Full talk · edited transcript

How We Launched a New Promotion Model: Matching and Regional Tests

Alina Savelieva25 May 2024

Alina Savelieva, Senior Monetization Analyst, Avito.

The speaker works on monetization analytics for Avito's transactional services (e.g., Avito Delivery and short-term rental) and develops causal inference approaches. The talk uses a real case to cover three things: how regional tests help overcome network effects, how to assemble a control group using matching when test assignment is not random, and how to validate such tests. The talk is useful for analysts, data scientists, and product managers who conduct A/B tests, encounter network effects, and cannot fairly split users into test and control groups. The speaker also notes that when comparing models, Avito tries to avoid regional tests (the previous speaker talked about this), but they are used in product launches.

Continue reading — it's free

Google or a one-time email code. No password.