Abstract:
Consumers often face a task to select a best option from a large set of alternatives,
such as choosing a car to buy1, an apartment to rent2, or an unforgettable trip to book3.
E-commerce sites frequently provide the possibility to search for structured items,
usually by asking a user to fill in a form asking about the requirements that a desired
product has to satisfy (preferences). This process is used, for example, when searching
for a used car, or a flight on popular websites, and is also referred to as preference-
based search or parametric search. Although such choice-based approaches are
prevalent, both users and retailers can find them unsatisfying. One of the major reasons
is that users are often not able to correctly transform their preferences into requirements
using online forms, and thus they are rarely provided with the information
they need.
On the other hand, consumers making purchase decisions in online shops are often
unable to evaluate all available alternatives in great depth, and so seek to reduce the
amount of information processing. Interactive decision aids that provide support to
consumers are particularly valuable in helping to determine which alternatives are
worth further, detailed consideration. Customers are being provided with a number of
different decision aids that, ideally, should enable them to search, browse and compare
vast numbers of available products. Retailers offer various recommender systems
that attempt to provide customers with manageable and relevant set of products based
on user profiles, history of interactions and product descriptions. Moreover, various
techniques for preference elicitation (e.g. dialogs in conversational recommender systems)
are used to enable better understanding of customers¿ needs. However, there are
many factors (e.g. the number of available products but also by the precision of information
preference elicited) that can impact the performance of decision aids in online
shops. We discuss the most popular decision aids in the context of online shopping
and decision-making.
We also introduce the concept of a soft-boundary pre-filtering decision aid. The decision
aid modifies the pre-filtration criteria provided by the decision maker and thus,
allows him to reconsider selected alternatives she initially eliminated. We propose a
model of such a decision aid, give an overview of its different configurations, and
provide the illustrative example of the scenario of use in the apartment selection decision
problem. We also hypothesise about the impact of the proposed decision aid on
decision quality and consideration set size and quality. We conclude the paper with an
overview of potential directions for future research and a discussion of benefits of application
of the soft-boundary pre-filtering decision aid.
