Ryan Jones Blog – dotCULT.com Ryan Jones Blogs About Internet Culture, Marketing, SEO, & Social Media

May 20, 2007

Search Engine Questions

Filed under: Main — Ryan Jones @ 12:51 pm

Over on SEOmoz, rand fishkin asks If You Could Ask the Search Engines Any Question and Get An Honest, Complete Answer… What Would it Be?

He then goes on to list about 14 questions he’d like answers to.

While I’m sure that the search engines will never answer his questions, I thought it would be interesting to guess at what the actual answers are.

So, here’s my guesses. ( clockwise Important: These are just guesses, in no way can I verify if any of these answers are correct or not. They’re just what I think the answers are. I’m only a computer scientist, I do NOT work at any of the search engines or have inside information)

What role do search quality raters play in determining rankings? I don’t think that the quality raters affect rankings in any way. I think that quality raters are used to determine a valid result set in which various algorithm changes can be tested against each other. Another possibility is that they’re used to view the “reported” sites – however I doubt this.

How much impact do the other domains owned by / registered by a site owner have on the way a site is viewed/treated algorithmically? This depends. If the owner isn’t doing anything shady then they shouldn’t matter. I, for example, own several unrelated sites all on the same server. They don’t cross link and they don’t have problems. If the “other” sites I own are part of a network of spammy sites then it can have an effect. The last company I worked for had a network of hundreds of sites (all of which were thin affiliates for the same site) that just targeted different keywords. I can see this type of setup hurting your sites. If it’s a normal relationship though, I wouldn’t worry.

What is the purpose/motivation behind obfuscating accurate, precise link data? The most likely answer is that it’s probably easier and quicker to do it the way they’re doing it. Another possibility is that they don’t want people to try to reverse engineer the algorithm. There’s 2 main things that make a search engine unique: The sites it has indexed, and it’s algorithm. Everything that goes into making a ranking decision is public; that is to say that there’s nothing Google has access to that you can’t get. If they made the links public, it’d be possible (for a small subset of sites…. say 1 query 30 sites) to try to reverse engineer an algorithm and guess what’s getting more weight.

Additionally, I don’t think it would be a good idea. When they released PageRank, webmasters went crazy trying to raise it. If they give total number of links, I’m sure there will be a huge race to achieve more links than your competitors – whether relevant or not.

In less than 100 words, describe why you choose to rank Wikipedia above accurate sources? I think this answer is simple too. Google’s main factor is links. Since wikipedia is so broad, it has links from TONS of places. Since PageRank carries on across the domain, links to one article can help boost the PR of other articles.

Another reason is that Wikipedia is peer edited. It’s not the view of one person (when you look at any more credible site, it’s only edited by 1 person.) I think though, that it’s all to do with links.

Do companies/sites that spend a lot with your engine receive any SEO benefits (free consulting time, a few tricks from an engineer, etc) This one I can honestly answer: No! My old company maxed out 2 credit cards / month on Google Adwords. We got cool things like pens, a shirt, and a USB key in the mail, and a phone number to call in regards to our adwords account – but we still had sites banned from the index and get penalized.

I don’t think that Google plays preferences to people who spend more.

Does your engine ever use the predictive abilities of search keyword demands to profit outside the world of search? This one is interesting. I don’t think they do it directly, but I’m sure it has something to do with what types of companies they acquire. For example: If you looked at a trend for job search site queries, I’d bet that they’re up. Why? Because MS, Yahoo, and Google are all buying job search sites right now. Either they’re predicting a big down turn in the market, or they’re noticing that a lot of people are looking for and using those sites.

I think that’s enough for now. For the full list of questions please see Rand’s site (linked above). There’s some more good questions in his comments. Feel free to leave some possible answers to some of the questions in my comments if you like.


  1. I’d like Google to explain the peculiar differences in the results they deliver for searches at Google UK when users choose the (default) approach, which provides results from “the web” and the alternative, which provides results designated as “pages from the UK”. (Similar peculiarities can be seen by users in Australia and other countries). Some sites that belong to businesses located in the UK and are hosted in the UK appear in results from “the web” but not in results from “pages in the UK” and some UK sites that appear in both results have more individual pages listed in “the web” results than in “pages from the UK” results. Another peculiarity is that the Home Page of a site (i.e. the default/index page of the site) is often missing from the “pages from the UK” results (although other site pages are included) but always appears in the results when “the web” option is used.

    Comment by Jon Dale — May 22, 2007 @ 2:54 am

  2. I’d ask “Is Jeeves Gay?”

    Comment by Alexander — May 23, 2007 @ 10:27 am

  3. Might be.. he is spending his retirement sailing the ocean:

    Comment by Ryan — May 23, 2007 @ 10:39 am

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.

Powered by http://toastmeetsjam.com/2013/08/28/definition-of-the-single-girl/ WordPress