At the heart of the technology, which has been crystallized from testing various search formulas on 50 million users, is Yandex’s new ability to analyze a user’s prior online search behavior and language preferences, and rank search results personalized for each individual user, offering one special tips even as he or she enters a query.
According to Denis Rogachevsky, the lead developer for Yandex search, tailoring search results to the time-tested online predilections of an individual “improves the quality of search results for about 75-80% of search queries.”
The Kaliningrad platform is fully available to Yandex users in Russia, Ukraine, Belarus, and Kazakhstan, and is undergoing a test run in Turkey. Applying the new technology has already enabled a 37% increase in searchers’ clicks on the top search result, Yandex claims. The personalized search function is active by default; however, a user can choose to disable it on his settings page.
‘Kaliningrad’ is an evolution of Yandex’s previous and more narrowly focused search platform, ‘Reykjavik,’ which the company introduced last year to boost search relevance for English-language queries.