Cheatsheet

The Shopify app math

A $49 app has to lift conversion by 32.7% at 500 visits a month to break even, and at that traffic the lift is not measurable in principle. Verified against vendor listings and Shopify's own documentation, September 2 2026.

Shopify ships 39 free first party apps. Most of what a new store gets sold in month one duplicates one of them.

The arithmetic that ends the argument

Monthly visitsConversion lift a $49 app must produce to break even
500+32.7%
50,000+0.3%

The app costs the same either way. The traffic is what decides whether it can ever pay for itself.

It gets worse. Detecting a 10% lift on a 2% baseline needs 161,362 visitors. A 500 visit store would need roughly 27 years to know whether the app did anything. A beginner buying conversion apps is paying for something that is unmeasurable in principle at their traffic, which is exactly why these get judged on vibes.

A lean stack runs about $29 to $44 a month. An over app'd one runs about $246. That gap is roughly $2,400 a year burned before the store has proven it sells anything.

What Shopify already does for free

The one thing worth paying for on day one

Reviews. Shopify killed its own Product Reviews app, delisted in September 2023 and removed in May 2024, with no first party replacement. The usual free tier covers unlimited reviews including photo and video with no order cap, and paid tiers start around $15.

Three things commonly taught that are wrong

  1. "The average store runs 6 apps" is 2021 folklore from an uncited blog post. Real data: stores under 50,000 visits average 1.8 apps, and 94% of stores running eight or more are on Plus.
  2. "Uninstalling leaves code behind" is mostly false. Script tags and app blocks remove themselves. Only direct edits to the theme file persist.
  3. Shopify retired the 0 to 100 speed score in January 2024. Anything telling you to chase that number is describing a system that no longer exists.

Two statistics to stop repeating

The often quoted "0.1 seconds equals 8.4% more conversion" study is correlational and excluded null results by its own authors' appendix. And in usability testing, a fake homemade trust seal outperformed real security seals from every vendor but one, which tells you those badges are measuring familiarity rather than trust.

The store gets built with the receipts

What actually moved a number, and what got uninstalled a week later.

Get in free More cheatsheets