Skip to content
Li Tan · 谭李 · Lin.
About On measurement, signal, and quieter craft.

Li Tan — on measurement, signal, and quieter craft.

Ten years ago I wandered into Silicon Valley asking a narrow question — how do we know this marketing spend is actually working? — and have been chasing variations of that question ever since: what is the true incremental impact, how do we measure when a randomized experiment is impossible, which metrics actually predict long-term growth.

The journey took me through different corners of tech. At NerdWallet I worked on helping people make smarter financial decisions; at Opendoor, understanding how people buy and sell homes. Different industries, same core problem: turning messy data into clear decisions.

Something shifted recently. I realized AI was not just another tool to add to the stack — it was fundamentally changing how data work gets done. Tasks that used to take a week now take hours. I have rebuilt my workflow around this insight, and this site is where I document what I learn along the way.

I also bring a perspective that is harder to find: deep roots in China combined with a decade in the Bay Area. I see patterns at the intersection of these two ecosystems that others miss. This site is, in part, my way of making those patterns legible.

How I think

  1. Rigor over speed

    A well-designed analysis that takes a week beats a quick dashboard that gives wrong answers. Shortcuts in measurement compound into bad decisions.

  2. Clarity over complexity

    The best analysis is one stakeholders can understand and act on. If your model only convinces other modelers, it does not matter.

  3. Impact over output

    Lines of code, dashboards, slide counts — none of it matters if it does not change what someone decides to do on Monday.

  4. Write to think

    The discipline of explaining a method in plain English is the same discipline that catches the flaw in it. The essays here are as much for me as for you.

Wide Bay Area landscape
Bay Area · 湾区 · the room where the work happens · image landing in Phase 7

Career at a glance

  • 2023 — present
    Senior Data Scientist Opendoor · Real estate marketplace
  • 2019 — 2023
    Senior Data Scientist, Marketing NerdWallet · Personal finance
  • 2014 — 2019
    Analytics & measurement roles Bay Area tech · Growth & monetization
  • 2013
    MS, Analytics University of San Francisco

Craft & tools

Methods
Causal inference (DiD, synthetic control, RDD) · A/B testing & experimental design · Marketing mix modeling · Multi-touch attribution · Retention & lifecycle modeling
Stack
Python · SQL · R · BigQuery · dbt · Looker · Airflow · Git
AI / ML
LLMs & agentic workflows · Prompt engineering · Retrieval-augmented pipelines · Predictive modeling · XGBoost & scikit-learn
Writing
Technical essays · Internal memos to execs · Teaching notes
Selected links