Patrick McKenzie switched Bingo Card Creator from a Swing desktop app to a web app and found the web version strictly dominates on every business metric.
Key Takeaways
Web app trial-to-purchase conversion was 2.32% vs 1.35% for the desktop version despite fewer features and only 2 months of development.
AdWords CPA dropped from $20 (desktop) to $9 (web), letting McKenzie outbid competitors who sell desktop shareware at the same price point.
Support requests from last 50 customers: 15 for the desktop app vs 3 for the web app; installation issues and lost registration keys drove the desktop volume.
Web analytics revealed counterintuitive product decisions: more features hurt sales, “baby shower” was the dominant use case, and 90%+ of purchases happen within 2 hours of first use.
The shareware funnel has 17 discrete failure points from Google search to checkout; web apps collapse most of them by eliminating download, install, and re-launch steps.
Hacker News Comment Review
The dominant thread is temporal context: multiple commenters flagged this was written in 2009, before Electron, before mobile-first B2C, and before AI-native tooling reshaped the desktop question entirely.
Commenters note the business-conversion framing only applies to commercial software; open source desktop apps face none of the funnel or piracy constraints McKenzie describes, making his conclusions largely irrelevant for that audience.
The implied follow-up question – what do these metrics look like now, with mobile as the dominant consumer platform and Electron blurring desktop/web – is unresolved; no commenter posted updated data.
Notable Comments
@sudb: Asks what the numbers say today and how Electron changed the calculus; frames mobile as the current best-metrics platform for B2C.
@ryandrake: “Almost all of Patrick’s points are great if your software development goal is to make a buck” – cleanly carves out open source as a separate domain where these arguments don’t apply.