Serving different content based on IP location

[ChatterBlock]

I have city centric website.  For sake of simplicity, say I only have 2 cities -- City A and City B.

Depending on a user's IP address, they will either get City A or City B.  Users can change their location through javascript on pages.  But there is no cross-linking between cities.  By this, I mean that unless you can read or execute javascript, there is no way for you to get from city A to City B.

My concern is this:  googlebot comes to my site, and we serve them up City A.  How does City B get discovered if Googlebot doesn't read javascript?

We have an xml sitemap plus plenty of backlinks to City B.  Is this sufficient?

Should I provide a static link to City B (and vice versa) on the homepage for crawling purposes?

[qurve]

I do the exact same thing (local business pages based on visitor IP) but you can change your location based on what search terms you enter.

What we also do is allow anyone to browse any state/city results through our faceted search and we have XML sitemap entries for each state/category landing page which will then link down to city level searches.

We have seen no problem with google indexing our site (currently almost 500,000 pages indexed.)

As long as you don't actively hide content that doesn't pertain to the requesting site IP and you provide some way for Google to find it, you should be OK.

[Dr-Pete]

Adding to Daniel's comment, I'd say the big difference "...through our faceted search." It's important to have both the XML entries and a crawl path. An XML sitemap may be enough to get the pages indexed, but they won't inherit any internal link-juice. That comes through your internal links. Somewhere, there needs to be a link that Google can crawl to the other cities.

The direct back-links will help, and should get you indexed and possibly ranking, but you're still losing the authority from the domain as a whole that you'd inherit via internal links. The upshot is that you'll lose ranking power.