Search Engine Optimisation For MSN Search
Posted on 5th July, 2007
Previously I’ve written a post where I points out how users coming from MSN are the best AdSense clickers. In response to that, one of my readers asks what’s the best way to attract traffic from MSN.
For a start, you can buy PPC traffic via Microsoft adCenter then direct the traffic to your pages with AdSense ads that offer higher payout than what you paid for your PPC. This is known as AdSense arbitrage where people make money by bidding on less expensive keywords to direct people to a page with AdSense.
However this is not what this article is all about. I am going to explain how you could use search engine optimisation for MSN to drive organic traffic to your blog.
While I know that common SEO practice such as keywords in the domain, keywords density, site relevancy etc. do play some roles, I can’t tell you for sure the exact MSN rankable traits. Essentially, this was my attempt to reverse engineer the algorithm used by MSN to identify the factors they used when ranking a page.
Reverse engineering the search engines used to work well in the 90s. This gave rise to the creation of many automated ranking software such as WebPosition.com. to help people create search engine optimised pages easily. Later, as the search engines became cleverer and complex, such software stopped working.
Reverse engineering is less effective nowadays because search engines update their algorithm more frequently, as a result, a discovery could be outdated by the time it’s published. While it’s not impossible to reverse engineer on page ranking factors, off site ones such as link popularity, relevancy, and their reputation are harder if not impossible to test due to the dynamic nature of the web.
Now this is not a scientific study nor do I claim the results definitive. I hope this article will help my readers especially those who are just starting out understand the basic of search engine optimisation.
The best way to reverse engineer a search engine is by studying lots of keyword rankings across a wide range of niches. Then, create several test pages. Each page tests a possible ranking factor such as keyword density, keyword proximity, anchor text and so on. Well, that’s one thing, trying to identify and test the off site factors are a different story all together.
Unfortunately, I don’t have the time to do detailed study or create test pages. So I had decided to analyse one keyword that I am already familiar with, and for that purpose I used “sabahan”. I compared the top three results from MSN.com and Google.com and tried to identify the traits that contribute to the their rankings.
The search for “sabahan” produces the following top three results from MSN Search.
While Google returns the following result. As you can see, the top result in MSN drops to the second place in Google.
Google alone uses hundreds of factors to rank a page. This information is highly guarded and never been divulged by Google. Same thing applies to MSN. So I came up with several actors which most people believe are used to rank a page.
Onsite SEO Ranking Factors
For the onsite factors, I’ve tested the obvious such as keywords in meta tags, page layout and coding, and keyword density.
Meta Tags & Title Tag
In general, most search engine optimisers agree that Google typically ignores the meta tags i.e. the description and keyword tags. So, placing your keywords in the meta tags won’t improve your ranking. In contrast, MSN does take into account the presence of keywords in the meta tags. The title tag is a different story as it’s one of the most important ranking factors used by most major search engines.
Interestingly, the results I got tell a different story as you can see below.
MSN Top Three Results Meta Data
The following shows the meta data from the top three results from MSN. As you can see, the keyword “sabahan” is only present in the keywords tag of the second page.
Google Top Three Results Meta Data
For Google, the keyword is present in the keywords tag of the top result only. The third page doesn’t contains any meta tags.
As for the title tags, "sabahan" is present in all titles of the top three results in Google. However, I discovered that sabahan.net which is the third on the ranking does not use any title tag. Google has automatically created a title for it based on the keyword being search.
Based on these findings alone, it’s not possible to conclude how important (or useless) the keywords and descriptions meta tags are to Google and MSN. One thing for sure, they are not as important as the title tag.
Page Layout & Code
The size of the top three pages vary greatly and I don’t see any preference towards smaller or larger page size when it comes to ranking them. That doesn’t mean that Google, and MSN will crawl over-bloated pages. The best thing you can do is create clean and lean HTML pages so as to avoid any potential problems when their crawlers visit your pages.
The size of each page is shown below:
- 45 KB
- 55 KB
- 2 KB
- 210 KB
- 47 KB
- 14 KB
On page keyword density
The following are keyword density reports for the top three results in MSN. The keyword here is "sabahan".
The following are keyword density reports for the top three results in Google using the same keyword.
Based on the top two results, it appears that MSN prefers higher keyword density (above 10% in this case) in the title tag while Google prefers relatively lower density. Similarly, MSN seems to prefer higher density in the body section and H1/H2/H3 tags compare to Google.
Then again, the third results in both Google and MSN contradict this possible preference in keyword density level. All I can say is that there must be another ranking factor (or a combination of them), likely an offsite one, that’s more important than on page keyword density for MSN.
Again I won’t be surprise if this is also the case for Google.
For this test, I used Yahoo! Site Explorer to determine the pages link popularity. Keep in mind that the tool counts feed URLs as well. This explains the higher than expected number for Google top result.
The following shows the number of inbound links pointing to each page.
- 50 Most Influential Blogs in Malaysia: 421
- About « Sabahan.com: 2497
- Sabah Tourism Board Official: 222
The problem with this limited amount of information is that we are unable to determine the reputation of those inbound links which I think is more important than the number itself.
Anchor Text Used In the Inbound Links
To help me find the anchor text used in the inbound links, I use Backlink Watch. In this case, I wanted to see if the frequency of the keyword “sabahan” appears more frequently in the anchor text of the inbound links.
MSN 1 : Most inbound links use “ 50 most influential blogs in Malaysia ” as the anchor text.
Google 1 : “sabahan” is used frequently in the anchor text of most inbound links. This is not surprising because the top ranking page is my homepage.
MSN 2 : Out of 18 inbound links, only one uses the keyword “sabahan” as its anchor text.
Google 2 : 3 out of 21 have “sabahan” as their anchor text.
MSN 3: Only few inbound links use the keyword as their anchor text
Google 3 : Most inbound links use “sabahan” in their anchor text
Based on this result, it appears that Google put a lot more importance in the anchor text of the inbound links when ranking a page. Conversely, this is not one of MSN’s major ranking factors.
From the findings, if all the other ranking factors are equal in weight, your page will likely rank higher in MSN with a higher keyword density in the body section and H1/H2/H3 tag. This unfortunately raises the problem that optimizing for MSN may have your page flagged as spam with Google.
In addition, it’s not surprising to see that Google put a lot more emphasis on the anchor text of your inbound links.
Some people claim that MSN tends to pick up and credit new content and inbound links more quickly than Google. I didn’t see this trend in this experiment but if that’s the case you can rank relatively quickly on MSN just by using the right tactics.
Lastly, more experiments are needed in order to derive more conclusive results. As I said earlier, reverse engineering the search engine does not work as it used too, so instead of trying to reverse engineer the search engine algorithm, your time is better spent on producing good quality content.