Most subscribers with no later OnStar charge still made observable fuel purchases
Share buying fuel in April–September 2025 and October 2025–March 2026
Share with an observed fuel purchase
Source: YipitData U.S. debit/credit card panel. No-later-charge group n=4,389; later-charge group n=44,241. Both groups had at least three OnStar charges in April–September 2025 and non-OnStar activity in both windows. Fuel purchases do not identify a specific vehicle.
Chart data
| Series | Share with an observed fuel purchase |
|---|---|
| No later OnStar charge — before | 84.3% |
| No later OnStar charge — after | 81.7% |
| Later OnStar charges — before | 76.6% |
| Later OnStar charges — after | 75% |
How did external spending change after OnStar payments stopped?
Mean non-OnStar card spending among subscribers with no later OnStar charge fell 3.3%, from $12,291 in April–September 2025 to $11,890 in October 2025–March 2026. It rose 2.8%, from $11,746 to $12,076, among subscribers with later charges. Median changes were −2.4% and +0.8%. Requiring non-OnStar card activity in both periods excludes people who stopped appearing in the panel.
Do fuel purchases prove that customers kept the same vehicle?
No. Fuel spending indicates continued card-visible vehicle activity but cannot identify the vehicle, driver, or whether the customer still owned a General Motors vehicle. Stopped OnStar charges can also reflect a payment-method change or annual billing rather than cancellation. The result shows that many observable subscribers without a later charge remained fuel buyers with broadly intact external spending; it does not establish why OnStar payments stopped or whether a win-back offer would work.
Assumptions & Methodologies
We analyzed YipitData’s U.S. debit/credit card panel. Established subscribers had at least three OnStar charges in April–September 2025. One group had none in October 2025–March 2026; the comparison group had at least three. Both needed non-OnStar activity in both windows. External spending excluded OnStar, and fuel came from service-station and automated-fuel-dispenser categories. The same calendar windows reduce seasonal bias. Card data cannot identify a vehicle, plan, cancellation reason, ACH payment, or card change, and the activity requirement excludes households that left the panel. Results show association, not cause.