Running your own experiments
You own the metric review and you're the call-it-or-not person. The team is watching what you ship, and you'd rather not learn the lessons in production.
An A/B testing simulator for what courses don't teach you. Before a real launch makes you do it live.
The gap nobody prepares you for
Your tool runs the math. Someone still has to make the call.
Picking the right primary. Calling a mixed result. Holding the line on day six when the CEO wants to ship. None of those are math problems.
Most PMs learn them in production, on a real launch, with the team watching. There is a better way.
Why this matters now
Statsig, Datadog, Eppo put PMs in the driver's seat. They tell you the p-value. They don't tell you whether the test was designed right, or whether to ship a mixed result.
Specs, summaries, status updates: those are AI's job now. The hard calls are still yours: when to ship, when to hold, when to revert. That's the work that gets you promoted.
Most PMs build that judgment on a real launch, in public, with the team watching. There is a better way to build it. One where getting it wrong doesn't cost you a quarter.
Who this is for
You own the metric review and you're the call-it-or-not person. The team is watching what you ship, and you'd rather not learn the lessons in production.
Your team ships two or three tests a week and you're not yet the one writing the readout. You'd like to be the PM who can.
You'll be asked to design a test on a whiteboard and read a mixed result out loud. You'd like to walk in with reps, not theory.
How it works
Every scenario simulates a complete experiment. You walk through the same six phases a real test goes through, and you get scored on the calls you make at each one.
You walk into a brief. The team has a change in mind. Stakeholders weigh in. You decide whether to invest investigation budget before designing.
Hypothesis, metric plan, statistical parameters. You make tradeoffs across MDE, power, and duration. The sample-size calculator updates as you decide.
The experiment plays out day by day. Curveballs appear in real time. You can pause to peek, or hold the line until the planned end date.
Final results land. You read the primary, secondaries, guardrails, and data-quality flags before making your call.
Ship, iterate, or revert. You write a rationale your team could read tomorrow morning. No second chances.
The ground truth is revealed. You see what you missed, what you got right, and which judgment habits hold up under pressure.
Featured scenarios
Each scenario is a complete experiment with stakeholders, data, curveballs, and a real decision at the end. You will probably catch yourself reaching for the wrong call at least once. That is the point.
Marketing wants to swap the homepage signup CTA from 'Get Started' to 'Start Free Trial'.
Preview
Control
Treatment
The product team consolidated a four-step checkout into a single page to reduce drop-off.
Preview
Control
Step 2 of 4: Payment
Treatment
Address
Payment
Summary
Growth shortened a five-step onboarding flow to two steps to lift activation.
Preview
Control
Treatment
Marketing replaced the homepage Featured Categories grid with a personalized 'Recommended for You' carousel.
Preview
Control
Featured Categories
Treatment
Recommended for You
A senior designer wants to change the mobile checkout button from blue to green based on a color-psychology blog post.
Preview
Control
Mobile checkout
Treatment
Mobile checkout
What changes
Six PM skills, before and after. Each one is something you'll do on a real launch in the next six months.
And one tangible thing
Finish the track and you walk away with a practitioner certificate. Put it on LinkedIn. Reference it in interviews. It is a record that you have sat through the decisions A/B testing actually rewards, made them under pressure, and seen what happened. Most PMs cannot show that.
Sample
Edition · 2026
Practitioner Certificate
This certifies that
with Distinction
has earned the Practitioner Certificate in A/B Testing Foundations, demonstrating rigorous experimentation judgment across five end-to-end A/B tests.
Curriculum
A/B Testing Foundations
Scenarios
5 of 5
Issued
May 5, 2026
Who built this

Senior Data Scientist · Founder, Experiment Lab
I'm a Senior Data Scientist with 8+ years in Data and Analytics, with a strong background in Product Analytics, working embedded inside product teams next to the PMs making the calls.
The work has been designing experiments, choosing metrics, making sense of inconclusive results, and translating them into the decisions that actually shipped. The same handful of judgment calls came up, team after team, unrehearsed and under pressure.
I've watched companies buy experimentation tools to help their teams ship faster. The tools handled the math. The judgment calls still landed on the PM, and most weren't ready for it.
Experiment Lab is the practice ground I wished every PM had before their first real launch.
Pricing
You finish the track in a few weeks. We don't think you should keep paying for that. Buy once, return whenever you need a refresher before a launch or an interview.
Individual
Launch price · First 50 buyers
For product managers who want the practice ground for themselves.
Five end-to-end scenarios · self-paced
Teams
For experimentation, growth, and product teams.
Frequently asked