App Store A/B Test 3D Simulator

Keyword & Ranking

Keyword relevance 72%

Competing apps 18

Creative A/B Test

Variant A conversion 3.2%

Variant B conversion 4.1%

Test budget (impressions) 1,200

Rank position

–

Impressions used

–

Installs A

–

Installs B

–

z-score

–

Significance

–

How it works

Every glowing particle is one simulated store visitor — a single Bernoulli trial, not an aggregate statistic. A particle spawns from the ranking gate (its spawn rate follows the same click-through curve real store search uses: better rank position → faster stream), is randomly assigned to creative variant A or B, and flips a weighted coin at that variant's conversion rate to decide install (green, joins the rising tower) or bounce (grey, fades away).

P(install | variant) = conversionRate(variant)
tower height ∝ cumulative installs
z = (p̂A − p̂B) / sqrt( p̂(1−p̂)(1/nA + 1/nB) )

The test runs against a finite impression budget, like a real capped A/B test campaign — once the budget is spent, spawning stops and the towers hold the final verdict. Because outcomes are drawn one particle at a time instead of computed as a day-level average, short runs show visible noise around each variant's true rate, exactly like a real test with a small sample.

Keyword relevance / competing apps — set the search-ranking position, which sets particle spawn rate.
Conversion rate A/B — the true per-visitor install probability for each creative variant (unknown to the test itself — only the sampled outcomes are).
Test budget — total visitors before the campaign ends; a bigger budget narrows the confidence interval and makes small true differences detectable.

Variant A visitor Variant B visitor Install (converted) Bounce

⚠ Couldn't load the 3D engineThree.js failed to load from the CDN. Check your connection and reload.