---
title: How much Meta budget does an app need? (the 50-events math)
slug: meta-ads-budget-math-for-apps
summary: Worked math for what a Meta ads budget buys an app — $1,000 at $2 CPI → 500 installs → 35 trials → ~10 subscribers (~$95 CAC) — and why trial optimization needs 50 events/week.
domain: mobile-apps
tags: [user-acquisition, meta-ads, budget, unit-economics]
related: [meta-ads-readiness-checklist, app-break-even-cpi-and-revenue-per-install, paid-social-creative-testing-framework, meta-ads-per-ad-set-and-budget-minimums, signal-engineering-and-conversion-event-audit]
sources: [yt:2jH64ziB7uM, yt:WaOrPQVLz08]
updated: 2026-10-23
---

# How much Meta budget does an app need?

## TL;DR
- Optimize for **start trial**, not installs — Meta gets you exactly what you ask for, and cheap installs ($1–2) often never pay.[^1][^2]
- Meta needs **≥50 optimization events per week**; size the weekly budget to hit 50 trials.[^3]
- $1,000 over 1–2 weeks at $2 CPI ≈ 500 installs → 35 trials (7%) → ~10–11 subs (30%) → **~$95 per subscriber**.[^4]
- So: weekly budget ≈ 50 ÷ install→trial rate × CPI (e.g. 50 ÷ 0.07 × $2 ≈ $1,430/week).
- Expect to lose money for a while; one founder took 2–3 months to reach positive ROAS.[^5]

## Details
**Steve's worked example (Oct 2026):**[^4]

| Step | Value |
|---|---|
| Budget | $1,000 (≈1–2 weeks) |
| CPI | $2 → 500 installs |
| Install→trial 7% | 35 trials (below 50/week) |
| Trial→paid 30% | ~10–11 subscribers |
| Cost per subscriber | ~$95 |

A $1,000 test can tell you whether a creative gets good installs, but it won't exit learning on trial optimization — you'll need a bigger weekly budget.[^6] That's why funnel conversion and LTV per install must be high before Meta (see [[meta-ads-readiness-checklist]]).

**Optimization event ladder.** Install campaigns are cheap but low quality; the goal is to move "as down-funnel as possible" — start trial is where Steve wants to be.[^1][^7]

**Patience.** Steve cites Ivy (founder of an app he calls "Resto Me"), who spent ~$20K/day at positive ROAS but only after 2–3 months of negative results.[^5]

## Disagreements / caveats
- The 50 events/week figure is Steve's statement of Meta's requirement as of Oct 2026; check current Meta guidance.
- Part 2 of his Meta series (campaign setup/budgeting) wasn't published at ingest — gap.[^8]

## See also
- [[paid-social-creative-testing-framework]] — test-phase budgets
- [[app-break-even-cpi-and-revenue-per-install]]

[^1]: yt:2jH64ziB7uM @ 07:35 — Steve P. Young
[^2]: yt:2jH64ziB7uM @ 08:30 — Steve P. Young
[^3]: yt:2jH64ziB7uM @ 04:02 — Steve P. Young
[^4]: yt:2jH64ziB7uM @ 17:00 — Steve P. Young
[^5]: yt:WaOrPQVLz08 @ 02:37 — Steve P. Young
[^6]: yt:2jH64ziB7uM @ 18:33 — Steve P. Young
[^7]: yt:WaOrPQVLz08 @ 01:35 — Steve P. Young
[^8]: yt:2jH64ziB7uM @ 20:17 — Steve P. Young