The CMO Podcast with Fexingo · 2026-07-06 · 10 min
Episode 97 of The CMO Podcast explores how senior marketers are deploying machine learning models to estimate customer lifetime value before a customer makes their first repeat purchase. Lucas and Luna break down the shift from RFM-based segmentation to neural-network predictions, using the case of a mid-market DTC brand that cut acquisition costs by 22 percent within six months by scoring prospect lists against predicted LTV. They discuss the data requirements - minimum 5,000 historical customer journeys, preferably 12 months of transaction data - and the organizational friction: finance teams often distrust CLV models because they forecast revenue that hasn't happened yet. The episode also covers how brands are feeding granular behavioral signals - scroll depth, time on site, email open patterns - into models that generate segment-specific marketing budgets. No software-name-dropping, just the architecture of how modern CMOs are shifting from CAC obsession to a blended unit-economics view that includes predicted retention curves. A concrete playbook for marketers who want to stop chasing high-intent but low-LTV customers.
Other episodes covering the same guests and topics, from across The B2B Podcast Index.