URL structure: what matters, and what is worth changing
You are deciding on a URL structure, or wondering whether your current one is holding you back.
Short answer
URL structure is a minor ranking consideration and a major migration liability. Readable, stable, lowercase URLs with words rather than IDs are worth setting up correctly at the start. Changing an established structure to a slightly better one almost never pays back the cost of redirecting every page.
What actually matters
That the URL is stable, unique per page, and readable enough that a human can guess what is behind it. Keywords in the URL are a very small signal at best. The larger effect is on click-through and on the link text people use when they paste a bare URL.
Depth is a proxy, not a cause
Deeply nested URLs are not penalised for being deep. What correlates is that deep pages tend to be far from the homepage in clicks, which means fewer internal links, which means less crawling and less authority. Fix the linking, not the slash count.
Words, hyphens, lowercase
Hyphens separate words in a way engines read; underscores historically did not. Mixed case invites duplicate URLs on case-sensitive servers. Numeric IDs work but tell nobody anything, including you when you are reading a report at speed.
Dates in URLs
They make content look stale the moment the year turns, and they make updating an evergreen page awkward because the URL contradicts the freshness. For news they are reasonable. For guides they are a slow-acting mistake.
Parameters
Tracking and filter parameters multiply one page into many URLs. Handle them with canonicals pointing at the clean version, and keep them out of your sitemap and internal links. This is where duplicate-content problems usually originate on otherwise tidy sites.
Why changing is rarely worth it
Every URL change is a migration, with the full cost: a complete mapping, permanent redirects, updated internal links, updated sitemaps, and a period of turbulence. The upside of a marginally nicer structure does not cover that. Change URLs when they are actively broken, not when they are merely imperfect.
If you are starting now
Pick something flat, readable and stable, and then leave it alone. The compounding value of a URL that has not moved in three years exceeds any structural cleverness you might design instead.
Questions
- Should my URLs contain keywords?
- A small benefit at most, and it comes mainly from readability rather than matching. Write the URL for a human reading it in a link and you will get the keyword effect as a side effect.
- Do trailing slashes matter?
- Only in that you should pick one and be consistent. A page reachable both with and without a trailing slash is two URLs, and that is the actual issue rather than which form you chose.
- Is a subdomain or a subfolder better for a blog?
- A subfolder keeps everything on one hostname, which is the simpler and generally safer default. Subdomains can be treated more separately, which matters more the weaker your site is. If you have a choice and no strong reason, choose the subfolder.
Measured, not asserted
ihatepdf.cv used flat, readable, one-word-per-task URLs and never restructured them across 224 days, reaching 100,542 users. Stability was worth more than structure: the five pages carrying 75.4% of all clicks accumulated that entirely at URLs that never changed.
Free tool for this: Internal Link Graph. No account, nothing uploaded.
Where this goes deeper
Every number on this page comes from one complete dataset: one product taken from zero to 100K+ users on search alone, with nothing spent on advertising. The full argument is Chapter 7 and Chapter 10 of the book. Five chapters are free to read.