Humble Choice cancellers bought more games on Steam before and after cancellation
Average paid Steam orders in 120-day windows around cancellation or a matched comparison date for retained subscribers
Average Steam orders per subscriber
Source: YipitData U.S. e-receipt panel. Cancellers n=437; matched retained subscribers n=380. Steam is not the full gaming wallet, and cancellation was not randomly assigned.
Chart data
| Series | Average Steam orders per subscriber |
|---|---|
| Cancellers before | 21.1 orders |
| Cancellers after | 20.2 orders |
| Retained before | 11.4 orders |
| Retained after | 13.5 orders |
Did Steam order frequency rise after Humble Choice cancellation?
No in this comparison. Cancellers averaged 0.9 fewer Steam orders after the event, while matched retained subscribers averaged 2.1 more after their assigned comparison date. The relative change was about three fewer Steam orders for cancellers. Because cancellation was not randomly assigned, this result describes divergent observed behavior rather than a causal effect of canceling.
How did the share buying on Steam change?
The active Steam-buyer rate among cancellers fell from 67.7% before cancellation to 64.3% afterward, a 3.4-point decrease. Among retained subscribers, it rose from 65.3% to 65.5%, a 0.2-point increase. The relative change was about −3.6 percentage points for cancellers compared with retained subscribers.
Does this measure every game purchase after cancellation?
No. The analysis covers paid Steam order receipts and excludes GOG, Epic Games Store, consoles, and other game purchases from the core metric. Dollar spending was also too sensitive to outliers to support the claim, so the article uses order frequency and buyer participation. The results do not establish why subscribers canceled or what happened to their total gaming wallet.
Assumptions & Methodologies
We analyzed our U.S. e-receipt panel from 2022 through 2026. Humble Choice cancellers had a first cancellation confirmation, an active subscription beforehand, complete 120-day pre/post windows, and continuous panel visibility. Retained subscribers had no observed cancellation, remained subscribed across both windows, and received comparison dates matched to the cancellation-date distribution. Steam orders were paid USD receipts; dollar totals were excluded from the conclusion. The cohorts are observational, Steam is not the full gaming wallet, and matched timing does not make cancellation a random treatment.