Skip to content

CONCEPT Cited by 1 source

Ride cancellation rate

Definition

Ride cancellation rate is the share of dispatched rides that are cancelled after driver match but before trip completion. In a two-sided marketplace it is conventionally broken out by party:

  • Rider cancellation rate (post-match) — rider bails after the driver is already assigned and en route.
  • Driver cancellation rate (post-match) — driver abandons the assignment after accepting it.

Both sides produce marketplace-level waste — the ride that doesn't happen consumes platform resources and hurts the counterparty's experience.

Why Lyft tracks it post-match specifically

Pre-match cancellation (rider cancels before a driver is assigned) is cheap — a request that never matched. Post-match cancellation is the expensive one:

  • The driver has already committed and accepted.
  • The driver may have started driving to pickup.
  • The rider has already been informed a driver is coming.
  • Re-dispatching produces a double-match pattern that harms driver efficiency.

The gated-community cancellation failure

Lyft's 2026-04-23 write-up uses cancellation rate as the primary business metric for the gated-community pickup fix. Both sides cancel more at gates:

  • Rider cancels because the wait ticks up, chat starts, they can't get the driver in, coffee goes cold → "forget this".
  • Driver cancels because they've circled the community trying to find a way in, the rider isn't responding with a working code, they've spent five minutes earning nothing.

The post reports measurable decreases in both post-match cancellation rates for gated community rides after the new pickup flow rolled out, without disclosing specific deltas.

Why post-match cancellation is a good metric for pickup UX

experiments

Cancellation rate captures the real-world failure mode the feature is meant to eliminate. Other candidate metrics have blindspots:

  • Rider survey satisfaction — subject to selection bias and non-response. Lyft did capture this too (~95% positive).
  • Pin-to-actual-pickup walking distance — measures a process change, not an outcome.
  • Wait time p50/p90 — contaminated by matching supply, not just pickup UX.

Post-match cancellation rate is an unambiguous outcome: the ride either happened or it didn't.

Seen in

Last updated · 550 distilled / 1,221 read