Applying PageRank to other Searches
http://www.wsj.com/articles/google-discloses-more-search-data-to-woo-retailers-1445384100
The article primarily focuses on Google’s new method of utilizing its search data. Google today released the “Shopping Insights tool.” This feature focuses primarily on Google Shopping, Google’s product search engine. Google has started collecting data using users searches based on location. Google can then offer this data to retailers so retailers are aware of what users are searching for in their area. The system has been implemented in more than 16,000 U.S. cities and towns. The initial version only covers over 5,000 products on sale through however over time, Google will add more products and update the data more often. The rankings offering varying data in demands based on location. For example, in Berkeley, California, the searches for Star Wars costumes is three times as high as for Minion outfits, while in Madison, Wisconsin, the opposite is true.
Google shopping is a conventional search engine liked described in lecture and could be using PageRank. Google shopping ranks products based on hubs and authorities. In this case the authorities being sites that are offering a product and the hubs being product listings. Here Google is offering a way to interpret and monetize the search data. However, it is apparent here that Google is using other methods of ranking the results. As mention in the textbook as well as the article, Google utilizes links, text, and usage data to effectively provide results.
Google’s search algorithms, specifically PageRank, have been the backbone of the company for years and has been adapted to many other uses beyond product design. Right now, smartphone users discover apps by searching the app store, by seeing ads, or from input from other users. Now, Google is designing a tool to apply a PageRank to app searches. Instead of using links as the voting system, Google plans to index apps by recording data on the users in-app behavior. This technology is enhanced by Android’s new “Now on Tap,” where you can tap and hold the home button for data without having to leave your current app. For example, if a friend emails you about seeing a movie, you can use Google Now without leaving the app. This will allow you to see ratings, watch a trailer, or buy tickets and then immediately transition back to the app you were using.
This is an application of the PageRank algorithm we learned about in class. Google is treating app transitions like hyperlinks and can rank relevant content appropriately. It works the same way PageRank does as described in class but replaces searches with on-screen content. For example, suppose a user has information about a movie on their screen from a text message. In this case, if the user links to or has similar keywords to a certain app (i.e. if the movie mentions a location or time), Google can use this to generate relevant search results to pull up once the user has activated Now on Tap, using different screens as hubs to get data in terms of links, texts, and usage data to power its search results.
Google was founded on the idea of offering PageRank. Rather than simply categorizing different information, Google was innovative for it used PageRank to effectively produce top results based on a search. Now Google is adapting a similar technology in order to expand beyond links and move into physical products and apps. It also can use this data to effectively monetize its searched for apps and products.