Days spent away from the product, and the ones that pay
The line sounds like a call to work harder. Read against Spolsky's own writing, it is more about what counts as working on the product at all.
Every day that we spent not improving our products was a wasted day.
The line fits the early years of Fog Creek Software, the company Joel Spolsky started with Michael Pryor in 2000 as a consulting shop that also built its own products. Another page here follows the marketing experiments of that period. This one looks at a different question the line raises: what actually counts as a day spent improving a product.
Spolsky’s best-known answer came in 2000, in an essay about Netscape. The company had decided to throw away its browser code and rewrite it from the beginning. For roughly three years it shipped no major new release while the rewrite went on, and Microsoft’s browser took the market. Spolsky’s point was that a rewrite feels like improvement to the engineers doing it, but for users it is a long stretch in which the product they have does not get better. By the standard of the line, every one of those days was wasted, however hard people worked.
The same test catches smaller habits. An internal reorganization, a redesign of the company logo, a week of meetings about the next year’s roadmap: each can feel productive while the software customers use stays exactly as it was. Asking at the end of a week what changed for the user is a blunt and useful check.
The line does overreach in one direction. Fog Creek’s consulting days did not improve its products, but they paid the rent while the products found customers. Days spent raising money, hiring a strong engineer or closing a first large sale can matter more than a week of polishing. A company that only improves the product and never sells it can have the best software in its category and still close.
There is also improvement that customers never see directly. Fixing the build system, writing tests or paying down code that slows every change will not appear in release notes. Done in moderation, it is what lets the visible improvements keep coming. The Netscape lesson is about scale: small repairs continuously, not one enormous rebuild.
So the line works as a default and a warning. Most days should leave the product better for someone using it. When a day does not, there should be a clear reason, and “the engineers wanted a clean start” is rarely a good one.