Cover Image for How I Amplitude: When Your Experiment Setup Truly Reflects Your Product Strategy
Cover Image for How I Amplitude: When Your Experiment Setup Truly Reflects Your Product Strategy
Avatar for Amplitude Community

How I Amplitude: When Your Experiment Setup Truly Reflects Your Product Strategy

Zoom
Registration
Welcome! To join the event, please register below.
About Event

At Photoroom, experimentation plays a critical role in how the product evolves — but not all experiment results are created equal. In this How I Amplitude session, Charlotte de Thiersant, Senior Product Data Analyst at Photoroom, shares how her team uncovered a subtle but impactful limitation in how experiment funnels are commonly calculated — and how that discovery directly influenced a new experimentation feature in Amplitude.

Charlotte will walk through why exposure events are not just a technical detail, but a core design decision that defines who enters an experiment analysis and how results should be interpreted — especially when using sequential testing. She’ll explain the difference between assignment and exposure, why many teams unknowingly optimize for the wrong denominator, and how counting unique users instead of funnel starts can lead to conclusions that don’t reflect how products are actually used.

Through concrete examples from Photoroom, Charlotte will show how this mismatch surfaced while optimizing repeated workflows, why it took time to identify the root cause, and how close collaboration with the Amplitude Experiments team led to a new funnel calculation methodology that better aligns experiment results with real product behavior.

This session is designed for teams who want more than default experiment settings. If you’ve ever felt your experiment results didn’t quite match what users were doing in your product, this talk will give you a practical framework for designing experiments — and analyses — you can truly trust.

What you’ll learn

  • The difference between assignment and exposure, and why it matters in sequential testing

  • How exposure events define experiment denominators and stopping behavior

  • Why many teams unknowingly measure unique users when they really care about workflow performance

  • When exposure choices introduce bias — and how to validate them safely

Who should join?

  • Analysts

  • Product Managers

  • Experimentation and Growth teams

  • SaaS teams optimizing multi-step or repeatable workflows

Can't make it to the live session? Register for the event and we'll make sure you receive the recap!

Avatar for Amplitude Community