Information Technology Reference
In-Depth Information
data can be continuously collected from user-media interactions and used for multiple
user profile updates.
The most popular technique of observing human computer interaction is eye track-
ing. Unfortunately, this technique cannot be used in real environment [6] - due to the
necessity of specialized equipment it is limited to laboratory controlled studies. How-
ever, implicit behaviour data on websites can also be acquired from other sources [7]:
on the server side, the proxy side or the client side.
Due to their inner limitations, server logs (and similarly proxy logs) store only data
connected with page requests to particular websites. Data which can be collected and
analyzed from the perspective of human media interaction is very limited. In contrast,
logging user behaviour on the client side allows a more detailed analysis of the inte-
raction. Highly accurate logging on that side is possible due to very good support of
Document Object Model (DOM) in Internet browsers. A number of user interactions
with the system can be controlled and monitored, e.g. mouse moves and clicks, scrol-
ling, selecting and copying content, saving and printing pages page and keyboard
inputs. That large amount of information describing the user-website interaction can
be used for many purposes, incl. the enhancement of usability and overall UX. It can
also be used to discover user preferences in order to dynamically adapt the interface
and website content accordingly. Such adaptation may be performed with regard to
individual or group needs. Personalization involves customization of web pages to
accommodate needs, interests, knowledge, goals or tasks of individual users. A perso-
nalized website can be described as a system that “adapts the content structure, and/or
presentation of the networked hypermedia objects to each individual user's characte-
ristics, usage behaviour, and/or usage environment” [8].
This article presents a tool for monitoring the preferences of e-commerce customer
preferences - ECPM (E-commerce Customer Preference Monitor). After some insig-
nificant changes, the tool may be used to monitor users on any kind of website. There
is a large number of interactions with the website that are monitored by ECPM, in-
cluding all those mentioned in the previous paragraph. The tool is an extension for the
Mozilla FireFox browser. However, the core monitoring code can be detached and
used also outside the extension.
The article also presents the results of human-website interaction study conducted
with the help of ECPM for 5 major business-to-customer online shops. The aim of the
study was to find links between user behaviour on certain product pages and prefe-
rences towards the products.
A review of literature shows that no great efforts have been made to assess the
efficiency of the implicit feedback approach in the field of B2C websites [9,10].
2
Related Work
The great volume of data gathered with implicit feedback techniques offers many
possibilities for observing and analyzing various aspects of human-website interac-
tion. In the recent years much effort has been made to discover user's preferences
from that stream of data. Among many studies, however, very few were dedicated to
online stores [10,11].
Search WWH ::
Custom Search