The Unfairness of Google’s Ranking Fairness

Lately, there has been a lot of discussion on search neutrality and whether or not Google’s results favor its own products. There’ been even more discussion on whether or not search engines should be regulated by some government agency, both domestically and all over Europe.

Danny Sullivan of Search Engine Land has been covering this story at great length. His most recent coverage actually tested the waters, comparing search queries and Google’s results against that of a relatively recent search bias survey conducted by a Harvard professor by the name of Ben Edelman. Last January, Edelman searched on 32 terms in each of the (then) big-three engines. For the purposes of his survey, he defined search engine bias as happening when a particular engine’s other vertical products appeared in one of the first three results.

Sullivan’s results and analysis significantly differed from Edelman’s.  I have to agree with Danny Sullivan and offer some counterpoints to Edelman’s survey.

In my opinion, Edelman’s conclusions are largely due to his flawed testing methodology. He chose a sample size of 32 terms upon with to search. A sample size of 32 is not statistically significant. Of those 32 terms, 12 were the actual name or a stem of a Google product name. This means that 37.5% of his terms are a result of  bias sampling. Additionally, 25 of the 32 terms (78%) were one-word terms, which are the most broadest of matches which make relevancy in results more difficult. There were no search phrases greater than two words used in his survey. Basic data analysis suggests poor sampling.

Sample issues aside, any decent SEO knows that Page Title is the first point of on-page optimization. Google’s products all have Edelman’s “one-word” search term both in the name of the product and in the Title tag. In most cases, the term is the second word in the title tag (right next to “Google”, ie: Google Maps or Google Finance).

Danny further contradicted Edelman’s conclusions, citing Google’s “most important product,” search. He showed that a Google search for search engine didn’t even list google.com. A secondary service, Google Custom Search, cracks the top-10. However, the main Google search product does not. Surprisingly, search engine was not one of the 32 terms Edelman used in his survey.

When speaking to students about search marketing at RIT, I use several examples on visibility of brand. I pick on Pepsi, who is never found for soft drinks or other related terms. I pick on iTunes, who does not rank for music service. Lastly, I pick on Google. Google’s default search is not found for search engine, as Danny astutely points out. However, looking at other long-tail terms, Gmail ranks below many other services for free email or free email account. In fact, Google has ads running to promote its email service. Clearly this is a nod to the fact they are not stacking the deck toward their product.

Here are more non-biased, “long tail” terms:

Less Stats, More Common Sense

If a company develops a product, then provides more specific products for particular verticals and has a method of integrating the two together to give a perceived better user experience, is that wrong? Each of the search engines are doing that. An image search or a video search are extensions of the search product. I will grant you that Email is not and neither are blogs or photo sharing.  However, if I perform a typical user result trying to actually find a service offering those products, Google does not appear to favor itself. Certainly not by Edelman’s definition.

Is this any different than a pharmacist? If I need a cough syrup recommendation, is the pharmacist really going to recommend a name brand over the store brand? I think not. This is true for many markets, tech or otherwise.

Should there be regulation? Hardly. Google publishes their guidelines, which companies should heed. There are still choices. While I have concern about the amount of information Google has indexed, I feel they have done nothing that merits regulation. They are one of many services out there. Clearly people are still using Dogpile, AOL search and even Ask. People know they have choices if they don’t like the results. Microsoft is clearly promoting that with their clever “what has search done to us” commercials.

Can we put this to rest and get the focus back on more relevent results?

What do you think? Leave a comment, please.

This site uses Akismet to reduce spam. Learn how your comment data is processed.