Introduction to Deferred Linking
Detour enables deferred deep linking for mobile apps, so you can preserve navigation context and campaign data across the install flow.
Detour Dashboard
Deferred deep linking lets you send users to specific in-app content even when they must first install the app. A deferred link records a click (and lightweight contextual signals) and then, after install and first app open, attempts to deliver the original destination into the app.
The core flow:
- A user clicks a Detour link (for example:
https://your-org.godetour.link/<app_hash>/campaign-x). - Detour records the click and redirects the user to the store.
- After install and first open, your SDK sends a match request (
clickIdwhen available, otherwise probabilistic fingerprint data). - Detour matches the install with the click and returns the matched link URL.
Why use deferred links?
- You preserve marketing context through the store-install friction, increasing the chance users see the exact promotion or content you intended.
- You get attribution for installs via link-based campaigns.
- You can improve onboarding by pre-filling flows or showing contextual content on first open.
Deterministic vs probabilistic matching
Detour uses two matching modes:
-
Deterministic matching uses an explicit identifier generated during click-time that is preserved through the store-install flow and read on first app open (e.g., Android Install Referrer). Deterministic matches are exact and preferred when available. Detour uses
click_idfrom the Play Install Referrer on Android for deterministic matching. -
Probabilistic matching (fallback) uses multiple signals collected at click time and again on first app open, then computes a similarity score. Signals can include user agent, screen size, system version, locale, timezone, IP and optionally the clipboard contents. If the best candidate's score exceeds a conservative threshold, we return the matched link.
Limitations and important notes
- Timing: probabilistic matching is time-sensitive and limited by a configured time window.
- Cross-device: if a user clicks on device A and installs on device B, matching will not work.
- Privacy / platform constraints: iOS does not provide an install referrer equivalent. Clipboard-based heuristics can help but require the user's paste permission on iOS and are not guaranteed. Always design your campaigns with these constraints in mind.
- Accuracy: probabilistic matching is best-effort and cannot guarantee 100% accuracy. Avoid using deferred links for sensitive data routing.