How to evaluate emoji reactions?

A messaging app plans to launch an emoji reactions feature. Users can react to a message by long-pressing the message for 5 seconds and selecting an emoji. The company wants to know whether this feature creates value and whether it should be launched broadly.

Describe how you would evaluate this feature from a product analytics and experimentation perspective.

Your answer should address:

1. Product objective: What user or business problem might emoji reactions solve?
2. Success metrics: What primary, secondary, and guardrail metrics would you track? Consider both sender-side and receiver-side behavior, short-term engagement, and potential negative effects.
3. Experiment design: How would you design an A/B test for this feature? Specify the unit of randomization, treatment definition, likely sources of interference/network effects, experiment duration, and how you would think about power and minimum detectable effect.
4. Measurement details: What events would you instrument? How would you define adoption, active usage, retention impact, and quality of conversations?
5. Risks and confounding: What biases or pitfalls could make the feature appear better or worse than it really is? Consider novelty effects, heterogeneous treatment effects, long-press friction, message volume differences, and power-user concentration.
6. Decision framework: Under what conditions would you recommend full launch, iteration, or rollback?
7. Executive communication: How would you summarize the results for a C-level audience that cares about growth, engagement, and user experience rather than statistical detail?

Assume the app is consumer-facing, has an existing messaging product, and can run a randomized experiment.