What is click-to-install time?
Click-to-install time (CTIT) is the time between an ad click and the first open of the app that the click is credited with. It is one of the most useful signals in install fraud detection, because fraud tends to produce timing patterns that real users do not.
How CTIT is measured
- Start: the timestamp of the click recorded by the attribution platform.
- End: the first app open after install, when the SDK reports the install.
- Unit of analysis: a distribution per source, network or sub-publisher. The shape of the curve matters more than the average.
On Android, the Play Install Referrer also records when the store click and the install began, which helps tell a genuine click from one fired mid-install.
What CTIT tells you about fraud
- Too short: a cluster of installs opening within seconds of the click points to click injection.
- Too long and flat: a wide spread of long times with no clear peak points to click spamming.
- Different from your other sources: compare each network's curve against organic and your trusted channels. A source that looks nothing like the rest is worth a closer look.
Where CTIT gives false positives
- Re-engagement and referral links: users who already have the app open it in seconds, with no download.
- Incentivised installs: users paid to install often open the app immediately. They are low quality, but they are real people. See incentivised traffic.
- One global threshold: a single cut-off for all traffic will catch these real users. Set timing rules per channel.
CTIT versus click-to-install rate
Click-to-install rate is the share of clicks that become installs. CTIT is how long each install took after its click. A very low rate and a long, flat CTIT together are a strong click spamming signal.
How Linkrunner does this
Fraud Control includes a click injection template that flags Android installs opening within 3 seconds of their Play referrer click. New rules can be previewed on the last 7, 14 or 30 days and run as Tag only before any postbacks are withheld. For what normal looks like, see the click-to-install timing benchmarks, and for the wider method see how mobile install fraud is detected.