Brands influence purchases; there is really no argument about that. The art of advertising has paired a brand with a single adjective since the ads moved beyond basic product descriptions. For example: Coca-Cola=Classic, Gatorade=Quench, and Volvo=Safety. In today’s market, Google=Search and Bing=Decide. What would happen if brands were to disappear and only the bare bones, no glitz or glamour services and products were left behind? Michael Kordahi has provided the public with just that, a way to compare search results without the big three’s branding.
What is BlindSearch?
Kordahi has taken the famous Coke v Pepsi blind taste test and applied it to modern day technology. The introduction of BlindSearch has given Internet users the ability to search for results from Google, Bing, and Yahoo! without any branding or layout.
BlindSearch provides the results in three side by side columns with a voting button at the top of each. Each column represents results from Google, Bing, and Yahoo! The user is able to vote on the column which most closely matches their desired results. After voting, the buttons are replaced with the search engine logos to reveal which search engine’s results most closely matched your search. Don’t get me wrong, this isn’t anything brand new. There are other search engine comparison sites out there. However, I found this one to be particularly blog worthy since the results are listed side by side rather than tabbed as in Zuula (not to mention a fun little test).
Why I’m writing about it
BlindSearch isn’t really a big deal per say. It is just a fun, quick way to see what search engine may be better for you without having a hundred experts telling you which one you should use. On the plus side, BlindSearch helps in beating out some lazy tendencies when doing deeper web research. Additionally, even if at a base level, it may be a helpful nudge to some users in considering other search alternatives in their every day routines.
In no way is BlindSearch a statistical tool that will be the end all be all of the search engine supremacy argument. Even if it was 100% effective, it could render the same results as the Coke v Pepsi test, in which Pepsi was more generally liked but Coke was still the dominating force in sales.
The first to clarify BlindSearch’s discrepancies is Kordahi with a clear disclaimer on the BlindSearch homepage.
“The system has many flaws that I know about already, the primary one of interest is the lack of localisation. So, all searches are going through the US as US searches. The other deficiency worth noting is that there is much missing from the actual experience of using these search engines eg, image thumbnails, suggestions, refine queries, etc.”
There are other arguments that can be made against the accuracy, relevancy, and even importance of this experiment.
First, it could be that modern search engines are already too much alike in terms of results. Search engines have begun to emulate the leaders, taking what is effective and applying it to their own engines.
A simple comparison could be described with handbags. Designer handbags are extremely popular, high end stores carry these high end bags. In order to provide an alternative, lower end stores emulate the designs, colors, and patterns and sell similar bags for a cheaper price. After a bit of time, knockoff purses are released that are identical to the high end originals and if done correctly, can rarely be told apart. Basically, the best was emulated, and now, even the competitors have a similar product.
A second argument is shown when viewing the top ten results. The list shows a fairly generic breath of search. None of the ten results show any type of long tail searches and therefore do not really replicate normal search. This could be in part to lazy testing and a desire to get a result as quickly as possible to “test” a user’s search engine preference. It could also be in part to the limited amount of data, only 600,000 queries to date.
Lastly, BlindSearch is a sort of site where users would try to get off the wall results and test the boundaries/parameters. The site is small and the audience is certainly a select sample.
Kordahi released the following results with roughly 8 weeks of data:
Google: 41%, Bing: 31%, Yahoo: 28%
Although an employee of Microsoft, Kordahi makes it abundantly clear that this project is not initiated by or affiliated with Microsoft.