Know Your Customer
summary
Ticketmaster implements RFM and probabilistic models (Bayesian, Beta-Geometric/Beta-Binomial, gamma for spend) to predict two-year LTV at the individual level and drive marketing choices. A recent A/B-like change — sending a second email — yields a 20% conversion lift but also 20% unsubscribe rate; modeling projects $3M incremental revenue, $1M added CLV in year one and a net impact ≈ $2M/year, so the feature launches. Models run in containers for reproducible deployment and target CLV:CAC ~3–4:1.
AI-written summary · about
chapters (6)
- 0:00Start
- 4:20What Is My Projected Clv
- 11:37Purchasing Journey
- 32:05Conditional Expectations by Frequency and by Recency
- 35:09Define Your Matrix
- 41:37Allocation of the Marketing Resources
related talks
36mCustomer Lifetime Value: Models, Metrics and a Multitude of Uses
17mdata science at Ticketmaster
40mNext Best Action: Maximizing Customer Value
24mUnlocking revenue: newcomer classification for boosted profits
27mPreventing churn like a bandit
44mEating Humble Py: From Toy Problem to Real-World Solution in Predicting CLV
28mRecent approaches in retail banking marketing optimization
36mML: Predictive analytics in Ad Engineering