Resources

Clear definitions of the business metrics investors ask about, such as bookings, recurring revenue and lifetime value.

Why surprising wins are usually wrong, why low-powered tests exaggerate effects, and how to read p-values correctly.

How Spotify uses inferiority tests on guardrail metrics to decide when a change should be rolled back.

A free, rigorous and readable book on causal inference with Python code, from randomized tests to methods for when you cannot run one.

How to watch people use the real product in their own setting and ask about it as they go.

The check that catches broken experiments before anyone reads the results, with a taxonomy of the causes.

The method to reach for when a change shipped to one group or market and you cannot run an A/B test.

A product studio's own short, opinionated rules for naming and documenting events. Good to copy as a starting point.

Written for product engineers. What guardrail metrics are, which ones to pick, and how companies such as Airbnb use them.

How to keep a group of users without new features to measure long-term impact, and the trade-offs of doing so.

Uses the familiar "users of this feature retain better" case to show how a confounder fakes an effect and how to adjust for it.

How users are assigned to groups with a salted hash, so the same user always gets the same variant without stored state.

The classic explanation of why peeking at results and stopping early produces false wins, and what to do instead.

A compact look at hashing-based assignment and the two-step decision of who enters an experiment and which variant they get.

A short explanation with business examples of drawing conclusions only from the users who stayed.

A step-by-step method for turning notes, transcripts and feedback into themes you can defend.

The original "we believe, will result in, we will know when" template for writing testable product hypotheses.

Shows exactly when to call identify, how anonymous events get merged, and the mistakes that split one person into two.

Treats identity as an engineering problem across browser, mobile, server and third-party data, which is what it becomes in practice.

Conversion windows, step order and why you compare absolute drop-off, not percentages, to find the step that matters.

How to find the early signals that predict slow outcomes such as retention and revenue.

The original HEART paper. Happiness, engagement, adoption, retention and task success, plus the goals-signals-metrics process for choosing them.

Explains the object-action naming pattern and shows how messy event names break analysis later.

How to go from business questions to a small set of events and properties, with naming advice.

A practical walk through deciding what to track, with a tracking plan template and the reasons behind each choice.

How to work out the sample size a test needs before running it.

The standard long-form reference for MRR, churn, LTV and CAC, with the formulas and the reasoning behind them.

Interactive visualizations of distributions, inference and confidence intervals. The fastest way to build intuition for the basics.

How masking and blocking work in session recordings, so you can watch real sessions without capturing personal data.

A measured look at what NPS can and cannot tell you, from a researcher who works with the data.

Why a result that reverses across segments, or a win in one segment, should lead to a new test rather than a shipping decision.

A short, visual explanation of what power is and why sample size changes it.

How to measure whether people complete their tasks, including partial success. Old, and still the clearest explanation.

How cohorts work, how to read a retention grid, and common mistakes such as cohorts that are too broad.

The most complete free guide to choosing a North Star metric and the input metrics a team can actually move.

Replaces the single DAU/MAU number with a histogram of how many days people are active, which shows what averages hide.

A quick overview of the same method, for when you need the idea before the detail.

The standard reference on running experiments at scale, written by the people who built experimentation at Microsoft, Google and LinkedIn.

Retention benchmarks by business type, collected from growth practitioners, and why the shape of the curve matters more than one number.