Pump.fun's $15 million callout program switches to paying for followers' profits

Alon framed the goal as broader "wealth effects", instead of letting most profits flow to a small group of skilled traders, according to Crypto Briefing's account of his post on X ^[3][5]. In market-structure terms, the caller's pay is moving away from a commission on flow, earned when the trade happens, and toward a performance fee, earned only if the trade works. The money still comes from the budget Pump.fun set aside when the program launched around August 2026 ^[6].

We think the change makes bad calls pay less. Whether they now pay nothing depends on how Pump.fun defines a successful callout, and it has not published that definition ^[11]. A follower's profit depends on when they buy and when they sell, and followers do not all act at the same moment ^[14]. Take a token that jumps in its first hour and falls by evening. Early buyers end up ahead and late ones behind, and the caller's score depends on which group the formula counts. Low-cap calls stay in the program at a reduced rate ^[1]. Pump.fun cut what they pay instead of dropping them.

If the profit test holds, income moves away from callers who post often on thin tokens and toward selective ones, as the change intends ^[2]. If callers work out what time window the test measures, they can time calls to look good inside it, though a private formula is harder to reverse-engineer ^[15]. The third possibility is another rule change. Pump.fun adjusted scoring on October 4 and again on October 10, six days apart ^[16]. On that record alone, we'd give this outcome the most weight in the near term. The counter-case is that the two fixes, one aimed at spam and one at volume, are a program about two months old settling into its design ^[17]. On that view, a third change comes only if the profit test fails in public.

The view that bad calls now pay less is wrong if accounts posting many low-cap calls still top the earnings rankings after October 10, or if followers of top-ranked callers keep losing money. Checking either is harder than it sounds. Historical leaderboard data does not reflect the latest algorithm updates ^[10], so the rankings from before and after the change were produced under different rules.