What I do, and how I do it.

Workforce Prediction Models

Predictive modeling for the outcomes that matter most — turnover, performance, promotion readiness, engagement risk. I build models trained on your data, validated rigorously, and designed to produce actionable outputs rather than just interesting statistics.

This might look like:

• A turnover risk score for each employee, updated monthly, with the key drivers surfaced
• A performance classification model for hiring decisions in high-volume roles
• A promotion-readiness index built from behavioral observations and assessment data
• A quantified estimate of how much your current selection process is costing you

I work in R and Python. I build models people can actually understand and use — not black boxes.

Measurement Design

Most organizations try to measure HR outcomes without first asking whether their instruments actually measure what they think they do. I help you define what you want to measure, build or select instruments that measure it, and validate that they work.

This might look like:

• Defining a construct (e.g., "high potential") rigorously enough that it can actually be measured
• Auditing existing metrics to assess whether they measure what you claim
• Building a measurement framework that connects people data to business outcomes
• Applying decomposition and Fermi estimation to quantities you thought were unmeasurable

Decision Analysis

Some HR questions don't need more data — they need better decision framing. I apply decision analysis techniques to help leaders think clearly about high-stakes people decisions under uncertainty.

This might look like:

• Quantifying the value of information — is it worth running that survey or assessment before deciding?
• Modeling the distribution of outcomes for a restructuring or workforce change
• Calibration workshops to improve how your team estimates and reasons under uncertainty
• Building a decision model for a make/buy/develop talent question

Grounded in the work of Howard Raiffa, Ron Howard, and Douglas Hubbard.

Survey & Assessment Design

Rigorously designed employee surveys and selection assessments — built on psychometric principles, not intuition and convenience. If you're going to ask people questions and act on the answers, the questions should be worth asking.

This might look like:

• Employee engagement or culture surveys designed with construct validity in mind
• Selection instruments developed and validated against job performance criteria
• Post-survey analysis that goes beyond means to model drivers of key outcomes
• Assessment audits — evaluating whether your current tools predict what you think they predict