Designing Product Lists for E-commerce: The Effects of Sorting on Consumer Decision Making

Sorting Effects in Product List Design 1
Running Head: DESIGNING PRODUCT LISTS FOR E-COMMERCE
Designing Product Lists for E-commerce: The
Effects of Sorting on Consumer Decision Making
Shun Cai
Department of Information Systems,
School of Computing, National University of Singapore
3 Science Drive 2, 117543
Singapore
Tel: (65) 6516 1360
Fax: (65) 6779 4580
caishun@comp.nus.edu.sg

Yunjie (Calvin) Xu
Department of Information Systems,
School of Computing, National University of Singapore
3 Science Drive 2, 117543
Singapore
Tel: (65) 6516 6562
Fax: (65) 6779 4580
xuyj@comp.nus.edu.sg

---

Sorting Effects in Product List Design 2
Abstract
One common information display design which appears in nearly all electronic shopping
sites is product lists on web pages. Many electronic shopping sites provide sorting
functions for product quality attributes. However, the extant literature has not been
particularly insightful on how consumers respond to product lists sorted in different
orders (i.e., ascending, descending and random). Despite the intuitive postulation that
products appearing in an early position of a list may draw more attention, it is not evident
whether and how different sorting of products in a list affect consumers’ purchase
decisions. The purpose of this research is to investigate how product sorting influences
consumers’ decision making. Specifically, we focus on the effect of sorting by product
quality attributes on the importance of quality and price in consumers’ product choice
and on the formation of their consideration set. Results from our experiment show that,
when product quality and price are positively but not perfectly correlated, subjects
perceive quality to be more important and they tend to include products with higher
quality in their consideration set when they are exposed to a descending list than when
they are exposed to a random or ascending list. Therefore, online vendors could apply
descending sorting to promote the sales of high quality products.

Keywords: Electronic commerce, product list design, sorting effect, quality
importance, price importance

---

Sorting Effects in Product List Design 3
Introduction
E-commerce is growing steadily over the years. Online vendors are embracing the
advantages of dynamic interface design to keep shoppers happy and spending. One way
to increase profit for electronic shopping sites is to encourage consumers to purchase
high-end products in a product line because those products often have higher markups
and profits (Marn & Rosiello, 1992). According to Marn and Rosiello (1992), for an
average company, improving sales volume by 1% yields a 3.3% increase in profit
assuming no decrease in price. But, a 1% increase in price, assuming no loss of volume,
can boost profit by 11.1%. Our study, therefore, attempts to demonstrate that carefully
designing a product list can help online vendors promote high-end products in the list, i.e.,
those products with higher quality and price. This is possible because the information
displayed in an online environment is very malleable (West et al., 1999) and consumers’
purchase decisions are often constructed on-site under the influence of product
information presentation (Bettman, Payne, & Staelin, 1986; Bettman, Luce, & Payne,
1998; Hong, Thong, & Tam, 2005).
This study focuses on the design of a product list, which is a very common form
of information display in e-commerce (Hong, Thong, & Tam, 2004; Hong et al., 2005).
In a product list, a number of products are displayed sequentially in a web page for
consumers to search and choose (Diehl & Zauberman, 2005). A product list could be
used to present a product catalog or the result of a keyword search (Diehl, 2005). While
some web sites do not explicitly manipulate their product list, recognizing its practical
impact on consumer decision, others list the most popular products upfront (Tam & Ho,

---

Sorting Effects in Product List Design 4
2005). Still others provide sorting functions by price or by product attributes such as
brand, customer rating or quality indicators.
With an aim to sell high-end products of a product line, this study is particularly
interested in the effect of product sorting by quality attributes on consumers’ purchase
decision. In this study, we investigate the effects of a special type of product quality
sorting -- hierarchical quality sorting, in which products are first sorted by the attribute
that is most important to consumers. For products that assume the same value for the
most important attribute, they are then sorted by the second most important attribute in a
nested fashion, and this process repeats. Quality attributes refer to the product attributes
that indicate the configuration of the product. For example, in a digital camera purchase,
the value of mega pixels is often the most important criterion, followed by optical zoom,
LCD screen and so on. In this case, a hierarchical quality sorting of digital cameras is to
sort all the models by mega pixel first, and for the models with the same mega pixels,
they are then sorted by optical zoom, and LCD screen and so on. In this study, quality
refers to the objective overall configuration of a product. We define the average
perceived quality of a product as the average of subjective quality perceptions by a panel
of judges based on the overall configuration of a product. Average perceived quality
summarizes the multiple quality attributes of a product into an overall perception. The
advantage of hierarchical quality sorting is that it makes product comparison much easier
for consumers, but it does not necessarily sort products by average perceived quality. For
example, in descending hierarchical quality sorting, a digital camera with high
megapixels and low optical zoom could be displayed before a camera with slightly lower
mega pixels but much higher optical zoom, even though the latter might have a higher

---

Sorting Effects in Product List Design 5
average perceived quality. Hierarchical quality sorting (hereafter quality sorting for
simplicity) could easily be realized online but, to our knowledge, has not been adopted.
Our focus is not on the design of this sorting tool. Instead, we are interested in
how consumers respond to a product list sorted by quality. The intuition is that different
quality sorting (i.e., in a descending, ascending or random order) would affect the
importance consumers attach to product quality and price in a purchase decision, which
in turn affects consumers’ product choice. In short, our research questions are: How does
quality sorting affect consumers’ perceptions of quality importance and price importance
in product choice and the formation of consumers’ consideration set? Consideration set
refers to those products that consumers seriously consider and regard as interchangeable
when making a purchase decision (Hauser & Wernerfelt, 1990). We term the effect of
hierarchical quality sorting as a sorting effect. Sorting effect should be differentiated from
order effect. The latter refers to the impact of position that a product assumes in a list on
consumers’ evaluation of the product (Hogarth & Einhorn, 1992), while sorting effect
investigates the impact of the sorting method of a list (i.e., ascending, descending, or
random) on consumers.
The extant literature on product display has largely focused on how to increase
consumer welfare such as decision quality and effort (Diehl & Zauberman, 2005; Haubl
& Murray 2003; Hong et al. 2004). In contrast, we approach this issue by focusing on
how to design a product list to influence consumers’ behavior which in turn affects
vendor’s sales. An understanding of sorting effect helps marketers to strategically
construct product lists that lead to better business performance. In the following sections,
we will first review the literature on product sorting and catalog design. We identify that

---

Sorting Effects in Product List Design 6
the principle of concreteness and loss aversion are particularly useful to explain sorting
effect. Hypotheses are proposed based on the two theories. We then report on an
empirical study and findings. We conclude with discussion of findings and implications.
Theoretical Background and Research Hypotheses
Product List Display and Consumer Decision Making
In both marketing and information systems literature there are two streams of
studies related to product list design. The first stream of studies has investigated the
effect of product price sorting on consumers’ purchase decision. Quality information was
not available to consumers (e.g., Kosenko, 1989; Garbarino & Slonim, 1995). For
example, Garbarino and Slonim (1995) reported that subjects exposed to descending
prices purchased more pens than subjects exposed to ascending prices and were more
likely to consider their final purchase as of good value. Bennett, Brennan, and Kearns
(2003) employed two types of products: fast moving consumer goods (FMCG) and
household appliances. Their study tested the effects of price sorting and other variables
on the average price respondents were willing to pay. Price sorting was found significant
for FMCG but not for household appliances. There was a lack of theoretical explanation
for the mixed result.
Going beyond price information and price sorting alone, Diehl, Kornish, and
Lynch (2003) investigated the effects of quality sorting on consumers’ evaluations of
electronic birthday cards. They suggested that consumers’ price sensitivity will increase
when products are sorted by quality because products close to each other are more

---

Sorting Effects in Product List Design 7
substitutable in terms of quality. Consequently, consumers would find a lower price in a
sorted list than in a random one.
The second stream of studies does not investigate sorting, but focuses on
consumers’ quality or price sensitivity as a function of website design. One of the earliest
and influential studies is consumers’ choice of supermarket products. Russo (1977)
showed that when product price information was easier to process, sales could increase
by 11%. Creyer and Ross (1997) examined how the availability of information about the
value of a product, expressed as a ratio of quality per dollar, influences preference
formation. Their results indicated that the presence of the quality-price ratio increases the
likelihood of consumers choosing a product with a lower price and higher value
compared to consumers presented with separate price and quality information. However,
in their purchase simulation, the quality rating of products was assigned by researchers
rather than perceived by consumers. In the e-commerce literature, Alba et al. (1997)
suggested that online retailing reduces the information search costs for price. Hence
consumers will become more price sensitive. Lynch and Ariely (2000) showed when an
online retailer makes quality information easier to search and compare, consumers
became less price sensitive. Haubl and Murray (2003) found that when a product attribute
is included in a product comparison matrix produced by recommendation agents, this
attribute becomes more processable, hence more prominent in consumers’ purchase
decision. Diehl (2005) proposed that, although a sorted product list lowers the search cost
for consumers, searching too much in a sorted list could degrade choice quality.
In short, the extant literature has found that price sorting affects consumers’
purchase decision and that the ease of processing quality information tends to boost the

---

Sorting Effects in Product List Design 8
weight of quality in consumer decision making. However, very few studies have
investigated quality sorting.
The effect of different quality sorting methods on consumers’ decision making is
expected to be a two-step process. First, we believe that quality sorting affects
consumers’ quality importance and price importance in product choice. Second,
consumers’ perception of quality importance and price importance affects the average
quality and price of products in their consideration set.
Consumers are guided by their perceptions of the importance of various product
attributes in information seeking, product evaluation and purchases (Goldstein, 1990;
Mackenzie, 1986). Attribute importance could influence consumers’ product selection
and evaluation process regardless of whether a compensatory (e.g., weighted sum of
attribute utility) or a non-compensatory (e.g., lexicographic or elimination-by-aspects
with products evaluated by attributes in descending importance) strategy is used in
decision making (Heath & Ryu, 2000; Sanbonmatsu, Kardes, Houghton, Ho, & Posavac,
2003). Quality and price are the most important yet competing product attributes. When
there are multiple quality attributes, consumers often have to consolidate them into one
“meta-attribute” and then make a trade-off between quality and price (Kivetz, Netzer, &
Srinivasan, 2004). This trade-off is determined by the subjective weights consumers
assign to product quality and price. In this study, quality importance (QI) refers to a
consumer’s perception of the importance of product quality in the consumer’s purchase
decisions (cf. Kalra & Goodstein, 1998). Similarly, we define price importance (PI) using
a similar conceptual base as sensitivity to price. Given product and price importance, the
notion of relative importance of quality over price (RIQP) can be defined as the

---

Sorting Effects in Product List Design 9
perceived relative importance of quality as compared to the importance of price. Relative
importance of quality over price represents a consolidated measure to capture consumers’
trade-off in purchase decision.
We chose consideration set composition as the ultimate dependent variable
because consumers often adopt a hierarchical choice process to simplify complicated
choice problems. They filter out unwanted products first to form a consideration set
which contains substitutable products, among which they make a final choice (Alba et al.,
1997; Bettman, 1979; Nedungadi, 1990). Empirical studies suggested that consideration
sets play an important role in quantitative models to predict consumer choice (Hauser &
Wernerfelt 1990; Simonson, Nowlis, & Lemon, 1993).
With these concepts laid out, we resort to the principle of concreteness and loss
aversion in consumer behavior to explain the effect of quality sorting on consumers’
decision making.
Sorting Effect: The Principle of Concreteness
The first mechanism in a sorted product list is the principle of concreteness
(Slovic, 1972). An important concept in the principle of concreteness is processability,
which refers to the ease with which information can be interpreted and used (Bettman et
al., 1986). Based on the constructive preference perspective of decision making,
consumers tend to construct quality and price importance as well as their preferences on
the spot when product information is prompted (Bettman et al., 1998). Based on the
principle of concreteness, decision makers tend to (1) use only the information which is
explicitly displayed in a stimulus environment and (2) process information in the

---

Sorting Effects in Product List Design 10
particular form in which it is presented. Decision makers are often unwilling to spend
more cognitive power to re-organize information or search for more. When there are
multiple information items, the one that is more processable is often picked up (Creyer &
Ross, 1997).
Compared to an unsorted (i.e. randomly ordered) list, when products are sorted by
quality, quality information should be relatively more processable because quality
attributes can be easily compared. Based on the principle of concreteness, the enhanced
processability of product quality information will, in turn, increase its importance in
purchase decisions. In other words, the processability of quality attributes makes quality
more salient as a criterion in selecting alternatives (Areni, 1999). Therefore, we
conjecture that when products are sorted by product quality, either in ascending or in
descending order, consumers will attach higher importance to product quality than when
products are in a random order. We hypothesize:
H1a
The quality importance is higher in consumers’ purchase decision when a
product list is sorted in a descending order of quality than when it is in a
random order (i.e., QID > QIR, where the subscript indicates the hierarchical
quality sorting method of Descending, Ascending, or Random).
H1b
The quality importance is higher in consumers’ purchase decision when a
product list is sorted in an ascending order of quality than when it is in a
random order (i.e., QIA > QIR).
Next, we consider the influence of sorting on price importance. Typically, in the
real marketplace, product price often positively correlates with product quality. In such a

---

Document Outline

• ??
• ??
• ??
• ??
• ??
  • ??
  • ??
• ??
  • ??
    • ??
    • ??
    • ??
• ??
  • ??
    • ??
• ??
  • ??
    • ??
• ??
• ??
• ??
• ??
• ??

---