“Information Overload: Evidence
from the Field in an Online Retailer Setting”
Working Paper with
Miguel-Villas Boas, UC Berkeley
Steven Huff
June 26, 2007
Abstract
This paper is exploratory in nature and investigates
the existence of information overload in the online search behavior of
consumers in a real-world setting by creating an experimental product category
in an online retailer in office supplies.
Subjects were sent e-mails with varying numbers of
products, each with descriptions and links leading them to an experimental
product category on the retailer’s website, where they were presented with a
random number of products. Data shows how the number of links presented in the
e-mail affects the likelihood of clicking one of the links, and how the number
of alternatives subsequently displayed on the experimental category menu page
affects browsing behavior.
Results indicate that the likelihood of click-through
initially increases with the number of products presented, but subsequently
decreases as the number of alternatives increases beyond four (i.e., there is a
local maximum of click-throughs when four products
are presented); this suggests that consumers are indeed affected by information
overload, and arguments in support of this conclusion are presented.
Results also indicate that the global maximum of click-throughs for both e-mails and the experimental category
webpage occurs when the maximum number of products is presented. This occurs
because the cost of a click is very low when compared to the expected benefit
of more diagnostic information, which makes clicking a link a viable
alternative strategy to cope with information overload when compared to the
option of exiting the choice situation. Thus, as the number of links increases
beyond the point of overload, so does the likelihood of clicking one of these
links instead of exiting the choice situation, and eventually this choice dominates the exit choice.