Postmates lapsers newly ordered through Uber Eats and DoorDash at similar rates
Share of eligible lapsers first appearing on each platform in July 2025–June 2026
Share newly ordering through the destination
Source: YipitData U.S. e-receipt panel. Postmates lapser cohort n=1,007. Eligible denominators exclude people who already used the destination in July 2024–June 2025: Uber Eats n=409; DoorDash n=507. New-customer counts were 80 and 99, respectively.
Chart data
| Series | Share newly ordering through the destination |
|---|---|
| Uber Eats | 19.6% |
| DoorDash | 19.5% |
How many eligible lapsers did each platform newly receive?
Uber Eats newly received 80 of 409 eligible Postmates lapsers, while DoorDash newly received 99 of 507. The denominators differ because anyone who had already ordered from a destination during the pre-period was excluded from that destination’s new-adoption rate. The comparison therefore measures first observed adoption after the Postmates lapse, not every lapser’s total delivery-platform use.
Did the newly captured customers order more through one destination?
Their observed order frequency was similar. Customers newly ordering through Uber Eats averaged 8.2 orders during the post-year, compared with 8.0 for those newly ordering through DoorDash. That repeated activity shows more than a single observed order on average, but it does not establish exclusive switching, the reason for adoption, or whether a customer used untracked restaurant-delivery channels.
Assumptions & Methodologies
We analyzed YipitData’s U.S. e-receipt panel. Established customers placed at least two Postmates orders in July 2024–June 2025 and remained active in the panel. Lapsers placed no Postmates order in July 2025–June 2026. For each destination, new adoption required no destination order in the pre-year and at least one in the post-year. This produced different eligible denominators for Uber Eats and DoorDash. Order receipts do not measure spending, exclusive platform use, or intent, and untracked delivery channels remain possible. The counts of 80 and 99 are panel estimates, not market totals.