Should SEOs Redirect or Park For Geolocation?
Geolocation is the process of associating a geographic area (usually a country) to a website or website page. This is done in order to help search engines provide results that are not only relevant to a searchers keyword, but also location.
For example, there are certain professions, like lawyers, that are restricted to geographic locations – a lawyer from Poland probably will not be helpful to a searcher in South Africa, regardless of how well the Polish lawyers website is optimized. Likewise, many people prefer to buy from vendors that are based in their own country – not only is shipping not as much of a concern, but local consumer protection laws are more likely to be helpful if their is a problem.
- This refers to Country Code Top Level Domain. Normal TLD’s such as .com, .net, and .org refer to they type of website the domain belongs to (commercial, network, or organization), whereas a ccTLD refers to where the domain belongs – ie a country. Common examples include Canada (.ca) UK (.co.uk) France (.fr) and so forth. These codes are assigned by the Internet Assigned Numbers Authority (IANA).
- The act of assigning a geographic location to a website or webpage.
- Incoming back links (IBL’s) that point from another website to yours, where that other website has a specific country code associated with it. For example, a link from a German website would be a ccLink, where a link from a normal .com with no country geolocated to it would be considered a normal link. If a site gets enough ccLinks related to a specific country pointing to it, it will often be assumed to be relevant to that country by some search engines.
In late 2003, I wrote a fairly well received article called “Only In Canada, Eh?” which outlined several issues and possible solutions regarding geolocation. However, it’s now late 2005 and the article is starting to show it’s age. The search engines have not stood still in the interim, and there are more rules to work with. A fairly complete discussion of the changes can be found at the High Rankings Forum in a thread called: “mcanerin – Only in Canada, Eh?“. This article hopes to summarize that thread and to clarify some parts of it, especially related to redirects and parks.
The main conclusion of the original article was that it would be best to host your website on an IP address in your country. This allows Google to associate your website with a specific country easily, regardless of what it’s domain extension is. This is still true today, but there are other considerations.
Other Search Engines Don’t Use IP Geolocation
Google is the only major search engine currently that uses IP address for geolocation. The reason that Yahoo won’t is quite simple – they know very well that many people host websites outside of their own country. As a matter of fact, Yahoo hosts websites from all over the world, and since it’s IP’s are US, following Google’s IP focussed method would not accurately reflect the reality that Yahoo knows is going one. Likewise, the folks at ASK are well aware that the preferred hosting location for large Chinese websites is Japan. Google’s IP fixation is apparently more due to a lack of experience outside of the US than an accurate reflection of the realities of web hosting in the rest of the world.
The non-Google search engines use 2 methods of discovering what area a web site is relevant to: 1) the ccTLD (Country Code Top Level Domain) and 2) link analysis – they assume that if a very large number of German websites link to you, that you must be relevant to Germans, regardless of your hosting location.
Geographic Metatags and GeoTags
A lot of people haven’t heard of Gigablast, which is too bad, as it’s a nice search engine. Gigablast is different from most search engines in that it actually indexes metatags, and lets you search for results using those metatags. It’s the only major search engine that actually uses Geolocation Metatags, which look like this:
<meta name=”zipcode” content=”87112,87113,87114″>
<meta name=”city” content=”albuquerque, abq, rio rancho”>
<meta name=”state” content=”new mexico”>
<meta name=”country” content=”usa, united states of america”>
If you want to experiment with geographic metatags, you can use my free Metatag Generator to do so. There is another option called GeoTags, which actually incorporates latitude and longitude information, as well. Unfortunately, it’s not well supported at all.
For now, geographic metatags don’t work with the big 4 search engines, and I don’t really expect that to change in the near future, so they are interesting, but not practical at this point.
Ok, I just made that term up. For the purposes of this discussion, ccLinks (Country Context Links) are incoming backlinks (IBL’s) that the search engine in question considers to be relevant for a particular country. So, for example, if my webpage is considered by Yahoo to be a “Canadian” page, and I link to you, then Yahoo will receive a hint that you are of interest to a Canadian.
If Yahoo get’s enough of these from other Canadian sites, they will eventually come to the conclusion that this site is somehow relevant to Canadians. Notice that it’s not necessarily assuming that it’s a Canadian site, but that Canadians would consider it relevant, which may be a better metric anyway.
So it’s possible to get your .com, US hosted site considered to be relevant to the UK if you have enough UK relevant sites (ccLinks) pointing to it. This won’t work on Google at this time, but does work on Yahoo, MSN and Ask. This of course can cause issues if you somehow buy/trade a whole bunch of links to your site from Russian link farms, of course.
Further, I believe from testing that each page can only have one country associated with it. In theory, you would have 20 pages on your site all considered belonging to 20 different countries. In theory. In practice, it’s something to be cautious about while link building – you want to focus on ccLinks that are from your target country as much as possible. Which should be no surprise since presumably that’s part of relevant link building.
ccTLD – The Most Accurate Method
Currently, the best way to make sure that a website is geolocated for a specific area is to make sure it has a ccTLD (Country Code Top Level Domain) for that location. For example, if you wanted to make sure that your website was considered relevant to Canadians, you would make sure that it had a .ca domain extension.
This works very well in all the search engines today, including Google. It’s simple and clear.
But there are problems. First, many people already have a well-branded .com site. Second, some countries make it very difficult to get a domain appropriate to that country unless you are incorporated or a citizen there, which can be an issue if you are a multinational corporation, or a travel company that is from the destination location but is marketing to the users location.
So What’s This Have To Do With Redirects?
Good question. The answer is that one of the most common use for domain redirects is geolocation using ccTLD’s. If you have a .com and want to show up as a country specific site, your best and easiest method of doing so is to use a ccTLD and associating it with the .comwebsite. You can do this by using redirects. I will use .com as the example from now on, but the information below also refers to .net, .org, and other non-country specific domains.
Geolocation is one of the few instances where I recommend a 302 or park instead of a 301.
I have an existing .com site and I just bought a ccTLD. I want to have the .com website associated with that country.
Answer 1: Use a 302 redirect from the ccTLD to the .com. Create a second account and point the ccTLD to it. Then redirect using a 302 from that account to the .com account.
This tells the search engine that your ccTLD is the “real” domain and that it’s being temporarily redirected to the .com. The search engine will index the .com, but keep the ccTLD as the “original” domain. In short, the .com won’t be considered. If you pointed the .com to the ccTLS with a 302, the .com would be kept and you would lose the benefit.
The problem with using a 302 redirect is that it applies on a page by page basis, which is good if you want certain pages of your site to be associated with different countries, but harder when you want the whole site to be associated. In this case, you would need to create a sitemap with the ccTLD coded as an absolute (not relative) link to each page of your site you wanted indexed as belonging to the country.
Answer 2: Park the ccTLD directly on the .com. This associates BOTH domains to the site. Since one is a ccTLD, the site will be considered geolocated. Remember that if you do it this way it may take awhile for the search engine to figure out that this is really the same site, so for a while you will split your link popularity between the 2 domains, until they are merged. This also creates a potential duplication issue during this period.
The good news is that if you use relative links, then usually both domains will eventually be indexed for each page, and your whole site will be considered associated with the ccTLD. If you don’t want this, you will have to control it using absolute links or 302 redirects.
I have a ccTLD website, but I also want to add my new .com onto it without messing anything up.
Answer: In this case, you would do a 301 redirect (not a 302) from the .com domain to the ccTLD. This tells the search engine to pass on all link popularity to the ccTLD, and to not consider the .com as the “proper” website.
Important Step for All Geolocation Redirect Scenarios
In all of the cases above, you need to do link building to the ccTLD. If you only do link building to the .com, there will be no opportunity for a search engine to index or know about the ccTLD. In view of the ccLinks issue that Yahoo, MSN and Ask use, I strongly recommend that you do as much ccLinking to the ccTLD as possible.
Geolocation Spidering Issues
Remember when I said that every page has it’s own country potentially linked to it? Each page can only have ONE country linked to it.
How would that be accomplished? Well, the simple method would be to consider all the pages under the ccTLD domain to be long to that ccTLD. Makes sense, right?
Well, it’s a little more complicated than that. You see, most sites use relative links within them. Some use absolute. Whether or not you use relative or absolute links can have a difference on whether or not your websites pages are considered geolocated or not.
Imagine that I point mcanerin.ca at mcanerin.com using a 302. In reality, what I’m doing is allowing the .ca domain to resolve any webpage on the site. But I’m not necessarily actually resolving all those pages.
Let’s say that a spider follows a link to mcanerin.ca to the default page of my website. Now, what country is going to be associated with that page? Why, Canada, of course. Now, let’s say that there is a link on that page that points to mcanerin.com/tools.htm. What country will that page be associated with? Nothing, really. Because the link to it is using the .com extension, not the .ca one!
That’s one of the dangers using an absolute link within your website regarding geolocation. The absolute link takes precedence over the previous redirect. Now, if the link had just been to /tools.htm, then the spider would have followed that relative to the .ca domain it was currently in and the page would have been associated with Canada in that case.
One use for absolute links within geolocation is the use of a geolocation sitemap. This is an ordinary sitemap to your site, but with 2 exceptions: 1) it’s pointed to with at least one (preferably many) absolute links using the ccTLD, and 2) it contains a list of all the pages in the site that you wish associated with the ccTLD listed using absolute ccTLD links.
In short, it points to every page using the ccTLD in the link, thus associating that ccTLD to each of those pages.
If you don’t do this, or alternatively do a lot of link building with the ccTLD and then use relative links throughout the site, you risk having only your home page, and maybe one or two others, being associated with the country, rather than your whole site, unless you use parking.
Geolocation can be a tricky area. I specialize in it and sometimes find it confusing. The key is to remember that each page is treated separately, and to follow the geolocation. Also, remember that a 301 throws away the old and keeps the new, a 302 keeps the old and ignores the new, and a park keeps both. If you keep these things clear in your head while you plan it out, you’ll get through it fairly easily.