Hi,

We are integrating the Experiments SDK into our existing iOS app, which already uses Amplitude Analytics.

Our experiment involves testing two different onboarding flows, and we want to measure which performs better.

Current approach:

• We initialize the SDK using initializeWithAmplitudeAnalytics.

• Our configuration sets fetchOnStart(false).

• We call experiment.start(user).

• We implement a short delay (displaying a splash screen briefly) to allow sufficient time to retrieve the variant assignment.

Could you please help clarify the following points?

1. Is this approach (using local evaluation with a slight delay) suitable for instantly evaluating onboarding experiments?

2. To avoid targeting existing users who've already completed onboarding, we're considering setting a user property (e.g., finished-old-onboarding = true). Is this the recommended method?

3. Regarding app reinstalls: Each reinstall generates a new device ID. For consistent testing, should we manually persist the device ID (for example, in Keychain storage)?

Thanks in advance for your support!