Six Thoughts from Google’s SEO Report Card

Posted by Fernando Chavez on March 16th, 2010

Two weeks ago on the Webmaster Central Blog Google released an SEO Report Card that they had distributed internally to Google product teams. They gave their teams some ideas about how they could improve their product pages from an SEO standpoint. After reading through it briefly, I thought it might be a good idea to summarize some of the key points for SEOs. There were a handful of concepts touched upon that fully justify recommendations that many of us routinely make. Here are the quotes from the document that I think are somewhat enlightening along with my analysis of what they mean.

Issue #1 – Title Tags

Google Quote (page 4):

Use the space provided – Most major search engines display approximately 60 characters from a page’s title tag in the title of a search result. These 60 characters are an opportunity to tell both users and search engines what the focus of the page is. There’s no need to go past this many characters, as most search engines will display ellipses ( … ) after this limit. Also, search engines may give less weight to words after a certain point.

Analysis:

This snippet might actually be the list revealing of all the quotes I will be discussing here. The problem is that they expanded their recommendation to include other search engines. I’ve seen 70 characters of a Title tag displayed in Google SERPs, so that’s the guideline I typically use. Google says that “there’s no need” to go beyond that character limit, but I suspect that they were talking about click through rates and not SEO. The next sentence says that they “may” give less weight to words after a certain point, which implies that you can get some benefit from words beyond their 60 character recommendation.

I see no reason to be bound by a hard length restriction. Remember that if your pages get scraped the full Title tag could end up being the anchor text of your inbound link (i.e. not just the truncated version). I won’t cut it really short for that reason. However, I do try to keep the length fairly reasonable. In a nutshell, I keep my Title tags under 12 words, but focus heavily on getting the top page keywords into the first 70 characters. Click here for more information about how to optimize Title tags.

Issue #2 – Meta Description Tags

Google Quote (page 9):

Take control of your snippet – The snippets above are great examples of why to use description meta tags. Although description meta tags don’t count in Google’s ranking (nor do keyword meta tags), the text contained in them is sometimes used in the snippet of search results. In the examples above, Google may have chosen to display a description meta tag, but as it didn’t exist, the alternative option for the snippet was the page’s content. Since the main page of Google Sky doesn’t have much text on it, the only text Google could find was the page’s navigational and boilerplate text. The Google Gadgets main page is light on text as well, so the legal disclaimer at the bottom of the page was chosen. Neither of these snippets are attractive to search users. Since you have the chance to take control of your snippet with a description meta tag, do it!

Analysis:

The underlines within the quote are Google’s, not mine. They wanted to be very clear to their product teams that Meta Description tags do not count in Google’s ranking calculation. I’m sure that most informed people knew that Meta Keywords tags are ignored, but prevailing opinion on Meta Description is still all over the place (myself included). I choose to optimize them to influence the snippets that appear in search results. However, I also think that in certain instances Descriptions can have an indirect influence on rankings as well. If spammers scrape your pages/search results often, then the Description may appear as part of their scrape along with a link to your site. In other words, you can possibly influence your off-page relevance with a well-optimized Meta Description. They may not have a direct effect on rankings, but there’s no reason you shouldn’t try to use keywords in your Descriptions.

Issue #3 – 301 vs. 302 Redirects

Google Quote (page 27):

Don’t use 302s when 301s would be more appropriate – When a 302 (temporary redirect) is used on a URL, the search engine is told that the destination URL is a temporary one and that the search engine should keep track of both URLs (the one with the redirect on it and the destination). In the example above, users will properly reach the canonical version through the 302, which is good; however, search engines won’t transfer the reputation from www.google.com/friendconnect to the canonical version through a 302, only through a 301. With a 302, the search engine says, “This destination is just temporary, so I’ll hold the reputation at www.google.com/friendconnect in case it’s used again in the future.” Avoid this situation by using a 301 to consolidate reputation.

Analysis:

I don’t recall Google summarizing the 301 vs. 302 redirect issue so clearly before, but then again I don’t read their blogs religiously. Essentially, a 302 redirect will not allow PageRank (or reputation as they referred to it in the Report Card) to consolidate at the destination URL. I’ve always stated this as essentially fact, but now we know for sure.

Issue #4 – URL Consolidation and Redirects

Google Quote (page 19):

Consolidate URLs – Although the highlighted results above all go to the main page of the service, users should never have to ask, “Which one do I click and how are these different?” Choosing one version of a URL and consolidating the others with 301 redirects or the canonical element makes everything easier for the search engine and for users.

Prevent dilution of reputation – If the same content is accessible through multiple URLs, this could cause duplicate content. This content may rank worse because its reputation is spread over multiple URLs. Consolidation of the URLs, as mentioned above, resolves this.

Allow crawling for 301 redirects – A URL must allow crawling before the search engine will recognize that it redirects. www.feedburner.com’s robot.txt file currently disallows search engines from crawling pages on the domain, so its 301 is never seen. In some cases, a site might choose to allow crawling for just its main page and block crawling on the rest of the site.

Google Quote (page 20):

Google products’ URLs take many different forms. Most larger products use a subdomain, while smaller ones usually use a directory form, with or without a slash (note that www.google.com/product is considered a different URL than www.google.com/product/ by search engines). With so many products and different forms of URLs, users face a daunting task knowing which URL format to use for which product. Such varied URL behavior can lead to 404 pages (users following broken links or directly typing the URL in) and split reputation between multiple URLs, hurting the ranking of content. The following recommendations can help with this:

  • choose the easiest to remember form of the URL as the canonical (likely product.google.com/)
  • be consistent with this canonical form across all products
  • think of the most common URL forms visitors may try and 301 redirect these to the preferred/canonical URL or use the rel=”canonical” link element if you cannot redirect

Analysis:

As SEOs we’ve all dealt with this issue to some degree. I don’t think that they’ve said anything too insightful here. Basically, you should choose a canonical that is easy to remember and 301 redirect the others to the canonical to prevent PageRank/reputation dilution. However, there are two sentences worth noting.

The first is from page 19 in the paragraph labeled “Allow crawling for 301 redirects”. I’ve had several clients that changed their URLs and decided to disallow their old versions to prevent the “duplicate” content from being indexed. But they neglected to consider that this would effectively eliminate any link relevance that may be going to the old page from external links. Disallowing the page prevents Google from following the links and consolidating the link relevance. The same issue could theoretically arise from non-preferred forms of URLs. A robots.txt disallow in that situation would also prevent the reputation consolidation.

The second noteworthy sentence is from the last bullet in the page 20 quote. They suggest using the rel=”canonical” link element “if you cannot redirect”. This implies to me that the rel=”canonical” is not as reliable as a 301 redirect. I rarely suggest using the rel=”canonical” because I always assumed that to be the case. Reading between the lines here confirms that for me.

Issue #5 – Logo Image Linking

Google Quote (page 39):

Logo image link destination
Logo images on our products’ pages play an important role in both users and search engines’ navigation of our sites (most logos link to the main page). Making sure these link to the main page’s canonical URL can improve the flow of internal reputation and prevent 404s for users.

Google Quote (page 42):

Logo image alt text
Using brief and descriptive alt text for our linked logo images helps search engines know more about our products’ homepages. Also, the search engine learns more about the image itself, as do users who don’t load images due to accessibility or device reasons.

Google Quote (page 43):

Good! – These product main pages have logo images with alt text that includes the product’s name. Since many of our logos are used as home/main page links, the alt text of these images is basically treated as anchor text by search engines.

Be descriptive – Like many optimizations in this document, using brief, descriptive text in the alt attribute-which is treated like anchor text for linked images-is best. Above, it’s good that “Google” is present, but this doesn’t tell the search engine which Google product is being linked to.

Google Quote (page 44):

Remember that alt text on an image link is essentially anchor text – You wouldn’t create a text link to Google Analytics with no anchor text (e.g. <a href=”http://www.google.com/analytics/”></a>), so the same applies to image links.

Analysis:

I’ll be the first to admit that I abused logo image ALT attributes on many sites that I’ve worked on in the past. A couple years back Google seemed to really clamp down on their effectiveness, so I backed off that tactic completely. I suspect they discounted their value by changing the way they interpret outbound links to pages that have both a text and image link. Prior to the change, logo images were treated as the anchor to the home page because they were generally the first link. I confirmed through testing that Google now uses the first text link to a page as the anchor text. The only time an ALT attribute is used is if the page is only linked to with images. The following quote from page 43 would seem to support this viewpoint: “Since many of our logos are used as home/main page links, the alt text of these images is basically treated as anchor text by search engines.”

I find it interesting that Google would suggest using ALT attributes in this way. I got the sense that they didn’t want people to abuse it. However, they do say to keep the ALT attributes brief and descriptive, so they are not exactly condoning keyword stuffing. Also, I think they know that they won’t inadvertently trigger filtering of any Google pages since all of their products presumably have a significant amount of external relevance to validate the keywords that would be used in the ALT attribute.

Issue #6 – Triggering Sitelinks

Google Quote (page 27):

Webmasters can’t choose when sitelinks are shown; however, they can optimize their site’s organization and internal linking to improve their chances. The following can help:

  • use a hierarchical site structure
  • use descriptive anchor text for links pointing to internal pages
  • avoid deep nesting of content behind many subdirectories

These optimizations assist both search engines and visitors as they navigate your site.

Analysis:

Okay, this actually tells us nothing in my opinion, but it is more than they’ve stated in the Google Webmaster Tools Help section. What they are essentially saying is build a user friendly and SEO friendly site with good anchor text and hope for the best. That’s pretty much what I suggest for my clients, so I guess technically this is a Google confirmation of that recommendation.

Share and Enjoy:
  • Print
  • Digg
  • Sphinn
  • del.icio.us
  • Facebook
  • Mixx
  • Google Bookmarks
  • Blogplay

3 Responses to “Six Thoughts from Google’s SEO Report Card”

  1. I came to your site from the Youmoz article, and then spotted this piece. Thank you for the analysis. With the ALT text for images, I have always been more concerned with making my site accessible, rather than keyword stuffing. I guess then that the practice of trying to make your pages usable and accessible helps with the SEO.

  2. Is there a reason you don’t allow rss subscription to email? I don’t use any one specific reader.

    Great post!

  3. Fernando Chavez says:

    Thanks for the comment. Actually, the lack of an email option for RSS subscription was not by design. I had simply neglected to think about it. I’ll look into how to implement that ability.