Folders vs Subdomains vs ccTLD in International SEO - An Overview

Thoughts, Graphics and Formulas to Improve Decision-making

The recent whiteboard Friday session from SEOmoz on International SEO: Where to Host and How to Target was a welcomed overview of how SEOmoz see this landscape. Together with Dan Taylors How To: International SEO and Matt Cutts on signal strengths of ccTLDs vs Geotargeting in Google Website Optimizer vs. IP-address, we start to see a clear picture (or less blurred), about when you should use each of the three methods of geotargeting:

• ccTLDs (Country code top level domains) - like mysite.fr
• Subdomains - like fr.mysite.com *
• Folders - like mysite.com/fr/ *

* It's assumed that you use Google Webmaster Tools (GWT) console to geotarget these under all circumstances.

This post is a summary of existing knowledge and intended to help you answering tricky questions from customers and colleagues. And perhaps as a reminder for yourself :-)

Main pros and cons of each geotarget-type

Domain authority: is any metric of trust and authority that search engines attribute to a domain. Contrary to PageRank and Trust Rank, Domain Authority does NOT flow between unique domains. Age of the domain as well as unique domains linking to a domain are obvious metrics that make up Domain Authority. You can expect Domain Authority to flow freely to subfolders, but only partly (or only sometimes) to subdomains. Creating a mycollsubdomain.blogger.com is not expected to give you all that credit from blogger.com you may hope for. Pages receiving good Domain Authority will rank without external links, just as you see with any article published on Wikipedia.

GEO targeting: As can be seen from Matt Cutts' wrapup, a ccTLD has bonus attached when it comes to geotargeting. With folders and subdomains you have to stick with the less effective GWT console. Most international SEOs agree this is correct.

Maintenance Cost: Some organizations are more efficient than others, but everything else being equal, buying, annual payment, DNS setup etc. has an attached cost og using more domains. Subdomains also require a correct DNS setup, special handling in the CMS etc. Folders on the other hand are supported by most CMSs.

PageRank: PageRank is listed here due to the often seen misunderstanding that your choice of geo-targeting affects PageRank and that you lose PageRank on your language specific pages if you use e.g. a subdomain or ccTLD. It doesn't. The PageRank algorithm works on links between URLs, not domains or subdomains. If you launch a new website it doesn't change the PageRank if you use http://www.mysite.com/fr/ over http://www.mysite.fr/ for your language specific content. It is only how you interlink all your pages that matter in this regard. Hence you will see that PageRank is static in the following graphic.

The image show that if you choose a folder-solution, you gain a lot of domain authority form your main domain, and you can expect lower maintenance costs. If you choose the ccTLD solution, you get a geortarget "bonus" that will help you rank in the local countries, however you lose the Domain Authority (0 years old and only one domain linking to it), and maintenance costs are higher. Subdomains are in between, without any obvious pros except on the maintenance part (which occasionally plays a part in the workings of real world SEOs).

Earnings

In the whiteboard Friday session the French surfer problem was mentioned, and it is still a real problem out there, when doing SEO for non US websites. If you use a ccTLD website you can expect higher CTR, higher conversion rates and higher spend per order. On the downside, you have to invest more in order to get a critical mass of incoming links to your domain. With subdomains you don't need to get as many links, but don't cater as well for the French surfer as with a ccTLD domain.

The following graphic shows you have you with the right initial investment (linkbuilding), can expect a better bottom line. The folder solution has no initial investment (and starts at 0). Subdomains and ccTLDs has an initial investment and start of on the negative side. However, do to better CTRs, conversion rates etc. they perform better in the long run.

Now this is all very simplified. The graph will look different for each website. Some websites will never reach break-even if the initial linkbuilding is too expensive, the local demand is too small or the French surfer still smells something fishy.

In general larger companies with larger budgets, better brand penetration, better resources for translation and support o the local customers will stand a better chance of making a good ROI of the investment.

Model for calculating Breakeven of a ccTLD project

The faint-hearted can skip this part.

In order to estimate or calculate break-even for an investment in launching ccTLD site we need to look at four KPIs (Key Performance Indicators):

• Impressions (how many times you are shown in Google)
• Click Through Rate (CTR )
• Conversion Rate (CR)
• Average Spend

By having a ccTLD you can expect CTR, Conversion Rates, and Average Spend to increase from day one. The initial investment (for link building etc.) will increase both impressions (when going from page 2 in SERPs to page 1) and CTR (when going from position 7 to position 3). As a paradox you may see CTR fall when linkbuilding when you move from the top of page 2 to the bottom of page 1. Due to increased impressions the net-effect should still be an increased number of visitors.

If you estimate the expected increases in these KPIs you can calculate an Increased Earnings Ratio (IER ) as:

Break-even can then be calculated as:

Conclusion

Even though the above graphs and formulas may be impractical (even impossible) to use to calculate a real-life break-even for using a ccTLD site, hopefully the model can raise awareness that this approach requires an investment of additional resources. Some websites and countries will be able to benefit from this approach, through higher monthly earnings. For others the ROI will be too small (or negative) to justify the project. No two websites are equal and only an analysis of a particular website in a particular country will answer if the expected ROI is great enough to pursue.