Travel Reference
In-Depth Information
considered more valuable and important and, thus, having a higher importance compared to the
ones with fewer inbound links. In addition, search engines use an iterative process to determine
the quality of links (Brin and Page 1998; Langville and Meyer 2006; Levene 2006). Other criteria
affecting page ranking include age of the site and the frequency of updating (Malaga 2007;
Sullivan 2006), page loading time and the popularity of the page (Fish 2009).
In general, link structures, web page content, frequency of updating, the loading time and
the implicit feedbacks in the form of click behavior from users are all determining factors in the
ranking of search results. Still, indexing, ranking and representing the enormous amount of
information on the Internet is a huge challenge for search engines. While the most popular
search engine, Google, currently claims to index more than one trillion web pages (Google
2008), the entire information space on the Internet can only be covered in small parts by a single
search engine (Levene 2006). The 'deep web' pages such as those buried in the databases and
dynamic pages are not indexable by many search engines and, thus, are inaccessible to users who
query a search engine (Bergman 2001; Lawrence and Giles 1999).
The travel information searcher
While travellers may use search engines at different stages of their trip, a majority of them fi nd
search engines particularly helpful in serving their trip planning purposes (TIA 2008; USTA
2011). Studies have shown that the process of using a search engine consists of two major
cognitive steps including query formulation and search results evaluation. Search queries have
been studied extensively in fi elds such as information sciences as well as travel and tourism
(Jansen and Pooch 2001; Jansen and Spink 2005b; Jansen, Spink and Saracevic 2000; Pan
et al
.
2007; Pan, Litvin and O'Donnell 2007; Xiang and Pan 2011).
Search queries are short strings of words or terms that refl ect a user's goals, information needs,
search intent, as well as his/her search strategies. Studies in information science, consumer
behavior and tourism have explored the characteristics of search engine queries such as the
length and depth of search, types of search and changes of search characteristics over time. For
example, a typical web session is around 15 minutes; 47 per cent of users only search once during
a session; about 20 to 29 per cent of queries only contain one term; in the United States, around
11 to 20 per cent of user queries contained logical operators; users only view a few result pages,
mostly the fi rst page (73 per cent of users) ( Jansen and Spink 2005b; Jansen, Spink and Saracevic
2000). Search queries refl ect users' goals including navigational goals (looking for a specifi c web
page), informational goals (trying to obtain a piece of information) and transactional goals
(carrying out certain action) ( Jansen and Molina 2006). Recently, Jansen
et al
. (2008) found that
users' queries in general are largely informational (81 per cent), followed by navigational tasks
(10 per cent) and transactional tasks (9 per cent).
In travel and tourism, recent studies indicate that travellers' questions tend to be short,
consisting of less than four keywords; and, most travellers do not go beyond the results provided
on the second page. As a result, only a relatively small number of websites are visible to the
traveller though millions of potential web pages were found (Xiang, Wöber and Fesenmaier
2008). Pan
et al
.'s (2007) study also indicates that searchers in the US usually focus on cities as the
geographical boundary instead of states or countries; and, travellers often combine their searches
for accommodations with other aspects of the trip, including dining, attractions, destinations, or
transportation. In addition, this study indicates that there are strong associations between
place names (particularly city names) and a specifi c hotel and a hotel brand. Xiang
et al
.
(2009) found that there are relatively few distinct words in travel queries beyond 'hotel' and
'attractions', indicating that there is a 'long tail' of words that represent travellers' heterogeneous
