Monitoring Human Website Interactions for Online Stores - New Contributions in Information Systems and Technologies - page 356

only to those which can be gathered via a regular JavaScirpt code attached to any
website. The analysis of possible solutions led to the conclusion that an ideal candi-
date for implementing ECPM is the FireFox browser, which is very popular among
Internet users. It allows to install extensions called Addons to enhance the functionali-
ty of the browser. The extensions mainly use two programming languages: JavaScript
and XUL for Interface (XML-based User-interface Language).
ECPM allows monitoring of various users interactions with a website by utilizing
DOM (Document Object Models) Events model. The ECPM extension contains the
Preference window, where a user can deactivate the extension, provide personal data
(optional), learn about the study (opening a page with study description) and finally
end the study by sending results to the server (until that moment all user data col-
lected during the study are stored on the user's machine).
ECPM was built for the purpose of the study among some of the biggest polish e-
commerce websites. Despite that after few simple modification it can be used to mon-
itor behaviors on any website.
ECPM was instrumented to actively monitor user's activity on product pages for
five major Polish online stores: Merlin.pl, Agito.pl, Electro.pl, Empik.com and Mo-
rele.net. Our extension collects data on the visited product pages and user interactions
with them. The product pages contain detailed information about particular products.
ECPM was built to register numerous parameters describing physical attributes of the
page visited (Table 1).

Table 1. Page parameters registered by ECPM

Parameter Description

document_length Number of characters within all texts on the page

desc_length Number of characters within product description

review_length Number of characters within product reviews

Number of characters within other recommended products
section

recommend_length

image_number Number of product images

page_height Page height in pixels

Source: Authors' study.

Moreover, during page browsing ECPM monitors user behaviors via DOM events
generated by their activity. The data is used to obtain 24 implicit indicators listed in
Table 2. They could also be calculated just by attaching a regular JavaScript code to
the monitored page. The indicators reflect user interactions with the website. The time
spent on a web page ( page_time ) is one of the most intuitive candidates for a user
interest indicator. Beside page_time and tab_active_time (time of the browser tab
containing a particular page being active), other indicators regarding time have been
proposed by the authors. For example, user_active_time is calculated as a sum of two-
second periods when the user is actively interacting with the website -it accounts for
the fact that is a very common situation when users open many pages in different tabs
or leave the computer while the page remains open.