Run Play Store native A/B listing experiments (titles, screenshots, icons, descriptions). Use when the user says 'listing experiment play', 'a/b test play listing', 'play store experiments'.
Installs into .claude/skills of the current project.
Are you the author of Play Listing Experiments?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/khadinakbarlabs-play-listing-experiments)
---
name: "play-listing-experiments"
description: "Run Play Store native A/B listing experiments (titles, screenshots, icons, descriptions). Use when the user says 'listing experiment play', 'a/b test play listing', 'play store experiments'."
---
# Play Listing Experiments
Native A/B testing in Play Console. Better than guessing.
## What you can test
- Default graphics (icon, screenshots, video) — global
- Localized graphics — per language
- Localized text (title, short desc, long desc) — per language
- Default text — global
## Setup
Play Console → Grow → Store listing experiments → Create experiment.
1. Choose what to test (e.g., icon variant)
2. Upload 2-4 variants (A vs B vs C vs D)
3. Set traffic split (typically 25% each for 4 variants)
4. Min 7 days, recommended 14+
5. Need 90% confidence for winner declaration
## What works in Play experiments
- **Icon variants** — single biggest install lift potential
- **Screenshot order/style** — first 2 drive most installs
- **Short description variants** — keyword effects
- **Long description variants** — less impactful, longer test
## What's harder to A/B
- Title (rarely changed, branding-tied)
- Category (massive change, not A/B-able)
- Pricing (not part of listing)
## Statistical requirements
- Min 1000 installs per variant for significance
- Top apps: 7 days for clear winners
- Long-tail apps: 14-28 days
## Common variants to try
1. Icon: cartoon vs photo vs logo
2. Screenshot 1: text-overlay vs phone-only vs lifestyle
3. Screenshot 1 hook: "TRACK X" vs "SAVE Y" vs "FINALLY Z"
4. Short desc: outcome-led vs feature-led
## Multi-locale strategy
Localized experiments per country. India + Brazil might have different winners than US.
## Apply winner
When confidence >90%, click "Apply" → variant becomes new default. Then start next experiment.
## Common gotchas
- Low install volume → no significance, run longer
- Tested too many variables at once → don't know which mattered
- Promoted "winner" with marginal lift → keep iterating
- Forgot to localize → US winner ≠ India winner
## Pair with
- `design-screenshots-play` for screenshot variants
- `design-app-icon-adaptive` for icon variants
- `custom-store-listings` for per-country variants
