The first feature is a small button in the shape of a pair of dice that you can add to the Google Toolbar in your browser.
The dice button enables you to get personalised recommendations from Google, based on your search history, your location and the search history of similar users. When you click on the dice, Google takes you to a site that you may find interesting based on your past searches. If you want another recommendation from Google, you simply need to click the dice again and you'll be directed to a new page. In fact, the search engine will give you up to 50 new sites per day that might be of interest.
The second feature works along the same lines and comes in the shape of a new recommendations tab that can be added to your personalised Google homepage. Simply click on "Add A Tab" on your Google Personalized Homepage, and type in "Recommendations" for the tab name (keep the "I'm feeling lucky" checkbox checked).
Google will then give you a page of recommendations that are updated daily. The difference to the recommendations button is that you'll be shown your information at a glance rather than one after another.
These features will only work if you are signed in as a search history user; and, to begin with, the recommendations you receive may not seem very impressive. But be patient, as Google's search recommendations will improve with time: once Google has collected enough information from your profile, the search engine will be able to highlight sites that are sure to match your interests.
The new feature from Google may sound familiar - and that's because this isn't the first recommendation system that uses implicit data collection. Internet retailer Amazon, for example, has provided personalised recommendations on products to its customers for years. However, the difference in this case is that Google doesn't want to sell anything to the user.
Even in Google's realm, however, this isn't the first recommendations system. "Recommended Google News", for example, works on a very similar mechanism: it shows news which may be of importance to you based on the stories you have viewed previously.
