Search Engine Q&A On Links

Aug 10, 2005 • 8:44 pm | comments (6) by twitter Google+ | Filed Under Search Engine Strategies 2005 San Jose
 

We got Tim Mayer from Yahoo, Matt Cutts from Google and Kaushal Kurapti from Ask Jeeves on this single panel. No one knows what is going to happen now. Danny is modding up.

Ask was up first, he gives his company info slide, yada yada yada. He briefly reviews how the link analysis works here. Page A links to page b and c, that is a recommendation from page at to page b and c. More links, better you are, in short. Then he discusses the Ask Jeeves approach, the whole community & hug/authority thing....He then describes that all links are not equal. Be careful of reciprocal links and purchasing links. Avoid link farms, cloaked pages, invisible or hidden links and links by images - text based links can be understood, but not image links. Become and authority on a specific subject. Focus on your business and content and the rest will follow. Blog links do not mean too much from the blogrolls to Mr. Jeeves.

Yahoo! now. Tim announces a new product named Site Explorer, http://sitexplorer.search.yahoo.com/ where you can get your linkage data. It is a place for people to go to see which pages Yahoo indexed and to let Yahoo know about URLs Yahoo has not found as of yet (submit URL or URLs). He showed an example, you basically type in a URL into it (this is also supported via an API, good good), then you hit explore URl and it spits out the number of pages found in Yahoo Index and also shows you the number of inlinks. You can sort pages by "depth" how deep pages are and you can also submit URLs here. You can also quickly export the results to TSV format. Links have been very popular, he said Yahoo! has been moving towards the social community aspect and probably will go in that direction with link pop. He said, create natural links, make it look natural.

Google up now. Matt goes up with no presentation. He said Ask Jeeves covered most of the basics, and basically recapped it quickly. If Matt was someone starting out, he would get links from...HousingMaps.com shows you craigs list stuff on Google Maps. It has been around less then a few months and it got tons of links quickly. Also think about useful services to offer (validation tools for rss feeds, etc.) Matt ran across one that makes signs dynamically, he built his own "watch out for falling spam." If you don't have to ask for links, that is awesome. Add one new page of content everyday. Syndicate my content. Matt Cutts started a blog today, www.mattcutts.com. Make a community, reviews, forums, and so on. One of the best ways is to think outside of the box. For example, one of the seo contests (cant spell it).

Q & A:

Q: When I do comparison across the engines of who is linking to me, I see differences between the two. His answer he got was that the link command isn't full accurate. A: Matt said that they used to show only important backlinks. But then someone suggested to show random samples of backlinks. They have never shown all backlinks. They do have all the backlinks at Google but they do not show them all. Tim shows more backlinks then Google, they do not show all links but a more comprehensive link. The new system will "be very comprehensive" he wouldn't say it is every link. Kaushal said you would see a difference, because different engines filter spam and dups differently. Also not everything is exposed and each engine takes a slightly different approach.

Q: I have a client that has a great site, lots of links but the anchor text being used throughout the web is the same. A: Matt said that is very unnatural. Most natural links are not 100% one exact phrase to the site. It won't hurt you, there is no OOP, but all the links might be devalued. Tim agreed with Matt on it being unnatural.

Q: Reciprocal links; we have them now, we have plans to do more, what should I do? There are 20 of them links. A: Matt said here is my rule of thumb, pretend you are my competitor, what would they think of it? Plenty of people have reciprocal links but if its excessive, then you need to be careful. Editorial given links and independent links are best. Danny then asked 4 people in the audience to point to each other and then asked several to point at each other. Matt said if you go into "graph theory" you have a "clique", that clique is when everyone in a network is pointing at each other, that is not natural.

Q: How do you know when too much is too much? A: Tim said that is the hard question. It is all about "intent". Matt adds that if you take this to random 5 people outside of the SEO community, they would agree.

Q: Do none clickable links count as back links? A: Matt said he has never been asked that, and he can see it both ways. Google has the code that they can flip the switch either way - but use the hyper link. Tim said its best to get the hyper link. Ask Jeeves said the same thing as Tim.

Q: PageRank; is it important or not? With the rel="nofollow" thing, if I cared about my PR, I would use nofollow on all my links to keep the PR within my sites. What are your thoughts? And is there a correlation between PR and number of pages indexed? A: Most Webmasters say PR is not as important. Google has always said there are many variables in the algo and they keep evolving. Very few people outside of a search engine can say exactly how valuable a specific link is from a page. In Matt's opinion, the nofollow has been a very valuable thing for the search engines. It gives the Webmaster the ability to say if I vouch for this link. So now we have this new type of data the search engines can use, he said its being used very responsibly. If you can authenticate or trust a comment poster, then there is no reason to use the nofollow. Tim just repeats what Matt said.

Q: Query strings at the end or URLs, when does that make it a problem for engines? A: Matt said 3 or more, its not great, but GoogleBot sometime is smart. Don't use id= in it, and if you have numeric parameters, dont go above 4 numbers. Kaushal agrees with Matt, but a limited set of parameters are ok. Tim adds that if you have inbound links to those dynamic URLs, they will more likely crawl it. Yahoo! is less considered with duplicate issues.

Q: We build directories, what is the proper way to link out from those directories? A: Matt said you love something, you need to set it free. It looks very weird to have tons of ibl and no outbound links. So static link those out.

Ok here is some fun. Danny was showing examples of high PR sites and Danny joked about sometimes Google goes up to a PR11. Tim Mayer then said, "Matt, that is only for advertisers." Good one.

Someone asked why Yahoo!'s PR went down from PR10 to 9, why did that happen? Tim said because of the index size announcement. Matt said we don't do that, we don't give ourselves (google) a 10 because we give a 10. He said, we are too busy for that.

Tired...Stopping...Good Night.

Forum discussion at SEW Forums.

Previous story: Usability Clinic
 

Comments:

Pradeep Chand

08/11/2005 05:19 am

hi, Thanks to SeoRoundTable as it is often that we get a chance to listen from the Search Engine giants from a single panel. But the interview was a little letdown because I could not find any exclusive information about any of those Search Engines. The most important and fundamental point that I could derive was to be or appear "natural". Rest is same as what we find all over the web.

Barry Schwartz

08/11/2005 05:40 am

You wont find exclusive information from the engines on this panel. I am sorry.

Aaron Rubin

08/11/2005 03:20 pm

Is http://sitexplorer.search.yahoo.com the correct address? It's bringing up the standard advanced search page.

Barry Schwartz

08/11/2005 03:27 pm

It is the right address, but it is not live yet.

Joe Griffin

08/20/2005 08:39 am

Matt Cutts new Blog is great. I hope it grows reguraly!

Uncover China

10/20/2005 05:09 am

I'm a big fan of his blog also, I'm surprised Google haven't shut it down..

blog comments powered by Disqus