Selina (Xian) Xiao

Building Scalable AI & Measurement Systems for Growth and Performance Marketing

I design and commercialize data-driven systems that quantify incremental impact, optimize cross-channel investment, and operationalize AI at enterprise scale.

Over 18 years across technology (Meta) and management consulting (Accenture, Merkle), I have led global marketing measurement initiatives supporting $12B+ in annual marketing investment, influencing executive decision-making, shaping growth strategy, and translating complex data into enterprise-level business outcomes.

Unlocking incremental revenue, accelerating long-term growth, and delivering sustained ROI performance.

Marketing Measurement Systems

Integrated incrementality testing, marketing mix modeling, and multi-touch attribution frameworks.

AI & Automation

LLM-powered creative optimization, ML-driven experimentation, and scalable AI production systems.

Customer Intelligence

Full-funnel audience strategy, personalization at scale, and lifecycle journey optimization.

Marketing Measurement System

A modern measurement framework integrating incrementality, MMM, and attribution to measure, learn, and optimize growth marketing investment.

Incrementality Measurement

A practical framework for evaluating marketing impact using experiments, quasi-experiments, and bias correction methods: focused on selecting the right approach under real-world constraints.

Experimentation Challenges & Solutions

Even well-designed experiments can produce misleading results. This session highlights common experimentation pitfalls and practical solutions for reliable decision-making.

Selected Work

Representative systems and case studies across measurement, optimization, and AI.

Case Study 01

Enterprise Marketing Mix Modeling & ROI Optimization

Built and commercialized MMM and incrementality frameworks across industries, supporting $12B+ in annual marketing investment. Designed to quantify true incremental impact and guide strategic budget allocation across paid and owned channels.

Designed Bayesian and mixed-model MMM calibrated with experiments
Quantified incremental sales, revenue, and ROI by channel
Built scenario planning tools for budget allocation and forecasting

A walkthrough of the MMM system

Click any tile to expand.

MMM End-to-End Workflow

MMM vs Experiments vs MTA

MMM Methods: Adstock and Diminishing Returns

MMM Dashboard: Contribution and ROI

Budget Allocation Tool

Case Study 02

Omnichannel Attribution & Next-Best-Action Optimization

Designed and operationalized a privacy-safe, user-level attribution framework to quantify cross-channel impact and drive next-best-action optimization.

Integrated de-identified user level 1P and 3P data via Clean Room
Applied Propensity Score Matching and Inverse Probability Weighting or Shapley Value approach to measure cross-channel effects
Translated attribution insights into next-best-action decisions across flighting and channel mix to maximize ROI.

User-level incrementality and in-flight optimization

Click any tile to expand.

Cleanroom Data Foundation

PSM + IPW Lift Estimation

Diminishing Return

Next Best Action

Case Study 03

AI-Powered Creative Optimization

Build LLM-based workflows to extract creative features, identify performance drivers, and auto-generate optimized assets for ongoing multivariate testing and production deployment.

Automated LLM-based creative feature extraction at scale
Identified high-impact creative assets through ML
Auto-generated creative assets for multivariate testing and production deployment

AI-powered creative optimization workflow

Click any tile to expand.

End-to-End Overview

Drivers of Performance

GenAI to Production

Contact

If you’d like to discuss marketing measurement, incrementality, or AI-powered optimization, feel free to reach out.

LinkedIn Email