What John is talking about would be beneficial for both humans and bots, though hopefully also a programmatic pattern that could be tailored and editable.
As for the URLs, there is a happy middle ground, but I would agree for the most part with Paolo here (likewise, I’ve done agency and in-house SEO now for a decade for some of the biggest global brands around). While “overly too long” URLs aren’t desirable, for the information structure for these kinds of sites, this hierarchical URL protocol is desirable, again for both humans and bots. This is a top down, broad to focused structure, so the ordering makes sense, even though it is written “backwards” compared to regular address structure.
e.g. either
/hotels/united-states/illinois/chicago/
or
/united-states/illinois/chicago/hotels/
would be preferred and the ability to drop non-critical levels, such as the country if all locations are limited within country is ideal.
Also, if your site is on a ccTLD (country level, e.g., somesite.fr, mysite.de, etc.co.uk) from a search engine point of view, you will be sending a strong geolocational signal just based on your domain, though also a bit more challenging if you are on a .com or other top level domain, especially if your hosting is also out of country. However, in the case of the ccTLD, note that if your site is targeting outside your country, this geo-signaling may also be a bit more challenging to overcome.
Cheers
Paolo 10 years, 5 months ago.
