StretchLearn Course

Run A/B tests you can actually trust

The statistics, process, and tools real growth teams use to turn traffic into reliable decisions

Beginner9 hr 40 minSelf PacedRegistered

Course Overview

What this course is designed to develop

Most A/B tests fail not because the idea was wrong but because the test was run wrong — too small, stopped too early, or read with the wrong math. This course gives you the statistics, process, and tooling that real growth and product teams use, with named platforms, sample-size numbers, and worked examples you can copy. You will leave able to design a valid experiment, calculate how long it must run, and turn results into decisions and a documented program.

Learning Outcomes

What the learner should be able to understand, build, or execute.

Write a testable hypothesis that names the change, the metric, and the expected direction and size

Calculate the required sample size and runtime for a test using baseline rate, MDE, and power

Interpret p-values, confidence intervals, and statistical power without the seven most common errors

Choose the right test type and tool — frequentist vs Bayesian, Optimizely vs GrowthBook vs VWO — for your traffic

Prioritise a backlog of test ideas using the ICE and PIE scoring frameworks

Run a clean experiment end to end and document the result so the whole team learns from it

Curriculum Preview

Inside the curriculum: a structured path from fundamentals to execution.

Preview the course structure, see how the modules build on one another, and understand the path this program is designed to take you through.

Module 1

Module 1: What Experimentation Is and Why Most Tests Lie

Build accurate intuition for what an A/B test actually proves, why so many tests mislead teams, and the mindset that separates a real experiment from guessing with extra steps.

3 lessons

The Logic of a Controlled ExperimentContent · 45 min

Preview Enabled

The Seven Ways A/B Tests MisleadContent · 45 min

LMS Access

Where Experimentation Pays Off and Where It Does NotContent · 45 min

LMS Access

Module 2

Module 2: The Statistics You Actually Need

Just enough applied statistics — hypotheses, significance, power, and sample size — to design and read a test correctly, taught with worked numbers rather than proofs.

3 lessons

Hypotheses, Significance, and the p-valueContent · 50 min

LMS Access

Power, Effect Size, and Confidence IntervalsContent · 50 min

LMS Access

Calculating Sample Size and RuntimeContent · 55 min

LMS Access

Module 3

Module 3: Designing and Running a Clean Test

Turn statistics into practice: generate good hypotheses, prioritise them, pick the right tool, and execute one experiment end to end without contaminating it.

3 lessons

From Insight to a Testable HypothesisContent · 50 min

LMS Access

Prioritising the Backlog with ICE and PIEContent · 45 min

LMS Access

Choosing Tools and Running It Without ContaminationContent · 50 min

LMS Access

Module 4

Module 4: Reading Results and Building a Program

Interpret outcomes honestly, decide what to do with wins, losses, and flat results, and turn one-off tests into a durable, documented experimentation culture.

3 lessons

Interpreting the Result HonestlyContent · 50 min

LMS Access

Acting on Wins, Losses, and Flat TestsContent · 45 min

LMS Access

Building a Culture of ExperimentationContent · 50 min

LMS Access

Built for Application

A complete learning path, not a one-off inspiration hit.

This program is designed around progression: focused lessons, structured modules, applied resources, assessments, and a course rhythm that turns information into usable capability.

a/b testingexperimentationconversion optimizationstatistical significancesample sizehypothesis testinggrowthdata-driven decisions