How would you measure the success of improvements made to Uber's referral system?

To measure the success of improvements to Uber's referral system, I would focus on key performance indicators (KPIs) that directly reflect the system's effectiveness in driving new user acquisition and engagement. These would include:

1. Referral Conversion Rate: The percentage of referred users who sign up and complete their first ride. This is a primary indicator of how well the referral incentives and process are working.
2. New User Acquisition Cost (CAC) via Referrals: Comparing the cost of referral incentives to the number of new users acquired through the program. A successful improvement should lower this cost.
3. Referral Program Participation Rate: The percentage of existing users who actively send out referral invites. An increase here suggests the program is more appealing or easier to use.
4. Retention Rate of Referred Users: Tracking how long referred users remain active on the platform compared to users acquired through other channels. Higher retention indicates the referred users are a valuable cohort.
5. Referral Invite Click-Through Rate (CTR): For digital invites, the percentage of recipients who click on the referral link. This measures the initial engagement with the referral message.
6. Average Rides per Referred User: To understand the long-term value and engagement of users acquired through referrals.

I would also consider A/B testing different referral program variations (e.g., different incentive structures, messaging, or user flows) to isolate the impact of specific changes and identify the most effective strategies.