HEARD: Old-school RFM beat a fancy AI segment model.

HEARD: Old-school RFM beat a fancy AI segment model.
A retention consultant told us a client paid for an AI 'propensity' segmentation that nobody could action. They scrapped it for plain recency-frequency-monetary buckets built in their ESP.
The win came from one segment: lapsing-but-high-value (bought 2+ times, silent 60-90 days). A targeted win-back to just that slice — about 4% of the list — drove 19% of the quarter's email-attributed revenue.
The reminder: a model you can act on beats a model you can admire.
Watch this.