External to internal search: Associating searching on search engines with searching on sites

We analyze the transitions from external search, searching on web search engines, to internal search, searching on websites. We categorize 295,571 search episodes composed of a query submitted to web search engines and the subsequent queries submitted to a single website search by the same users. There are a total of 1,136,390 queries from all searches, of which 295,571 are external search queries and 840,819 are internal search queries. We algorithmically classify queries into states and then use n-grams to categorize search patterns. We cluster the searching episodes into major patterns and identify the most commonly occurring, which are: (1) Explorers (43% of all patterns) with a broad external search query and then broad internal search queries, (2) Navigators (15%) with an external search query containing a URL component and then specific internal search queries, and (3) Shifters (15%) with a different, seemingly unrelated, query types when transitioning from external to internal search. The implications of this research are that external search and internal search sessions are part of a single search episode and that online businesses can leverage these search episodes to more effectively target potential customers. It is known that people frequently use multiple information platforms (Hoa, Lin, & Chen, 2012), such as search engines and websites (Kumar & Tomkins, 2010), in order to address a task that involves searching. For example, a substantial percentage of Web users view major search engines (e.g., Google, Baidu, Yandex, NAVER) as entry points to the Web. These users may then traverse from a major search engine to a particular website (e. These searches are obviously related via a common task. However, searches on these separate platforms have typically been analyzed in isolation (i.e., treated as separate searching sessions). In reality, given the single underlying task, the searches likely comprise a single search episode that should be examined holistically (i.e., not treated as separate searching sessions) in order to understand the underlying user task. In this research, we examine one aspect of these multi-platform searching episodes, namely the search that brings the user from a major search engine to a particular website. Understanding this type of searching is important, as the majority of websites rely on search engines for substantial percentages of their traffic, with more than 80% of Web and Internet users employing a search engine as their starting point