6 Adımda Referral Program Deneyi Tasarla

Act as a Khan Academy tutor specializing in business experiments and academic research methods. Guide me step-by-step through designing a rigorous experiment for optimizing a referral program, using the Hypothesis-Test-Results format. Use [COMPANY NAME] as the practical example: a SaaS tool for [TARGET AUDİENCE], current referral conversion rate [CURRENT CONVERSİON RATE]. Step 1: Hypothesis Formation. Propose 3 testable hypotheses based on referral best practices (e.g., Dropbox-style incentives). Example: 'Offering [REWARD TYPE] to referrers will increase conversions by [EXPECTED LİFT]% over [TEST DURATİON].' Make them SMART and falsifiable. Step 2: Experimental Design. Detail the setup: Define control vs. treatment groups (e.g., 50/50 split of [SAMPLE SİZE] users). Specify variables: Independent ([INCENTİVE VARİATİON]), Dependent ([REFERRAL SİGNUPS]). Include randomization, blinding, and tools like Google Analytics or Mixpanel for tracking. Step 3: Data Collection Plan. Outline metrics (e.g., [KEY METRİCS]), sample size calculator for 95% confidence, and timeline ([TEST DURATİON]). Address biases like selection or novelty effects. Step 4: Analysis Framework. Teach statistical tests: t-test for means, p-value