Names and aliases
TTFT, first-transaction latency
Formula
first_successful_transaction_at - eligibility_started_at; report median, quantiles, and the no-event share
Metric card
| Unit | time: minutes, hours, or days |
|---|---|
| Data source | Eligibility-start event, first successful transaction, time zone, and observation-window closure. |
| Frequency | By entry cohort after the maximum observation window closes. |
| Segments | market side, micro-market, category, channel, participant type, cohort |
Interpretation errors
- Starting the clock at registration for one cohort and listing publication for another.
- Dropping participants without a transaction and ignoring right censoring.
- Comparing immature cohorts with fully observed cohorts.
Example
Illustrative example: observed first transactions have a 6-day median while 20% of the cohort has no event by window close; both results must be shown.
Use
Define the market side, start event, terminal success, and maximum window before analysis. For incomplete cohorts, use a survival view or report the censored share explicitly.
Anti-example
The median among transacting participants may improve simply because slower participants stop reaching the outcome and disappear from the calculation.
Limitations
- The metric depends on the start point, terminal event, and censoring treatment.
- A faster first event does not establish repeat behaviour, quality, or positive economics.
Related metrics
seller-activation-rate, time-to-match, activation-retention, transaction-completion-rate
Context sources
These works provide two-sided-market context; the formula above remains an author-defined working definition.
- Kong-Pin Chen, Yen-Chi Huang. A Search-Matching Model of the Buyer-Seller Platforms (2012).
- Jun Li, Serguei Netessine. Higher Market Thickness Reduces Matching Rate in Online Platforms: Evidence from a Quasiexperiment (2019).