GitHub's Silent Problem: Why 14,000 Volunteer Code Fixes Never Get Finished

When you use a free app, a chunk of it was probably built by volunteers. They suggest changes to code through something called a "pull request" — basically, handing in homework for a teacher to grade. Someone reviews it, asks questions, and either accepts it or sends it back. But a new study of 19 popular GitHub projects found that thousands of these hand-ins just... go silent. Nobody rejects them. Nobody accepts them. They simply sit there.

The researchers dug into 14,234 stuck pull requests and 164,562 pieces of review conversation. The surprise: most of these weren't abandoned because the volunteer lost interest or the code was bad. They stalled because of everyday coordination failures — a reviewer never replied, feedback came weeks late, nobody said who should do what next. Many were also blocked by technical snags: automated tests failing, mismatched software versions, or settings that didn't match the project's setup. In other words, the code was fine; the plumbing around it wasn't.

That silence has a cost. Only 39.56% of the people whose work got stuck ever submitted another fix, and reviewers re-engaged with those same volunteers just 21% of the time. So one bad experience often ends the relationship permanently — and the pool of free helpers shrinks a little more.

The team recommends simple fixes: flag different types of contributions differently instead of treating them all the same, give clearer feedback, assign someone to own the next step, and make those automatic "stale bots" (reminder programs that nudge quiet threads) smarter about why something went quiet, not just how long it's been silent. This isn't just a coder problem. Every free tool you rely on — from browser extensions to small business websites — runs on this unpaid pipeline. When it clogs, updates slow down, security holes stay open longer, and the software you depend on quietly gets worse.