Is a research repository worth keeping if nobody reads it?

Answered by Falk Gottlob · Reviewed · 1 min read

THE SHORT ANSWER

Only the part with a date and a reader. Jakob Nielsen's November 13, 2026 roundup says to leave every readout with 3 committed fixes and treats the rest as filing. I keep one row under each fix: the claim, its strength, a source someone can re-open, and a re-check date. His own roundup gives the reason, since findings about people keep for decades and findings about agents spoil within months, and agents now read documentation. In my creator scan, 42 reports since September 19 with a three-run expiry, every row that went stale was a claim about what a tool could reach: three blogs recorded as having no feed that had one, a handle that returned a 404, and a profile recorded as unreadable that loaded on November 13.

Keep the part that has a date on it and somewhere to be read. A 40-page deck in a shared drive earns nothing, and Nielsen is right about that. A one-line claim under each shipped fix, with its strength, its source, and the day it gets re-checked, is what stops the next team from building on something that quietly went false.

The rewrite and the row format are in Nielsen Calls Unused Research Expensive Filing. Ship the Three Fixes, Then File the Three Claims.. The other half, checking whether a shipped fix landed, is in The Three-Call Theme Was the Renewal. Count What Landed, Not Fixes.

SOURCES

THE LONG VERSION

RELATED ANSWERS