Monitoring Human Website Interactions for Online Stores - New Contributions in Information Systems and Technologies - page 358

Nowadays almost every product page contains sections such as: description, im-
ages, reviews, and recommendations of related products. Authors hypothesize that
moving mouse pointer into those sections may indicate user interest in particular con-
tent. In order to express user engagement, four indicators have been proposed. They
define time of mouse pointer being present within certain areas. Because of varied
HTML structure of particular websites, there are hard coded patterns for calculation
of such indicators.
Apart from non-relative indicators which are calculated on the basis of user beha-
viors, several indicators taking into account inherent webpage characteristics were
also constructed. In order to calculate those indicators, the following measures are
calculated for each website: total page characters length, total recommendation area
characters length, total recommendation area characters length, page height, number
of product images, visible page height. By using behavior data and the above-
mentioned page measures, relative indicators presented in Table 3 are calculated.

Table 3. Relative indicators calculated by ECPM extension

Relative indiciator Description

rel_page_time page_time / document_length

rel_user_active_time user_active_time / document_length

rel_tab_active_time tab_active_time / document_length

rel_prod_desc_time prod_desc_time / description_length

rel_prod_recommend_time prod_recommend_time / recommendation_length

rel_prod_review_time prod_review_time / review_length

rel_prod_image_time prod_image_time / image_number

rel_mouse_distance mouse_distance / page_height

rel_vertical_scroll vertical_scroll / page_height

rel_horizontal_scroll horizontal_scroll / page_width

Source: Authors' study.
Additionally, in order to measure user interest in products, every time a user
leaves a product page, an interest evaluation pop-up window is displayed. This win-
dow contains two questions concerning the product. The first one asks how much a
certain product is of interest to the user. It allows the user to express their explicit
interest in the product in a five-point scale. The second question checks if the user has
been familiar with the product before the study. It checks if the user has known the
viewed product before (in a Boolean scale). After filling the form the answers are
stored so that id the user revisits the product page, the form displayed upon leaving
will contain previously selected answers.