Skip to main content
Growth Opt Playbook
Growth Opt
Playbook
Library
◎Home
Learn & explore
▤Blog▧Guides & SOPs◇Start with a question
Analyze & review
⇧Start with my data▦Browse analyses
My workspace
◉My account◷My projects▣Data on this device▤Plans & service
Library
Marketing metric calculatorsTemplatesGlossaryMethod comparisons
Uploaded data is processed only in this browser.
Growth Opt PlaybookGrowth Opt Playbook/Guide 2 · Campaign Execution/Guide 2-1 · Google App Campaigns (UAC) Guide
About Pro◷My projects
My account

What can you actually control in Google UAC?

Campaign structure, bid-stage progression, and asset group diversity.

Loading…

Related resources

Related guides and terms

Check this with your data
  • Budget Allocation for CPI, CPA & ROAS→
Guide
  • When Google UAC Slips: Diagnose Learning and Bids→
  • Ad Learning Phase: Why CPA Swings While the Model Learns→
Key glossary terms
  • CPI (Cost Per Install)→
  • ROAS (Return On Ad Spend)→
⌕
↑↓ move↵ openesc close
Growth Opt Playbook

A browser-only weekly decision workspace for performance marketers.

AnalyzeOperations dashboardPerformance varianceCreative fatigueBudget allocationMarketing metric calculatorsDiagnose performance
LearnMy projectsWeekly reportTemplatesMethod comparisonsMethodology manualsBlogGlossaryOperating guides
LegalPrivacyTermsContact
공드리공작소
Representative
신기훈
Business registration number
856-07-03210
Business address
서울특별시 동대문구 천호대로41길 26, 402호(용두동, 천우네오젠)
Customer service
010-8829-9034
Email
gondry.montauk@gmail.com
Growth Opt Playbook
Growth Opt
Playbook
Library
◎Home
Learn & explore
▤Blog▧Guides & SOPs◇Start with a question
Analyze & review
⇧Start with my data▦Browse analyses
My workspace
◉My account◷My projects▣Data on this device▤Plans & service
Library
Marketing metric calculatorsTemplatesGlossaryMethod comparisons
Uploaded data is processed only in this browser.
2 Campaign · 2-1

Google App Campaigns (UAC) Guide

UAC is a black-box campaign where machine learning decides bidding, targeting, and creative combinations. The operator's three levers are (1) campaign structure, (2) bid-stage progression, (3) asset group diversity.

Channel · Google AdsBidding · tCPI/tCPA/tROASLearning phase ≥ 7 days
Summary

UAC splits into Install campaigns (I) and In-app Action campaigns (II/III). New apps start with I, move to II (tCPA) after securing 30+ installs/day, and graduate to III (tROAS) once revenue signal stabilizes. Safe bid changes: at most once a week, within ±20%.

§1 Campaign structure (OS × country × stage)

Create app campaigns for a selected platform and app. Run iOS and Android apps in separate campaigns. Countries need not always be separate: locations with similar goals, languages and conversion values can share targeting. Consider separate campaigns for country-level budget control or materially different performance, while checking the loss of learning data from excessive segmentation.

Campaign type↕Bid option↕Conversion definition↕Min daily conversions↕Recommended timing↕
UAC for Install (UAC-I)tCPIInstall10Launch ~ Day 30
UAC for In-app Action (UAC-II)tCPATier-1 event (sign_up etc.)10After Day 30
UAC for ROAS (UAC-III)tROASRevenue30After monthly revenue stabilizes

§2 Bid strategy — staged progression

  1. Phase 1 · tCPI (Day 1–30) — set the initial target CPI at 1.3–1.5× market average (to accelerate learning). After reaching 30+ installs/day, step it down by −10% at a time.
  2. Phase 2 · tCPA (Day 30–60) — designate a Tier-1 in-app event (sign_up, purchase, etc.) as the conversion. Initialize CPA = CPI × (inverse of install→event CVR).
  3. Phase 3 · tROAS (Day 60+) — enter after D7 or D30 revenue stabilizes. Start the ROAS target around LTV/CAC = 1.0 and raise +10% weekly.
!
Learning Phase

After any bid change, the campaign is learning for at least 7 days + 50 conversions. Changing again during that window restarts the ML and craters cost efficiency. Change at most ±20% at a time.

§3 Asset group rules (creative diversity)

UAC auto-combines assets into ads. Low diversity = weak learning signal. Run 2–3 asset groups per campaign, each enforcing these minimums:

Asset type↕Minimum↕Recommended↕Spec↕
Headline (Text)510≤30 chars; networks auto-truncate
Description (Text)510≤90 chars
Image381200×628 + 1200×1200 + 1200×1500
Landscape video1316:9, 15–30s, MP4
Portrait video139:16, 15–30s (YouTube Shorts placement)
Square video121:1, 15–30s
HTML5 (optional)01300×250 or 320×480 interstitial

§4 Troubleshooting

!
The campaign never exits the learning phase

With low conversion volume, first check tracking gaps, conversion delays, budget and target bids. Consider an upstream event that still relates to the business goal. Countries with similar goals, languages and conversion values may be consolidated, but iOS and Android apps remain separate campaigns. Do not treat a fixed conversion count or bid increase as a guarantee of learning success.

!
A tROAS campaign reports 0 revenue

Check the revenue-sending toggle for the purchase event in Adjust → Google Ads partner. Without a Firebase connection revenue can't flow — switch to Adjust direct-send mode.

!
iOS campaign installs report only via SKAN

This is normal. All iOS 14.5+ users who declined ATT are classified into the SKAN channel; total installs = deterministic + probabilistic + SKAN summed in the Adjust dashboard.

가이드 2-1. Google App Campaigns (UAC) Guide Last updated · 2026-05-22
Contents
Campaign structureBid strategy stagesAsset group rulesTroubleshooting
Sources & reviewReviewed · 2026-10-08

Growth Opt Playbook editorial review

  • Set up Google Ads App campaigns for installs ↗
  • Google Ads App campaigns help ↗
  • Google Ads Help Center ↗