visualization AI Workflows XFN Leadership
Marketing analytics
92
Ads measurement and incrementality
91
Data pipelines to visualization
88
Product, Sales, and Engineering collaboration
87
AI-assisted analytics workflows
84
Executive storytelling and sales enablement
86

Selected project logic

The project story is strongest when it connects measurement method, data infrastructure, stakeholder adoption, and business impact. These cards translate resume bullets into reusable proof patterns.

Case Type 01

Advertiser measurement and learning agendas

Designed experiments, Bayesian analysis, lift studies, matched-market testing, MMM/MTA thinking, and recommendation frameworks across Meta, Horizon, and PHD.

Case Type 02

Data pipelines to executive visualization

Built Tableau dashboards and Snowflake/Python/Alteryx pipelines that integrated media, first-party, CRM, and performance data, tripling reporting efficiency.

Case Type 03

Sales, Product, and Engineering collaboration

Partnered with Sales, Product, Engineering, MarTech, Marketing, and Content Strategy teams to turn advertiser context signals into CRM insight tools and scalable guidance.

How I explain analytics work

For measurement and analytics roles, proof is not just screenshots. It is the chain from business question to method, data system, stakeholder adoption, and measurable decision impact.

01 / QUESTION

Business question

Examples: why advertiser performance is decaying, what would have happened without media, which customer segment is under-served, or where Sales should focus.

02 / METHOD

Measurement method

Choose the right method for the decision: incrementality test, Bayesian analysis, matched-market design, MMM/MTA, segmentation, or journey analysis.

03 / PIPELINE

Data to visualization

Connect SQL, Python, Snowflake, Alteryx, Tableau, CRM signals, and media data into dashboards or tools that stakeholders can actually use.

04 / ADOPTION

Adoption and impact

Measure whether Sales, Product, Engineering, Marketing, or clients use the recommendation, and connect that adoption to revenue, efficiency, or better decisions.

Ask about my analytics experience

This first version answers from the resume-backed profile: Meta Marketing Science, Horizon analytics management, PHD media measurement, SQL/Python/Tableau pipelines, ads experimentation, and cross-functional GTM collaboration.

Hi, I am Leon Agent. I can answer using Leon's resume-backed profile: Marketing Science, ads measurement, data pipelines, visualization, AI-assisted analytics workflows, and cross-functional collaboration.