How each candidate is scored

Two weighted signals

Each score blends two factors: value, weighted 60%, and recency, weighted 40%. Value rewards a higher total purchase history; recency rewards how recently the fan was last seen buying.

Value is measured against a $500 total-purchase cap; recency counts down across a 90-day window since the fan was last seen.

Estimate methodology

Each candidate shows an estimated range, not a single number. The conservative figure is the median of the purchase amounts behind that candidate; the optimistic figure is the mean of the same amounts. Whichever is lower becomes the low end of the range. When fewer than 3 related purchase events exist, the range collapses to a single-point estimate (the mean) and the row is marked as a small sample.

The thresholds we use

• High-value: a total purchase history of at least $100.
• Dormant: 30 days or more since the fan's last purchase.
• Recent paid-content window: a paid-content purchase inside the most recent 14 days.
• Expired lifecycle: a status marked expired, cancelled, churned, or inactive.

The three candidate lanes

High-value dormant

A fan with a total purchase history of $100 or more who has been quiet for 30 days or longer is a candidate: past value is high and recent activity has gone quiet.

Paid-content follow-up

A fan with a paid-content purchase inside the most recent 14 days and no refund or chargeback afterward is a candidate for a neutral follow-up note.

Expired reactivation

A fan whose lifecycle status reads expired (or cancelled, churned, inactive) with prior paid history is a candidate for a reactivation note.