Search engine Yandex this week added personalization capabilities for Eastern European users’ search results. It analyses their online behavior including their search history, clicks on search results, and language preferences for its suggestions.
Kaliningrad is the name of the latest edition of Yandex’ personalized search engine. It uses that information to make suggestions and rank search results individually tailored for each user, showing book lovers that do a search on Harry Potter links related to the books, while those who prefer movies get film-oriented link fare.
Semantic markup didn’t play a role in the development of the technology, Yandex technical product manager and developer advocate Alexander Shubin says. But it can be applied for future enhancements, he notes. The new personalization reportedly leverages Yandex’ machine-learning-based query and search results algorithms “Spectrum” and “MatrixNet” to train the results to users’ requirements.
That said, Yandex has been diving deeper into semantic web waters. Beyond taking advantage of sites using schema.org markup to improve the display of search results, Shubin provides this update: “We enhanced our markup validator to understand all the markup (Open Graph, schema.org, RDFa, microformats). It is universal now (as Google’s or Bing’s instruments).”
Also on the docket: translating all its documentation about markup usage into English, he says.
Yandex has had some other interesting coverage in the media recently. Last week, for example, Bloomberg posted a story with a headline that read, “Yandex Dominating Google as Ads Boost Outlook,” noting in the piece that record online ad sales this quarter might be spurred on prospects of its domination over Google in the Russian search market.
Yandex, it says, got 90 percent of revenue from text-based advertising in the last quarter, and has averaged a 60.4 percent share of the Russian Web search market since September, compared with 26.5 percent for Google. “Yandex will post fourth-quarter revenue of 8.9 billion rubles ($288 million), according to the median of three estimates compiled by Bloomberg,” the article notes.
No tears need to be shed for Google, though: comScore this week released its monthly comScore qSearch analysis of the U.S. search marketplace, finding that Google Sites led the explicit core search market in November, with 67 percent of search queries conducted. Nearly 17 billion explicit core searches were conducted in November, with Google Sites ranking first with 11.4 billion, it notes. In November, 69.4 percent of searches carried organic search results from Google.
Google this week also announced its Data Highlighter for helping webmasters get into the structured data act (read more about that here.)
- Automatic Hashtags & Machine Learning: The New Google+
- Gmail, Meet JSON-LD
- Talking the Talk -- And Walking The Walk -- About The Beauty Of Search At Google I/O
- Bing Gets More Social with Facebook Likes