₦airaland Forum

The point of this session was to let the audience pose questions to Google's chief search quality engineer. The particularly valuable queries from these first forty-five minutes included a question from Patrick Sexton about Google's much-disputed guidelines.

Webmasters and SEOs often agonize over the brevity of the guidelines, wishing that they were both more detailed and provided real examples of both good and bad practices and Matt agreed that the guidelines do not allow for much interpretation.

Another interesting question came up about click-through and what effect is has on the rankings: Matt's response sounded as though he was trying to indicate that Google does not take clicks into account, as that sort of user-data would create a certain amount of clutter. I know many of us have wondered about click-through in the past, and Matt's response neither confirmed nor denied very much. In fact, this particular question may have left the SMX audience members even more interested in click-through than they were before.

Obviously, after the Greatest Living American scheme, the Googlebomb algorithm came up. The short version of Matt's reply is that there has been no human intervention in the algorithm; however, it does not run at every minute of every day and thus a “bombing” may appear to work for a while and then be crushed.

A later question related to a site-review panel that took place at PubCon in Las Vegas last year. During the site-review, Matt looked up how many domains a certain audience member owned in addition to the site that was being reviewed. This move was controversial in that people wanted to know why that was any of Google's – or Matt's – business.

Matt's response probably didn't calm many of the people who were unhappy with the PubCon incident, but he did give us a good explanation of why Google might look up a webmaster's existing domains, and what they do with that information. For example, if Google finds that a site that appears to be doing something wrong is run by a small outfit, they may take different actions than they would regarding a site whose owner deals in thousands of different domains. However, Matt made it clear that the number or type of domains a person owns does not directly impact the rankings of any one of those domains.

Other questions revolved around human-based search engines, the quality and accuracy of Google's image search and, predictably, the value one obtains from being listed in the Supplemental index.


Rebecca:
I dashed into the conference center at about 9:05 am, just in time to catch the tail end of Danny Sullivan's welcome speech. He thanked the Academy, wiped away a tear, all that good stuff. It was nice to have an official conference “Welcome”--I can already tell that SMX is striving to be a more personal conference than some of the others I've attended.

Now that we've gotten settled in, Matt Cutts joins Danny at the front of the room for a little “You&A” action. After some ribbing from Matt that he's dressed too stuffy, Danny shed like eight layers of clothing (I'm assuming he left his snowsuit at home) and chatted with Matt looking like a man-child in a t-shirt, shorts, and checkered sneakers. Casually dressed speakers and moderators? Oh man, this conference seems too good to be true.

Anyway, let's get down to business with a few things I learned from the You&A session:
1. Google really is watching you. An audience member asked how confident Google is in knowing the location of their users, and Matt's response was that they have an idea of where about 90% of their searches come from. He thinks that, overall, this is a good thing, and he gave an example of someone with a plumbing emergency searching for “plumber” instead of “seattle plumber” and still being able to find a relevant result.
2. Google defends Wikipedia's dominance in the SERPs. Someone asked why Google wants to have 10,000 of Wikipedia's babies (okay, the question wasn't exactly phrased that way, but you get the idea: why so much love for Wikipedia?), and Matt argued that SMX attendees, by definition, aren't “regular” users and searchers. Matt's mom would love Wikipedia (she probably also loves pinching Matt's cheeks and telling him to call more often). He then insisted that “regular” users like Wikipedia a lot. [I wonder if this is a “Chicken or the Egg” scenario—do regular users like Wikipedia because it's informative, and that's why it ranks so highly, or because it ranks so highly, meaning users get used to seeing it in the SERPs?] Apparently there is an awareness at Google that Wikipedia is a “good result most of the time, but not as accurate for more advanced searches.”
3. Google swears that the Google bomb algorithm is completely automatic. Jonah Stein asked about the Stephen Colbert “Greatest Living American” Google bomb that was recently defused, and Matt said he was pretty sure that the algorithm hadn't changed from a few months ago to after it fixed the Colbert bomb. It doesn't run every day, only about every 2 to 4 months. Google reserves the right, however, to take manual action on spam and handle a situation accordingly.
Danny then asked about how Bush was removed for “miserable failure,” then ranked again when the word “failure” was used on the site, and then went away again in about a week's span. Matt said that basically the algorithm analyzes queries and results and looks for a valid match. Bush was removed for “miserable failure” because the site didn't match the query, but he showed up in the SERPs when the word “failure” was used, thus making the site a potentially valid match. He was later removed again when the algorithm decided that the site wasn't a terribly relevant match, after all.

Overall, I really enjoyed the You&A session. Matt did a good job of answering the audience's questions as best and as truthfully as he could, and the attendees themselves did a good job of peppering Matt with some pretty compelling questions.

Duplicate Content Summit

Jane:
Next up was the Duplicate Content Summit, which has drawn some criticism as being too basic for this style of show. However, there were some tips to come out of this session:
• Eytan Seidman pointed out that it's okay to have similar content on pages that are targeted at specific audiences, such as when your site has a .com and a .co.uk version. However, the content should be similar, not exactly the same. If the content is exactly the same, you should only have one version.
• Eytan also pointed out that duplicate content can create problems when you're trying to gain organic links as well. Knowing which page to link to on your site should be “totally unambiguous” for users.
• To avoid your content being stolen, you can block unknown IPs from crawling your site, but you can run into problems here when you end up blocking legitimate users.

Peter Linsley from Ask mentioned some editorial ways in which publishers can cut down on content theft. Although many sites will reproduce content no matter what, those who hand pick articles may be turned off by posts that mention specific circumstances, people and events that are unique to the author.

Amit Kumar from Yahoo made the point that identical content in different languages won't be detected as duplicate, and neither will “boilerplate” features, such as common site elements, navigation, legal disclaimers, etc. On the surface, these two issues seem quite obvious; however, it's true that there really is a lot of duplicate content on any given site. Headers, footers, navigation, and sidebars usually appear on multiple pages, and although we'd never assume that the search engines are considering that duplicated, it's good to know for sure. As for different languages being treated as unique content, it would be very unfair for the engines to treat identical content in different languages as duplicated, since the majority of the population cannot speak multiple languages fluently. Again, it's nice to hear that the engines aren't wielding their linguistic advantages over us! Google's Vanessa Fox addressed the question of combining very similar content into one page or radically modifying similar content in order to cut down identical pages.


Visit This: getloadedinthepark