Samuel Martin, stated plainly.
Lead, Product Analytics and Experimentation at Apree Health. Eight years of designing the experiment, building the measurement system, and then watching the decision change.
San Francisco Bay Area. Samuel Martin has led analytics across a book of 750+ enterprise clients and 5.5M users, built the unified attribution warehouse that cut financial reporting latency from weeks to minutes, and designed the outbound randomized trial that priced a channel at $99 cost per acquisition against $549 of first-year value.
He is looking for work where the measurement is the hard part and someone has to own it.
Every case study on this site runs in the same four steps.
Not a framework sold to anyone. It is the order the work actually happens in, and stating it here means a reader can check whether each case study kept to it.
Name the decision before touching data. If nobody can say what would change, the measurement is decoration.
Find the step nobody instrumented. It is usually not the step being blamed, because the blamed step is the one that was visible.
Design the experiment or the quasi-experiment, build the measurement so each step is countable, and report the weak points as weak.
A decision moved. If none did, the work is not finished, whatever the analysis says.
Direct, no form.
Email is read and answered. If you want the résumé first, all four variants download without a form or an email gate.
Most of this work sits behind protected health information and an enterprise agreement. Figures are redrawn on synthetic data, no client is named, and the method and reasoning are exact.