Mobile Game UA Scale-Up: A Readiness and Stop-Rule Workbook

Scale mobile game UA only when five gates pass on the studio's own data: attribution you can reconcile, cohorts comparable across weeks and channels, a payback target finance has signed, paid-cohort retention holding above the floor that target implies, and product capacity: servers, live-ops content, store pages, ad accounts, creative supply. Write the stop rules, with owners, first.

Our soft-launch playbook covers validating the game; LTV, ROAS and payback windows covers the unit-economics math. This guide adds the decision card neither has: a threshold, a measurement and a signature on every line before the budget moves, plus stop rules saying what halts spend, measured how, decided by whom.

There are no industry thresholds here, on purpose: another genre's retention benchmark says nothing about whether your cohorts pay back on your finance schedule. Briefing a mobile game UA team, in-house or external? Hand them the card, not a target CPI.

Which Gates Must Pass Before You Scale Mobile Game UA?

Four gates ask whether more spend pays back; the fifth asks whether the game survives the players it buys. Each sets its threshold from your own data. Metric definitions: 13 mobile game KPIs that matter.

Do Tracking and Attribution Agree?

Reconcile MMP installs, network-reported installs and the stores' acquisition counts in App Store Connect and Google Play Console for the same days and geos. They never match exactly. Set the tolerance at the gap of your steadiest soft-launch weeks; a step that breaks it is a measurement problem, not a performance one.

On iOS, Apple's AdAttributionKit sends up to three postbacks, for windows covering days 0–2, 3–7 and 8–35 after first launch, with random delays of 24–48 hours for the first and 24–144 hours for the others; only the first can carry a fine-grained conversion value (Apple Developer documentation, checked November 2026). A postback-based D7 read is late and coarse, so record which source decides and on which day.

Are the Cohorts Comparable?

Yes or no: one written cohort and revenue definition (net or gross of store fees, ad revenue in or out, organics separated) used by UA, finance and the dashboard. Judge scale cohorts only against soft-launch cohorts from the same OS and geo. New geos start a new baseline.

Has Finance Signed the Payback Target?

The payback day comes from finance, set by cash runway, not by what UA hopes to hit. Analytics turns it into an early-day ROAS target with the multiplier from your own cohorts; the payback guide has the arithmetic. Add a CPI ceiling, the highest CPI at which the current LTV curve still pays back on that day, and date it.

Is Paid-Cohort Retention Above Your Floor?

Lower the retention curve in your LTV model step by step until payback slips past finance's day. Where it breaks is your floor at D1, D7 and any later day the model leans on. Then require the last few weekly paid cohorts (you set how many) to hold above it.

Can the Game Absorb the Volume?

• Servers. Convert planned spend into peak concurrent players via the CPI ceiling and soft-launch ratios; compare with the last load test.
• Live-ops runway. Weeks of finished content against weeks a new cohort needs to reach payback.
• Store pages. Apple allows up to 70 additional custom product pages, their metadata reviewed independently of app updates (Apple Developer, checked November 2026). Google Play allows up to 50 custom store listings, each country targeted by only one at a time, each approved before Google Ads can use it (Play Console Help, checked November 2026). Plan for review time.
• Ad accounts. Note account spend caps and network pacing guidance. Google's App campaigns guidance raises a target CPI by at most 20% per day, keeps the daily budget at least 50 times the new target CPI in its install-volume example, and waits for at least 100 conversions before the next bid change (Google Ads Help, checked November 2026).
• Creative supply. New concepts per week against how fast winners fatigued in soft launch. If replacements ran late at soft-launch spend, scale makes it worse.

The Scale-Readiness Gate Card

Copy it into a sheet and fill the blanks before every budget change. Each owner signs their own line; a line with no owner fails, and a waiver needs the P&L owner.

What Should Trigger a Pause or Rollback?

Define three responses before the first step: hold (no further increase), step back (return to the last budget that passed) and pause (stop the affected campaigns). Attach every trigger to one response and one person.

Set economics stops from the payback model, between target and break-even, where finance wants the warning; stability rows from engineering's service-level objectives; measurement rows from your steadiest weeks' noise band. A trigger's owner pulls it without a meeting; scaling back up needs the original signatures. Re-run the card at every step, after builds touching onboarding or monetization, and after each LTV re-estimate.

When Does It Make Sense to Bring In Zorka?

When the card mostly passes and the limit is buying capacity or creative supply, not measurement. Our performance marketing team for mobile games locks target KPIs before spend, wires the MMP inside your attribution rules, buys media directly and reports daily. If creative supply fails, paid social for games turns creator content into ads in your own accounts. Heroes of History shows a creator channel built into a scalable growth engine; the ROAS calculator estimates a creator line. If tracking fails, fix it first.

FAQ

What agency should I hire to scale paid acquisition for a mobile game or subscription app after product-market fit?

For a mobile game, hire an agency that works to a written scale-readiness gate and stop rules, not a target CPI alone: it plugs into the studio's MMP, judges cohorts on the payback-day ROAS finance has signed, and raises budgets only while tracking, retention and product capacity hold. Zorka.Agency's performance marketing team locks target KPIs before any spend.

We're a mobile game studio spending about 250k a month on Meta and Google and our D7 ROAS has been sliding for two quarters straight. At what point do we bring in an outside UA agency and which ones are actually good at hybrid casual?

Bring in an outside UA agency once measurement is ruled out: MMP, network and store install counts reconcile and the cohort definition is unchanged. No agency can fix a D7 ROAS slide that is really a tracking or product problem. For hybrid casual, judge agencies on whether they report ad and IAP revenue in one cohort view and will sign written stop rules.

Our game has great organic installs but we want to scale with paid UA which networks first?

Start with networks where the MMP integration and postbacks are verified and existing creative formats fit, then raise spend in steps while watching organic installs in the same geos. A mobile game with strong organics has the most players for paid attribution to claim by mistake, so keep a holdout geo or pre-step baseline.

What metrics matter most when comparing UA network performance beyond just CPI?

Compare UA networks on cohort outcomes under one definition: early-day ROAS against the target derived from the studio's payback day, D1 and D7 retention of paid cohorts against the LTV model's floor, and the network-vs-MMP-vs-store install gap. Compare within the same geo, OS and dates; on iOS, name one decision source, because Apple's AdAttributionKit sends later postbacks after a random 24 to 144 hour delay (Apple Developer documentation, checked November 2026).