Google Explains Site Penalties

This is Part 2 in an 5-part series of posts on a recent SEMNE session that gave the audience access to a Google staffer who gave useful tidbits of “insider” knowledge. Dan Crow, Program Manager of Crawl Systems, was the guest speaker to an audience of about 100 who ranged from SEO novices to industry experts.

Here’s a summary of the answers Dan could give “without getting fired:”

  • Supplemental Search Results – Dan explained that supplemental results are based on pages that “don’t change often”, get crawled only every few months and only show when there are not enough relevant pages in the results to display. Dan explained the supplemental results aren’t actually an active penalty against your site, but rather they’re an artifact of Google’s old IT infrastructure, which had to make some pages “low priority” to handle the enormous task of indexing the Web. Google is actually working on doing away with Supplemental Results, but no date was given.
  • CSS – The current crawling system at Google partially understands CSS, and they’re working to get their systems to understand it better. This means that Google can understand what information is being shown at the top of the page, even in a CSS-based layout. It also means that CSS spamming will soon be obsolete (something we do not practice at Catalyst anyway.)
  • Duplicates – Google systems find it hard to discern what file is actually a duplicate; however they are working to better clarify this process because they know that webmasters have had issues.
  • Flash fixes - As mentioned in a previous post, Google tries not to misinterpret Flash workarounds like SWFObject as spam, but Dan characterized it as a “dangerous” technique and couldn’t guarantee that penalties wouldn’t be applied, even for innocent applications.
  • Too much navigation – Pages with too many outbound-to-offsite links can be interpreted as a link farm, so it’s best to moderate your outbound links.

2 Comments

  1. Posted August 2, 2007 at 9:23 am | Permalink

    The partially understanding CSS could be somewhat dangerous. How much do they really understand? For instance - I don’t classes and ID’s all over the place, does it understand inheritance and the cascade?

    I would be interested to see how much it really understands of any given CSS implementation? Inline? In the head? External sheet?

    Oh, and what is ‘Flask’ - is that anything like ‘Flash’ - he.

  2. Posted August 2, 2007 at 9:32 am | Permalink

    Nate-
    Good point. I actually think the reality is that they (Google) don’t “understand” CSS at all, because it’s too complex. I think they may be able to flag things like “visibility:none” for HUMAN editors to manually check, but I doubt the AI is good enough parse out the good from the bad.

    Also, FLASK is what you get when your copy editor doesn’t proof well enough.

    -Francis, Copy editor for search Matters. ;)

One Trackback

  1. [...] more news from the session, along with some options for how to circumvent the issues raised. On toPart 2… Share and Enjoy: These icons link to social bookmarking sites where readers can share and [...]

Post a Comment

Your email is never published nor shared. Required fields are marked *

*
*