There could be some serious flaws in the algorithm that Microsoft's new search engine Bing uses to index South African web sites.


This is according to Howard Rybko, owner and director of Syncrony, writing in an opinion article reproduced here:
The launch of Bing, Microsoft’s new search engine, has brought to the fore the question of what strategy is best to adopt in order to secure and capitalise off the highest rank possible on any Bing search.
In June, heralded by a worldwide marketing campaign, Microsoft launched Bing, its new search engine.  Could it be possible that at last Microsoft has learned how to do search?
Awakened by the marketing hype, many South African website owners no doubt wondered how Bing would affect their site rankings and the flow of Web leads. The launch has left many a website owner contemplating how to best prepare websites to ensure the highest possible ranking on any Bing searches?
Unfortunately at this stage, a great deal more research must be done before we can solicit a comprehensive answer to this question.
However my early research attempts lead me to believe that the Bing algorithm used for indexing SA websites (local search) has some serious flaws.
It is not easy to determine how Bing evaluates and rates content that local searchers will deem to be valuable.  All we can do to assess this with any credibility is by looking at what pages appear as a result of a search.
It seems to me that the current state of Bing’s local indexes represent a heady mix of well past sell-by-date content spiced with a good dollop of fresh, poorly selected links.
I acknowledge that my tests targeted a miniscule portion of the possible search indexes and results available in Bing. I am also sure that these results will improve with time, but at the moment, despite a fresh new look, Bing has a lot of homework to do before it can match the skills of Google.
I did a few quick tests using the search term “web design” to test how a Bing search compares to a similar Google search. My scenario was to suppose that Company X planned to build a web site and was searching for a suitable web design company. I wanted to compare the results a web user would get by using either a Google or a Bing search.
A Google Search for “Web Design” (South Africa) on 9 June 2009, produced a top five listing that looked like this:
* www.web-design.co.za
* www.tm4y.co.za
* www.intoweb.co.za
* www.o5webdesign.com
* www.coza-web.co.za
The first address points to a portal targeted at web designers & web design companies, not the web design company we were looking for. The other four suggested sites belong to serious web design organisations.
These companies no doubt expend significant effort ensuring that they attain and retain their high Google ranking. Their efforts probably include ensuring Web pages are Google optimised, creating original content on a regular basis, posting original content Blogs and news articles. All this hard work also ensures that they steadily accumulate inbound links.
What if Company X carries out the same search in Bing? Do the results look similar? Not even close:
* www.branchmedia.co.za
* www.starwebdesign.co.za
* www.webdesignsa.co.za
* www.getwebdesign.co.za
* www.web-design.co.za/blog/?p=18
One similarity – the www.web-design.co.za link appears in both listings. The content on the site is certainly relevant to our test searches, but look carefully and one sees that Bing seems to be doing something that is not very clever.
I can see no reason why a search engine should point first to a deeper link in preference to a home page that has at least the same relevance to the search term. To make matter worse, the link points to an OLD blog entry from August 2008.
I used an outside party, namely Yahoo!, to report on the number of inbound links and indexed pages for each of the top five pages returned in the above lists.
Inbound links are an important indication of online reputation. A link pointing from one site to another web site (especially without a reciprocal back link) represents a recommendation and is in effect a thumbs-up for that site’s reputation.
The number of unique and relevant content pages a site has is an indicator of a site’s value and shows how much effort has gone into creating and providing content.
It does not take a rocket scientist to realise that there is a significant order of magnitude in disparity between the two results. And yes, I realise that there are many other ways to gauge the value of a page when calculating a ranking, but these results make no sense to me at all.
Please do not get the idea that I am in any way disparaging the intent and content of any of the sites above. I am sure that all the above sites represent hard work and effort. These examples were used only to demonstrate the differences and deficiencies that appear to exist in the current implementation of Bing in South Africa.
By the way, my look at blogs and blog indexing seem to show that there is a major issue with the frequency and method that Bing uses to index blog entries in South Africa, but that is for another discussion.