---
title: Integrated growth — moving user signals between store page, ads, paywall and CRM
slug: integrated-growth-signal-federation
summary: Andy Carvell's (Phiture) "signal-led" playbook — mine first-party data for intent, conversion-driver, engagement and value signals, validate each with an experiment, then reuse it on other surfaces (Adobe Japan custom product page +30% impression-to-download, Odyssey drive-time creatives, HBO Max recommendations in push, ASO ↔ UA creative swaps).
domain: mobile-apps
tags: [product-strategy, growth-systems, first-party-data, custom-product-pages, crm, personalization]
related: [paywall-personalization-from-onboarding-signals, custom-product-pages-for-ads, signal-engineering-and-conversion-event-audit, ad-landing-page-message-match, product-experiment-process-okrs-and-activation]
sources: [yt:Tid8-QMfpEE]
updated: 2026-10-24
---

# Integrated growth — moving signals between surfaces

## TL;DR
- Post-ATT, bought audiences and lookalikes are weak and ad networks are black boxes; **"creative is the targeting"**, so the edge is learning faster from **first-party data**.[^1]
- Data is abundant, **signal** is scarce: the part that says what one user wants and will pay for — and it usually stays trapped on the surface where it was collected.[^2]
- **Four signal types** to look for: intent, conversion driver, engagement, value (value signals go back to ad networks — signal engineering).[^3][^4]
- Loop: observe → hypothesis → **validate by experiment on one surface** → re-validate and adapt on each next surface → automate once proven.[^5][^6]
- Quick win to try this week: put your **best UA creative on the App Store page**, and your **best ASO test wins into UA creative**.[^7]

## Details
**Cases (Phiture clients, [[people/andy-carvell]]).**
- *Adobe Acrobat Reader, Japan*: analytics showed Japanese users mostly **edited** PDFs, unlike elsewhere. A PDF-editing custom product page, a main listing that stressed editing, and Apple Ads keywords about editing routed to it → **+30% impression-to-download**.[^8][^9]
- *Odyssey (US radio/music)*: listening peaked at commute "drive time"; surveys confirmed in-car use; driving-themed paid creatives and App Store description/creative performed well.[^10][^11]
- *HBO Max*: CRM pushes were generic "new this week"; the in-app ML recommender already had per-user picks. Engineering built an internal API so **Braze** could fetch each user's recommendation → large lift in push engagement and hours watched (no figure given).[^12][^13]
- *ASO ↔ UA*: Photoshop's ASO wins worked in paid creative; Onyx (hunting/outdoor maps) ran it the other way, UA winners lifting store conversion. Coherence between ad and store page adds a "halo" on paid spend; scale with a custom product page per winning creative.[^7][^14]

**Who does it.** A "growth intelligence layer" — at first just the growth team checking each surface for transferable signals; later rule-based automation and AI agents for user-level personalisation.[^15][^16]

**When a test fails** it is either the wrong signal or the wrong application of a good one; redesign before discarding.[^6]

**Hot take:** product-led growth (build 2–3 years and hope) has "expired"; siloed teams are "tired", integrated growth is "wired".[^17][^18]

## Disagreements / caveats
- Few numbers besides Adobe's +30%; most results are described as "huge" or "very meaningful".
- The paywall application is in [[paywall-personalization-from-onboarding-signals]]; matching ad and destination is the same idea as [[ad-landing-page-message-match]] and [[custom-product-pages-for-ads]].
- Feeding value signals to networks: see [[signal-engineering-and-conversion-event-audit]].

## See also
- [[product-experiment-process-okrs-and-activation]]

[^1]: yt:Tid8-QMfpEE @ 05:34 — Andy Carvell
[^2]: yt:Tid8-QMfpEE @ 00:00 — Andy Carvell
[^3]: yt:Tid8-QMfpEE @ 23:56 — Andy Carvell
[^4]: yt:Tid8-QMfpEE @ 24:50 — Andy Carvell
[^5]: yt:Tid8-QMfpEE @ 22:20 — Andy Carvell
[^6]: yt:Tid8-QMfpEE @ 27:16 — Andy Carvell
[^7]: yt:Tid8-QMfpEE @ 19:55 — Andy Carvell
[^8]: yt:Tid8-QMfpEE @ 02:07 — Andy Carvell
[^9]: yt:Tid8-QMfpEE @ 02:59 — Andy Carvell
[^10]: yt:Tid8-QMfpEE @ 10:41 — Andy Carvell
[^11]: yt:Tid8-QMfpEE @ 12:14 — Andy Carvell
[^12]: yt:Tid8-QMfpEE @ 13:38 — Andy Carvell
[^13]: yt:Tid8-QMfpEE @ 14:30 — Andy Carvell
[^14]: yt:Tid8-QMfpEE @ 20:42 — Andy Carvell
[^15]: yt:Tid8-QMfpEE @ 25:36 — Andy Carvell
[^16]: yt:Tid8-QMfpEE @ 26:23 — Andy Carvell
[^17]: yt:Tid8-QMfpEE @ 07:05 — Andy Carvell
[^18]: yt:Tid8-QMfpEE @ 08:16 — Andy Carvell