Add Peer Benchmarks section to Track Application Startups

hugo/content/en/real_user_monitoring/setup/enable_rum/track_application_startups.mdoc.md

@@ -20,12 +20,14 @@ RUM measures how fast your mobile app becomes usable after launch. Time to initi
 
 {% if equals($platform, "android") %}
 {% partial file="sdk/track_app_startups/android.mdoc.md" /%}
+{% partial file="sdk/track_app_startups/peer_benchmarks.mdoc.md" /%}
 {% /if %}
 
 <!-- iOS -->
 
 {% if equals($platform, "ios") %}
 {% partial file="sdk/track_app_startups/ios.mdoc.md" /%}
+{% partial file="sdk/track_app_startups/peer_benchmarks.mdoc.md" /%}
 {% /if %}
 
 <!-- Flutter -->

hugo/layouts/shortcodes/mdoc/en/sdk/track_app_startups/peer_benchmarks.mdoc.md

@@ -0,0 +1,28 @@
+## Peer Benchmarks
+
+Peer Benchmarks compare your application's TTID to the performance of comparable applications monitored by Datadog. Beyond your isolated value, you get visibility into your relative performance against other organizations using Datadog.
+
+### Use Peer Benchmarks
+
+{% alert level="info" %}
+To use this feature, your organization must use [RUM without Limits](/real_user_monitoring/rum_without_limits/).
+{% /alert %}
+
+Peer Benchmarks are displayed on the Explore Vitals page.
+
+Benchmarks adapt to the filters you apply on metric dimensions in the UI, letting you compare your performance for a specific end-user base.
+
+Filtering on `env`, `service`, or `version` has no effect, because these dimensions are not shared across organizations.
+
+### How it works
+
+The peer applications you are compared with are selected so that they share a similar audience base:
+
+- **Development platform**: peer applications are built on compatible development platforms.
+- **User traffic**: peer applications share a similar volume of ingested sessions.
+
+### Data privacy
+
+Peer Benchmarks data is fully anonymized: only aggregated indicators are displayed, so no individual application can be isolated.
+
+Benchmarks are only shown when the number of peer applications is large enough to ensure statistical anonymity and accuracy.