10 new changes to Google algorithms

New features from GoogleYesterday, over on the Google Inside Search blog, Matt Cutts shared 10 recent changes to the Google search algorithms from the last few weeks.

As always these posts can get a bit technical, and anyone subscribed to the feed can just get it from the horses’ mouth. The goal of this post is to put the changes into clearer terms from a SEO perspective:

Translated search titles:
When searching with languages where limited web content is available, Google can translate the English-only results and display the translated titles directly below the English titles in the search results. This also translates the result automatically, thereby increasing the available web content for non-English searchers. If you were selling products that appealed to a global market, but hadn’t yet invested in translations/global site structure, this could drive fresh traffic to your sites/products.

Better Snippets:
Google’s mantra is always ‘content, content, + more content’, and now the snippet code is focusing on the page content vs. header/menu areas. Because of the way sites use keywords in the headers/menus, coding the snippets to seek out body content will result in more relevant text in search snippets.

Improved Google generated page titles:
When a page is lacking a title, Google has code in place to assign a title to the page using various signals. A key signal used is back-link anchor text pointing to the page. If a site has a ton of duplicate anchor text in the back-links, Google has found that putting less emphasis on those links creates a far more relevant title than previously. In this way the titles in the search results should be much less misleading.

Improved Russian auto-complete:
Languages are a constant headache for search engines, and new features like auto-complete can take a very long time to mature in languages outside of English. Recently the prediction system for auto-completed queries was improved to avoid overly long comparisons to the partial query to make auto-complete function much better in Russian, and closer to how well it works for English queries.

More information in application snippets:
Last week Google announced a new method of improved snippets for applications. The feature’s pretty technical and looks like an entire blog post is coming on just this topic. Here’s an example image that hopefully gives you a gist of how the snippets are giving details, like prices, ratings, and user reviews.

Example of application snippet from Google search results.

The feature has been very popular and Google recently added even more options that will elicit a full blog post soon here.

Less document relevance in Image searches:
If you look up search engine optimization in Wikipedia and look at the entry for Image search optimization you will note that there’s really nothing to say about SEO tactics towards images. This hasn’t been true, there are signals that Google has to look for when deciding what image to show for a particular keyword.
Previously, an image referenced in PDF or other searchable documents multiple times would get higher placement in the results. Google has done away with this signal as it wasn’t giving improved results and could easily be abused. *Innocent whistling*

Higher ranking signals on fresh content:
Consider if you will, how Google would look if they never gave new sites/fresh content a shot at the top, or a moment in the limelight? By default most ratings systems will show you the ‘best of the most recent’ by default just to avoid older content dominating the results. As a person on the phones taking SEO leads I can tell you there’s always been a ’10 mins of fame’ situation on Google where the explainable happens in the search results with fresh sites/content, only to return to normal later on when the dust settles. Google claims the recent change impacts roughly 35% of total search traffic which could be a significant boost for sites that take the time to publish fresh content, or for new sites looking for a chance to be seen.

Improved official page detection:
We’ve blogged recently about the importance of the rel=author attributes, tying your content to a G+ profile, and completing the circle with a back-link from the profile to your site. Google’s added even more methods to establish ‘offical’ pages and is continuing to give ‘official’ pages higher rankings on searches where authority is important. If you missed our article on this topic from last week, here’s the link.

Better date specific results:
The date a page is discovered may not always be the date the information is published. Google has the difficult task of sorting out the ‘date’ relevance for search results, and they keep improving on this where possible. A good example would be using duplicate matches to avoid showing you a 3 year old article that was posted two days ago if you specify that you only want results from say ‘last week’.

Enhanced prediction for non-Latin characters:
You’d think it’s hard enough to get a predictive query straight when the character set is limited to Latin, and you’d be right. When it takes several keystrokes to complete a single character in non-Latin, a service like Google’s auto-complete would be hard pressed to know when to start guessing. Previous to this update predictions in Russian, Arabic, and Hebrew were giving gibberish results as the user was forming characters.

These are 10 changes out of 500+ made so far this year. We try to document the most important changes for you but there’s lots of times where Google can’t release info because of exploits/cheating. When that happens you’ll see us chime in with experiments and our personal experience when we can. So while I’d normally suggest folks interested in this topic subscribe to the inside search blog, we know that you’ll only be getting part of the story by doing so. 😉