Details sanitized and shared with appropriate discretion. Specific metrics and artifacts available upon request.
Portfolio-Level Quantitative Study Informing Loyalty Strategy
Led a portfolio-level quantitative research program with 500k+ participants; SQL and Spotfire analysis contributed to a 7% NPS lift among premier members and improved executive reporting efficiency by 33%.
Detailed case study available on request
This project contains confidential process details and internal metrics. The full case study — including methodology decisions, process documentation, stakeholder artifacts, and specific outcome data — is available to hiring managers and recruiters.
Summary
United’s premier and loyalty strategy affects the travel experience of tens of millions of customers annually. I led a portfolio-level quantitative research program with 500,000+ participants to give the loyalty team and executive stakeholders the data they needed to make high-stakes decisions with confidence. The analytical work — using SQL, R, STATA, and Spotfire — directly informed premier strategy changes that contributed to a 7% NPS lift among premier members, and the reporting infrastructure I built improved executive reporting efficiency by 33%.
Role: Senior Research & Insights Analyst · Timeline: Nov 2021–Nov 2022 · Team: Loyalty & Premier product teams · Methods: large-scale survey, SQL analytics, statistical modeling, Spotfire reporting
Problem & context
[PLACEHOLDER: describe the strategic context — what decisions did the loyalty team need to make, what data did they have (or not have), and why was portfolio-level quantitative research the right approach here vs. qual or smaller-scale studies?]
At the scale of United’s premier program, intuition doesn’t work. Decisions about benefit structure, tier thresholds, and member communication affect hundreds of thousands of high-value customers whose behavior is highly sensitive to perceived value. The team needed research that could both characterize the current experience at population scale and model the likely impact of proposed changes.
My role
I owned the quantitative research end to end — study design, instrument development, SQL analysis, statistical modeling, and executive reporting. I also built the Spotfire reporting infrastructure that made these insights continuously accessible to leadership.
Approach & decisions
[PLACEHOLDER: describe the key methodological decisions — how did you design the study to capture both behavioral data (SQL) and attitudinal data (survey)? How did you handle the scale and quality of a 500k+ participant study? What statistical methods did you use and why?]
The study combined survey-based attitudinal measures with SQL-extracted behavioral data to triangulate the relationship between what premier members valued, how they actually used their benefits, and where the gap between the two was creating NPS drag. Linking attitudinal and behavioral data at this scale required careful design to ensure the analysis was interpretable without being misleading.
Process
[PLACEHOLDER: describe the study design phase, fielding, QA process, analysis pipeline, and how findings were delivered to executive stakeholders]
Deliverables
- Study instrument (survey design, screener, quality controls)
- SQL analysis pipeline for behavioral data extraction and joining
- Statistical models (R/STATA) quantifying NPS drivers and segment differences
- Spotfire executive dashboards for ongoing loyalty KPI monitoring
- Insight reports delivered to loyalty leadership and executive stakeholders
- [PLACEHOLDER: sanitized example of a Spotfire dashboard or executive report]
Impact & outcomes
- 7% NPS lift among premier members following loyalty strategy changes informed by this research.
- 33% improvement in executive reporting efficiency through the Spotfire infrastructure.
- [PLACEHOLDER: describe how specific findings shaped the loyalty strategy decisions — what changed as a result of the research, and how was impact measured?]
- Research at this scale established a quantitative research baseline that informed subsequent qual studies.
Reflection
[PLACEHOLDER: what did you learn about large-scale quantitative research, executive communication of statistical findings, or the relationship between attitudinal and behavioral data?]
The most useful thing I did in this project wasn’t the modeling — it was making the output legible. Executives don’t act on regression tables; they act on a clear narrative about what’s driving the number and what would move it. Investing time in the visualization and narrative framing was as important as the analysis itself.