Danny Sullivan introduces this session as the best session to be at now. He said the other sessions are for those who are not serious about SEO. In NY they did the first SES indexing summit and they got some weather reports out of it. Danny said he wants to talk about two things; (1) standards with redirection and (2) standards on how search engine create titles and descriptions in SERPs.
Topic One:
Redirect Issues
Tim Mayer from Yahoo! up first. He said people were able to use some redirects to hi-jack sites. He put up the Yahoo! Redirect Handling Rules (covered in many past conferences, will link to document later). Just to touch on how they handle it.
Danny then goes into explaining what this issue of redirect hi-jacking is, http://blog.searchenginewatch.com/blog/050801-130330.
Rahul Lahiri from Ask Jeeves was next up. He wanted to add, that Ask Jeeves had been looking at the problem and did not see a major issue with redirects at Ask Jeeves. He said if you run into these problems, let Ask know. He said again that people don't use the standards or guidelines. Ask realizes this and they make assumptions.
Matt Cutts was next up and he asks the audience a few questions. He then brings up the history of the Indexing Summit where Danny guilt-tripped the search engines to work together. It encourages Google Sitemaps for a paid inclusion, without the money, and any search engine can use the sitemaps standard. The robots.txt is from ages ago, and a good standard, nofollow attribute. He then moves into the redirect issue, like Danny did. A year ago, Google picked the URL with the highest PageRank. That worked very well most of the time but not always. So search engines had to pick the best URL, and PR wasn't the best way all the time. Right now, they are open to doing what Yahoo! does, just indexing the target URL. BUT, you don't have the pretty URLs. He said, Google will have lots of beer tonight at the Google Dance. But also there will be an event "Meet the Engineers" where you can discuss this with them. State at Google right now, is that Google is not aware of any hi-jackings, they need to tell Google about them.
Q: When is a 301 redirect spam? A: Matt said a 301 is NOT spam when its a misspelling, redirected to the main site. A 301 spam redirect would be a doorway page, which are then redirected to the root page (i.e. cloaked pages).
Q: Do we need to use the robots.txt to block PageRank to them to be careful about spam? A: Matt said just 301 redirect those. You do not need to exclude them in the robots.txt
Q: Is the redirect issue solved? A: Matt said always go with the destination to be careful, but his engineers said, but we fixed all the issues with this. So he wants to show these engineers that there is still an issue.
So Danny then summarizes again, basically, all the search engines are doing different things. But this session is about the search engines working together. Randfish, said he is not cool with them all doing different things. One person stood up and said, if both are my sites, then allow me to denote so in the meta tags or in the robots.txt with the Yahoo! method.
Danny summarizes, we have different standards, a good chunk of us like how Yahoo is handling it now, but a good chunk wants to be able to specify the URL that should be listed in the search engine results. Clarity; give the Webmaster a way to define the URL they want anything listed.
DavidN looks at Google.com results for san francisco giants and shows that both redirected URLs are listed. He says that I can throw all 10 of my domains there and dominate the SERPs. Yahoo! has the same issue. This will be fixed, Danny said.
Nacho adds, once you figure out the domain you want listed, what information (link pop, age, etc.) flows through the the destination URL.
Some guy in the back adds there are legal issues why I might want to remove the URL from the search engines. So true.
Topic Two:
Title & Descriptions Management
Google dynamically created titles and then used some ODP directory titles sometimes, Yahoo! used the Y! directory and then stopped --- same with the description. Danny is being very very funny today, he goes through the various ways it is handled and he chatters incredibly fast to make a point. He brings up a forum thread, thread id=5759, the "Proposed Search Engine Standard for Titles & Descriptions" thread (see the poll there). Issue is, many pages have no title - so what do you do there? What do you do when a "created by adobe go live" in the title? What do you do when you have a CMS that inserts "Title Goes Here"? This is what should be listed in the SERPs? The search engines do not want it. The searchers do not want it. Do you?
Tim Mayer from Yahoo up first. The current behavior at Yahoo: (1) Feed titles and descriptions, (2) Yahoo! Directory descriptions, (3) Best match to query between contextual abstracts (on page) and meta title and meta description (on page), (4) If we cannot generate anything they use the ODP (implemented 1.5 years ago) or anchor text for the title. Abstract challenges; sites don't provide titles, sites use the same titles and descriptions for every page, meta tag T&Ds are not query specific, many sites over optimize their T&Ds and it does not accurately describe the page's content and different types of abstracts are appropriate for different types of queries (e.g. navigation versus informational). Going forward, Yahoo! will take all these different inputs and decide which is best based on user experience.
Rahul Lahiri from Ask Jeeves is up. If the page has a title, they use that one. If it does not, that is when they look elsewhere. They look at ODP, then they create a title from the text in the page (query words and content). The description, they look at ODP description and if the query term is in there, they will use the ODP description. If not, they will use the content of the page to build a description. Ask Jeeves prefers a single snippet over a longer one (like Google & Yahoo) if the query term is short, if its a longer query term they will try to match all the keywords in one or more snippets for the description.
Matt Cutts from Google is now up. He says all search engines want to show the most relevant listings title and snippets in the SERPs. When Google added the cache, people loved how you highlighted the keywords in that cache page. Matt hates "your browser does not support frames" as a snippet, so they will do what ever they can do avoid showing those snippets. A real issue with meta tags, you can not trust what people say. They have been doing a lot more experiments with the snippets. He said they tried and try things every day and test user experience. He said it will be time until you see standardization.
Forum discussion at Search Engine Watch Forums.
