You work on a B2B SaaS product with a product-led growth motion and want to understand the incremental impact of lifecycle nudges, in-app prompts, and upgrade surfaces on activation and paid conversion. Rather than testing one surface in isolation, you are considering keeping a small group of eligible users out of these growth interventions to measure the overall lift of the PLG system.
How would you use a holdout group in this product-led growth motion? Walk through how you would define the hypothesis, choose metrics, design the experiment, and decide whether the measured lift is strong enough to keep the PLG treatments on.