From 856b2badf7be5ae9c9f2feb03d5ed83bd2255158 Mon Sep 17 00:00:00 2001
From: Sean Oliver <882952+seanoliver@users.noreply.github.com>
Date: Thu, 14 May 2026 13:47:36 -0700
Subject: [PATCH] feat(telemetry): mirror signup_timestamp to PostHog person
 property (#45951)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

## I have read the
[CONTRIBUTING.md](https://github.com/supabase/supabase/blob/master/CONTRIBUTING.md)
file.

YES

## What kind of change does this PR introduce?

Feature — tiny telemetry addition. +7/-1 in one file.

## What is the current behavior?

[PR #45946](https://github.com/supabase/supabase/pull/45946) added
`org_count` as a PostHog person property to unblock targeting on org
membership. But `org_count` is set on every authenticated session (the
`Telemetry` component fires identify whenever `user.id` +
`organizations` resolve), not just at signup completion.

That means flag filters like `person.org_count == 1` match two
populations:
- Brand-new dashboard signups currently in their first org (intended)
- Returning users who happen to have one org and just signed in (not
intended)

For the upcoming `dataApiRevokeOnCreateDefault` experiment, this
contaminates the activation comparison because the "returning
single-org" group can't activate (they already did, months ago),
diluting the measured effect.

## What is the new behavior?

Adds `signup_timestamp: user.created_at` to the existing
`posthogClient.identify` call in `useTelemetryIdentify`. Since gotrue's
`user.created_at` is immutable, the value stays constant across sign-ins
— no need for `$set_once` semantics, no race with anonymous activity, no
cohort refresh lag.

Flag targeting can now combine `person.org_count == 1 AND
person.signup_timestamp >= <experiment_start_date>` to cleanly scope to
brand-new signups.

## Testing

No new unit tests added — the existing `useTelemetryIdentify` function
has no test file, the change is one additional field on an existing
call, and the property's correctness is verifiable end-to-end (sign up →
check PostHog person record). Adding to the test ticket
[GROWTH-854](https://linear.app/supabase/issue/GROWTH-854) for coverage
along with the broader posthog-client wrapper tests.

## Additional context

Ref: [GROWTH-853](https://linear.app/supabase/issue/GROWTH-853)

This is the follow-up gate before the 5% rollout of
`dataApiRevokeOnCreateDefault` — without this, the experiment would mix
brand-new signups with legacy single-org users on sign-in.

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

* **Chores**
* Enhanced telemetry and analytics data collection for signed-in users
by improving user identification tracking and adding signup timestamp
information for better analytics insights.

<!-- review_stack_entry_start -->

[![Review Change
Stack](https://storage.googleapis.com/coderabbit_public_assets/review-stack-in-coderabbit-ui.svg)](https://app.coderabbit.ai/change-stack/supabase/supabase/pull/45951)

<!-- review_stack_entry_end -->

<!-- end of auto-generated comment: release notes by coderabbit.ai -->
---
 packages/common/telemetry.tsx | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/packages/common/telemetry.tsx b/packages/common/telemetry.tsx
index 3dc49c6e669..b92aae494cc 100644
--- a/packages/common/telemetry.tsx
+++ b/packages/common/telemetry.tsx
@@ -446,7 +446,13 @@ export function useTelemetryIdentify(API_URL: string) {
         ...(anonymousId && { anonymous_id: anonymousId }),
       })
 
-      posthogClient.identify(user.id, { gotrue_id: user.id })
+      // user.created_at is gotrue's immutable signup timestamp — safe to $set on
+      // every identify because the value never changes per user. Lets flag
+      // targeting distinguish brand-new signups from returning single-org users.
+      posthogClient.identify(user.id, {
+        gotrue_id: user.id,
+        ...(user.created_at && { signup_timestamp: user.created_at }),
+      })
     }
   }, [API_URL, user?.id])
 }