URL Structure in 2026: Hyphens, Parameters, Case and What Google Recommends

Key takeaways
- Use hyphens, not underscores, to separate words in URLs. Google clarified this in its URL guidance on 23 October 2026.
- URLs are case sensitive.
/Appleand/appleare different URLs, so pick one form and stick to it. - Keep parameters to a minimum. Strip tracking tags, sort options and session IDs, because they multiply near-identical URLs.
- Do not use fragments (the part after
#) to change page content. Use the History API with real paths. - Do not change working URLs without a good reason. If you must, redirect every old URL with a 301 and update your links.

Why URL structure still matters
URLs are not a ranking trick. Their job is to be a clear, stable address for a page, one that crawlers can handle efficiently and people can read. Google’s guidance asks for a URL structure that is simple, logical and organised in a way humans can understand. When a site gets this wrong, the damage is usually quiet: duplicate versions of the same page, wasted crawling and links that split between variants.
Google reorganised its URL structure documentation in June 2025 without changing behaviour, then clarified the hyphen and underscore rule on 23 October 2026. The rules below are Google’s current guidance, with our practical advice on top.
The rules, one by one
1. Separate words with hyphens
Google recommends hyphens instead of underscores. Programming languages often use underscores to join terms, while a hyphen clearly marks a word boundary. Joining words with no separator is not recommended either.
| Avoid | Prefer |
|---|---|
/web_design_services/ |
/web-design-services/ |
/webdesignservices/ |
/web-design-services/ |
2. Use readable words, not ID numbers
A URL such as /services/web-design/ tells a person and a crawler what the page is. /p?id=48213 tells them nothing. Google’s own example prefers a readable path over a URL full of random identifiers.
3. Think about your audience’s language
Write the words in the language your audience searches in. A German page can use German words in its path.
4. Be consistent about case
URLs are case sensitive, so /Apple and /apple are treated as different URLs. If your server returns the same page for both, convert to one consistent case, and redirect the other. Lowercase everywhere is the simplest policy.
5. Keep parameters to a minimum
Use = between a key and a value and & to add more. Google’s advice is to shorten URLs by trimming unnecessary parameters, and to remove irrelevant ones such as referral tags, sorting parameters and session IDs, since they multiply the number of URLs with near-identical content.
- Never link internally to URLs with tracking parameters. Use them only on links you share externally.
- Give filtered or sorted versions a canonical URL that points to the clean page, unless they have their own search demand.
- For online stores, plan filters and facets carefully. They are the biggest source of URL explosions.
6. Avoid session IDs in URLs
Use cookies instead. If session-ID URLs are already being crawled, Google says you can block them in robots.txt.
7. Never use fragments to load content
Google generally does not support content that changes after a #. For JavaScript apps, use the History API so each view has a real path. We cover the related rules in our post on JavaScript SEO.
8. Percent-encode non-ASCII characters
If your URLs contain non-ASCII characters (for example Arabic, Chinese or emoji), they should be percent-encoded. Plain ASCII characters can stay as they are.
9. Use root-relative links
Google recommends root-relative links over parent-relative ones, to avoid creating bogus URLs. A link such as /services/ always means the same thing. A link such as ../services/ can resolve to different addresses depending on where it appears.
WordPress: what to check today
- In Settings, then Permalinks, use Post name (
/%postname%/). Avoid dates and numeric IDs in your main structure. - Edit auto-generated slugs. Drop filler words and keep them short and descriptive.
- Choose either a trailing slash or none for all non-file URLs, and redirect the other form. Whichever you pick, link to it consistently.
- Check that tag, author and archive pages are not producing thousands of thin URLs.
- If you use tracking parameters, make sure your canonical tags point to the clean URL.
Should you change your existing URLs?
Only when there is a clear benefit, such as fixing a structure that creates duplicates. Every change risks losing signals while Google works through the redirects. If you do change them:
- map every old URL to its new equivalent
- use permanent redirects (301 or 308), straight to the final address with no chains
- update internal links, canonicals, hreflang and your sitemap to the new URLs
- keep the redirects in place for as long as possible, generally at least a year
Our guide to a website migration SEO checklist walks through the full process.
A quick URL audit script
This script reads your sitemap and flags URLs that break the simple rules above.
import re, urllib.request
SITEMAP = "https://example.com/sitemap.xml"
req = urllib.request.Request(SITEMAP, headers={"User-Agent": "url-audit"})
xml = urllib.request.urlopen(req, timeout=30).read().decode("utf8", "ignore")
for url in re.findall(r"<loc>([^<]+)</loc>", xml):
path = re.sub(r"^https?://[^/]+", "", url)
problems = []
if "_" in path: problems.append("underscore")
if path != path.lower(): problems.append("uppercase")
if "?" in url: problems.append("parameters")
if "#" in url: problems.append("fragment")
if len(path) > 100: problems.append("very long")
if problems:
print(", ".join(problems).ljust(28), url)
URL checklist
- Lowercase, hyphen-separated, readable words.
- No session IDs, and no tracking parameters in internal links.
- One canonical form per page, with a self-referencing canonical tag.
- HTTPS everywhere, with one consistent host (with or without www).
- Real paths for app views, with no hash routing.
- Redirects for any URL you retire, and no redirect chains.
- Sitemaps listing only canonical, indexable URLs.
Clean URLs are the kind of boring work that quietly pays off. If you want us to review the structure of your site, our CMS development team can help, and you can get in touch.
Frequently asked questions
Should I use hyphens or underscores in URLs?
Hyphens. Google recommends hyphens to separate words because underscores are commonly used by programming languages to join terms. Joining words without any separator is also discouraged.
Are URLs case sensitive?
Yes. /Apple and /apple are treated as different URLs. If your server returns the same page for both, convert to one consistent case and redirect the other.
Do URL parameters hurt SEO?
Not by themselves, but they multiply near-identical URLs. Google advises trimming unnecessary parameters such as referral tags, sorting options and session IDs.
Can I use # fragments for page content?
No. Google generally does not support using fragments to load different content. For JavaScript apps, use the History API with real path URLs.
Should I change my URLs to include keywords?
Only if the current structure causes real problems. Changing working URLs risks losing signals while Google processes the redirects, so redirect every old URL with a 301 and update your internal links.