Weaken task: the commentator treats a higher average revenue per active creator as proof that the typical creator earned more. But an average can rise without anyone earning more if the makeup of the group being averaged changes. (C) supplies exactly that alternative: many low earners went inactive, so the mean of the smaller remaining pool rises mechanically while no remaining creator earned a cent more, so the typical creator need not have gained. That directly breaks the average-to-typical inference.
(A) only removes a measurement confound, which steadies the comparison rather than undercutting it. (B) names new payout options, a possible cause of higher pay, not a reason a higher average could coexist with flat typical earnings. (D) is about content variety, which is unrelated to how the average moved. (E) is the tempting one: a right-skewed distribution does let a mean sit above the median, but that skew is a standing feature present both months, so it cannot explain why the average rose this month. Only (C) supplies a month-over-month change in who is counted, which is what the conclusion requires.