Free tool
Creative testing budget planner
A creative test for app installs is a structured batch — 3–5 concepts × 2–3 hooks each — with a spend ceiling you can compute before launch. Pick your concepts, hooks, and target CPI below and the planner builds your test matrix, sizes the budget, and lays out the week-one kill/scale plan. Every threshold is a practitioner heuristic from our 3×3 Ladder framework — a starting point to tune, not a platform rule.
Distinct angles, personas, or problems. The framework batches 3–5.
The first 1–3 seconds. The framework uses 2–3 per concept.
Your guess — no target yet? See the CPI calculator.
Practitioner heuristic — the framework calls for ~20+ installs (a double-digit count) before CPI means anything.
full-read test budget: 9 variants × 20 installs × $3 CPI
Capped-ladder worst case instead: $84 — 9 variants × 2× CPI on rung 1 + 2 survivors × 5× CPI on rung 2 (upper caps of the 1–2× and 3–5× heuristics; the framework's worked example lands at ~$85 for nine variants at a $3 CPI).
Your test matrix
Concepts × hooks — structure tells you why something won: three failing hooks kill a concept; one hook winning across two concepts reveals a reusable pattern.
3 concepts × 3 hooks = 9 ad variations (the framework's sweet spot is 6–15 per batch)
The week-one plan
Every threshold below is a practitioner heuristic — see the framework — tune to your account.
- 1
Day 0 — launch
Launch all 9 variants simultaneously with equal rung 1 caps of 1–2× target CPI ($3–$6 each) and automated stop rules. Caps are automatic, not aspirational. Write down kill/scale thresholds before launch.
- 2
Day 2–3 — fast kill (rung 1)
Kill the bottom half of the batch on hook rate and IPM at the rung 1 cap; read Android / top-of-funnel first (iOS SKAN postbacks arrive days late). Pause and log which hook or concept failed.
- 3
Day 4–7 — validate (rung 2)
Give the top 2–3 survivors a larger cap of 3–5× target CPI ($9–$15), or ~20+ installs. Standard kill: CPI still >1.5–2× target at the rung 2 cap — keep the learning, not the ad.
- 4
Scale (rung 3)
Scale when CPI is at or under target across ~20+ installs and the ad absorbs extra budget without CPI collapsing. Promote to the scaling campaign; feed the next batch with new hooks on this week’s winning bodies plus one brand-new concept.
Frequently asked questions
How many ad variations should one test batch have?
Six to fifteen net-new variants — 3–5 concepts × 2–3 hooks — is the practical sweet spot: enough for the win-rate math to work (roughly 5% of creatives become true winners per Motion's Creative Benchmarks 2026), few enough that each variant gets readable spend. Structure the batch as concepts × hooks rather than ten random videos, so every result tells you why something won or lost.
How much budget does creative testing need?
Plan for testing to consume roughly 10–20% of your total UA budget (a heuristic, not a law), and size each batch as variants × per-creative kill cap: 1–2× target CPI on the signal-test rung, 3–5× target CPI or ~20+ installs on the validation rung. The framework's worked example: at a $3 target CPI, a nine-variant batch costs at most about $85 of media for a full weekly cycle — the ceiling is computable before you spend a cent.
When do you kill or scale a creative?
Kill on data, not taste, with thresholds written down before launch. Fast kill: bottom half of the batch on hook rate and IPM at the rung 1 cap. Standard kill: CPI still 1.5–2× above target at the rung 2 cap. Scale: CPI at or under target across ~20+ installs, and the ad absorbs extra budget without CPI collapsing. All heuristics — the full rules are in the creative testing framework.
Full disclosure: AppVids is our product — a 10-video AI UGC pack (€249, 48-hour delivery) covers one 3×3 batch plus a spare, which is exactly the production bottleneck this planner exposes; see our pricing.