Negative SEO blog : SEOassassin

Lots of people spend lots of time trying to work out how search engines “work”. Most of the time they fail, not because they don’t understand the answers they get, but because they ask the wrong questions in the first place. Anyone with a scientific background knows that formulating the right question is a vital first step in getting a meaningful answer.

For a start, I see far too much discussion and “analysis” that is purely Google-centric. Granted, they have the greatest market share, but they aren’t the only game in town. Comparison of data from diverse sources can also bring additional insight into data from a single source, so when researching SE issues, remember that there are alternatives.

Let’s take an example. One of the most common, if not the most common words in the English language is “the”. Since it’s so common, but meaningless on it’s own, searching for it should tell us a fair amount about the general operation of each SE. No-one targets the term “the” on it’s own, so we should get fairly natural SERPs, and those results not only stand on their own, but form a corpus of data for comparison.

First things first, the links to the searches are :

Google
Yahoo
MSN
Ask

Take some time to look them over, and we’ll come back in a few minutes.

Done? Good. Firstly, notice the similarities in sites appearing in the top 10 in Google and Yahoo, and Ask to a lesser extent - 60% commonality between G and Y, and although Ask only shares 1 result (whitehouse.gov) with the other 2, they have the same “sort” of sites, big, powerhouse domains with lots of trust / authority. MSN’s list is so different, we’ll consider it separately.

MSN

Hmmmm, lots of occurences of the word “the” in titles and domains. On-page factors are still very much in evidence here. It’s interesting to note that there are LOTS of Flash movies ranking well… so I guess keyword density really rules here. If there’s only 2 words of text, and one of them is “the”… that’s MSN’s current definition of relevancy.

Also, note that many of these sites will attract link text containing “the”, either directly from the domain name, or because it makes sense to point to them that way (The Killers, The Veils, The Witchery Restaurant).

I’m particularly amused to see Napalm Death topping the list, with a framed site, no less. Classic, in every sense. Grindcore, man!

Google / Yahoo

Lots of interesting things going on here. I think the key point to pick out is more social than technical. You’d expect the results for these SE’s to be more link powered - and I think they are, but not quite in the way you might expect. I think the key is to consider the sociology of linking here - if you were to link to X site, what anchor text would you use? How is that organisation referred to?

In many cases, it’s the “the XYZ” - the Onion, the White House, the New York Times. www.guardian.co.uk is a clue, I feel - it’s a UK based newspaper for those that aren’t familiar with it, and it’s notable because “the” appears in neither the domain or the title. It’s known in the UK as “the Guardian” (or The Grauniad if you happen to be politically opposed to it), and that is frequently the anchor text used to point at it.

Yahoo ranks it #1 - possibly related to it having it’s own Yahoo Directory category, and Google has it fairly high at #5. Gimp.org seems to demonstrate the same effect, as GIMP is an acronym for The GNU Image Manipulation Program. See where this is going? Both engines rank this site highly too.

Google has some wrinkles all of it’s own though. Check the #3 result - see the title? “Education news & resources at the Times Higher Education Supplement” (bolding Google’s). That bolding effect is used to pick out words that match your query, so Google obviously “knows” that “the” can be considered as an acronym for Times Higher Education, and Technology Horizons in Education. A search for SEO shows the same behavior.

So, Google and Yahoo, concentrate more on anchor text than on-page elements, it seems, although it’s wise to remember that the 2 *can* go hand in hand. Creating a desire to link to a site using the terms you want to rank for is a handy skill.

Ask

Perhaps my favourite #1 result - The Simpsons. Overall, a bit of a puzzle. The Simpsons, and The AA (a UK based automotive recovery service, in this case) seem fairly obvious, and anchor text driven. The remainder of the SERP is more of a puzzle. The only thing that seems a good explanation is that Ask relies more on LSI (text surrounding a link) to determine the “meaning” of that link. Since “the” is fairly ubiquitous, that would tend to make their SERP more or less a list of the worlds English language websites, sorted by general authority, which is pretty much what #3 - #10 look like to me. If any Ask specialists have a better idea, I’d love to hear it - I’m always keen to learn.

So, overall Google and Yahoo love your anchor text, Ask loves what’s around your anchor text, and MSN loves your keyword stuffing, viagra, viagra, cialis, viagra, viagra, casino, online poker, viagra, poker, viagra, viagra, mesothelioma.

Can I have my money now?