You are viewing the demo workspace with sample data. Connect your own app to see your real audit.
asopilotDriftly

Experiments

Apple's own product page tests, run one at a time. Every test declares its hypothesis before it starts, and Amelia waits for a real answer before calling a winner.

running50% of visitors · US + CA
Day 4 of 12

Screenshot order: app first vs people first

Hypothesis
Showing the listing flow first will lift conversion, because visitors arrive from 'sell' searches.
Metric
Page conversion
Smallest lift worth finding
5%
B is ahead so far, but it is too early to call. Amelia checks with statistics that stay honest when you look early, so she will not stop on a lucky streak.
26%32%Day 1: treatment 29% vs control 28.1%Day 2: treatment 30.2% vs control 27.4%Day 3: treatment 29.6% vs control 28.6%Day 4: treatment 30.4% vs control 27.9%AB
A · controlB · treatment
wonAll storefronts · 14 days

Search asset: benefit headline vs logo

+18% taps from search

5%9%Day 1: treatment 6.8% vs control 6.1%Day 2: treatment 7.1% vs control 6%Day 3: treatment 7.2% vs control 6.3%Day 4: treatment 7% vs control 6.2%Day 5: treatment 7.3% vs control 6%Day 6: treatment 7.2% vs control 6.1%Day 7: treatment 7.4% vs control 6.2%Day 8: treatment 7.1% vs control 6.1%Day 9: treatment 7.3% vs control 6.3%Day 10: treatment 7.2% vs control 6%Day 11: treatment 7.4% vs control 6.2%Day 12: treatment 7.3% vs control 6.1%Day 13: treatment 7.1% vs control 6.2%Day 14: treatment 7.2% vs control 6.1%AB
A · controlB · treatment

Rolled out to every storefront. Saved to your playbook.

lostUS · 10 days

Header: text overlay vs clean photo

-4% page conversion

20%24%Day 1: treatment 21.8% vs control 22.4%Day 2: treatment 21.5% vs control 22.9%Day 3: treatment 21.9% vs control 22.1%Day 4: treatment 21.2% vs control 22.6%Day 5: treatment 21.7% vs control 22.8%Day 6: treatment 21.4% vs control 22.3%Day 7: treatment 21.6% vs control 22.7%Day 8: treatment 21.3% vs control 22.5%Day 9: treatment 21.8% vs control 22.4%Day 10: treatment 21.5% vs control 22.6%AB
A · controlB · treatment

Rolled back in one click. Lesson saved so it is never tested again.

Playbook

What has worked for you and for apps like yours. Amelia picks the next test from here.

  • Benefit headline beats logo in search resultsDriftly · Oct+18%
  • People using the app beat device-only screenshotsMarketplace apps · 78% of tests+11%
  • Text in the product page header gets cropped by iOSDriftly · Sep-4%
  • Local currency in screenshots lifts single-country languagesShopping apps · 64% of tests+7%