Prior to purchase, many consumers tend to search information that is related to products or services and consider various things by online, the main place to produce and exchange. Also online information search behavior is becoming more frequent and helps consumer to search in detail by using not only PC but also Mobile. However it does not always lead to online purchase, instead, ‘Cross-over Shopping’, which is known as a purchase pattern that using Online and Offline at the same time, comes to the fore.
This research will explore the online information search behavior by consideration set, and explain features which lead to online purchase on online information search based on the concept of consumer purchasing behavior on online using existing results of prior research about online information search behavior.
For this study, an renowned advertising company in Korea provided the web log data set collecting pannels’ online behavior on both PC and mobile. The volume for the data set is composed of 1,301,785 rows of personal internet log data that people answered to purchase cosmetics in 3 months and considered one more brand during the consumer decision process. Therefore the number of pannel for the analysis is 108 that might be thought as a small sample. Nevertheless the sample size, its overall volume of web log which were produced from those sample were 109, and it is enough to provide validity and its statistical significance.
This research used various classification standards for search queries. At first, search query for various search engine was classified into 4 categories according to its characteristics: 1) brand related query, 2) products related query, 3) promotion related query, and 4) retailer related query. After these categorization, we calculated total amount of search queries which were generated by pannel. Also we estimated the level of search queries by how many keywords were conjointed within a search query. Second, to analyze the overall consumers’ online behavior, we adopted 1) total amount of page view as a representative variables for how many the consumer search, 2) site diversity that means how the consumers access to the various types of website, 3) depth of search as for a search quality. Third, the web page which was access after consumer run their certain query, was classified into 6 different groups: 1) official homepage for the brand or product, 2) informational web site including news and magazine, 3) online shopping site, 4) online community, 5) entertainment web site like youtube, and 6) portal site.
In order to analyze the effect of keywords related factor for a online purchase, this study applied discriminant analysis as a methodology. This study adopted 2-stage analysis. at first the size of consideration set was used as a dependent variable prior to actual purchase, then we used actual purchase as a dependent variable. The results reported in 3 Table(11, 12, 13).
As a first result, when consumer used brand related keywords or promotion related keywords it could be a reliable predictor of size of consideration set. But this result appeared only on PC, it means mobile consumers’ information search behavior using keywords might not directly related to building consideration set.
Then we conducted follow-up study using online purchase as a dependent variable at two perspective each. One is of the condition in which consumer considered only one brand for their purchase, and the other is of consideration set size of two and more. Between PC and mobile, only retailer related keywords was significant variable for online purchase prediction under consumers considered only one brand. Other variables in both condition of consideration set size of one and two more could not discriminate purchase on online or not.
<Table. 12> Predictors of online purchase based on the consumers’ information behavior <Table. 13> Predictors of online purchase based on the website types As a second result, consumers’ overall information behavior on PC affect how many brands are considered within decision making process, especially for the numbers of page that consumer accessed during informational search and site diversity across six different web site categories. We also tried further analysis same as the keyword types analysis that were conducted previously. We could find significant variables in conditions of which consumer considered two and more brand as alternatives. Online purchase could be predicted with site diversity and depth of search on PC, on mobile on the other hand, quantity of search was a effective predictors.
At Last, we used online website types to discriminate size of consideration set and online purchase. To Access blog or community, entertainment site and official brand website induced positive effect on the performing consideration set on PC, but none for mobile. For the second stage of this analysis, mobile was not significant with all of variables. If consumer consider certain brand meerly, their visit on portal and informational site raises the possibility of online purchase.
As a summary for results, PC affects more than Mobile in online information search process. Specifically, prior to purchase, there are factors that influence consideration set in the keyword search behavior; (1) the keywords that is related to ‘Brand’ or ‘Promotion’ (2) the amount of total keywords in search behavior, (3) the amount of visit pages, (4) the various visit sites in visit behavior and (5) community, shopping, entertainment and official homepage visit tendency. Second, the conspicuous feature that online purchases are affected by consideration set in terms of searching and visit behavior can provide more features of consumer online decision journey. Although there were some constraints on analysis from huge data size, this research could have implication on potential research about consumer online decision journey by panel data containing all online search process.
Which Coupons to use where? The Impact of Different type of Coupons on..
16 페이지
