Old links never fully die
Bookmarks, documents, emails, forum posts, and other people’s articles keep pointing at URLs long after you removed them. Every one is a visitor who wanted something specific from you and got an error page.
Read the 404 log as a to-do list
The URLs being requested and failing tell you exactly which redirects are missing, ranked by how often people hit them. Almost nobody looks at this, and it is among the cheapest recoveries available.
Your 404 log is a list of people who wanted something from you and didn’t get it.
Redirect to the nearest thing, not to home
A visitor looking for a specific article does not want your homepage. Point them at the closest equivalent, which is also how pruned content should be handled rather than simply removed.
Make the 404 page work
Some requests genuinely have no equivalent. That page should explain plainly, offer search and the main sections, and keep the site’s one action available — a dead end with a route out converts better than an apology.
Keep redirects forever
Redirects are cheap to store and the traffic they rescue keeps arriving for years. Removing them during a tidy-up is a common self-inflicted wound, particularly after a redesign.
Fix the internal ones properly
Internal links should point at the current URL rather than relying on a redirect. Chains accumulate, slow every request, and eventually break — the reason link targets deserve automatic validation.