P1: 61
Nyberg WL040/Bidgoli-Vol III-Ch-11 July 11, 2003 11:49 Char Count= 0
PROTOTYPING FORE-COMMERCESYSTEMS 139applet, or application) as a tool for reaching consensus on
the system’s modes of operation, inputs, and outputs.
Object-oriented analysis (Bruegge & Dutoit, 2000;
Coad & Mayfield, 1999) and the Unified Modeling Lan-
guage (Pooley & Stevens, 1999) provide useful tools for
use case analysis and requirements modeling. However,
the resulting model diagrams cannot be discussed directly
with a customer or user (unless he or she is skilled in UML,
which is unlikely). It is more useful to construct a proto-
type that illustrates an initial version of each user function
in a familiar interface (e.g., Web pages or applets). Once
such a prototype has been constructed, the engineers can
demonstrate the system for the customer, talking him or
her through each proposed function for each user role.
Feedback from this kind of prototype demonstration can
be revealing. Often the initial requirements from the cus-
tomer are vague or ambiguous; a prototype review can
uncover potential misunderstandings. For example, an
initial requirement might be specified as “provide a prod-
uct search capability for the user,” without being specific
about the search criteria. If the prototype omits certain
search criteria (e.g., particular product characteristics
that are important to the customer), the customer will
notice in the review, and the requirements document can
be updated to reflect the more specific requirement.Interface Prototypes
A typical Web-based system will include several pages or
“screens” in support of each user interaction or user func-
tion. Each of these pages will contain a variety of user
interface elements, such as hyperlinks, pull-down menus,
fill-in forms, checkboxes, and radio buttons. The pages
will be arranged in a particular hyperlinked structure,
starting from the home page of the application. If pos-
sible, navigation should be transparent and effortless, so
that the user can focus on the task at hand without be-
ing distracted by poorly designed navigation, or content
that is difficult to scan visually (Nielsen, 2000). A proto-
type can be used to evaluate a set of criteria regarding the
interface; questions to consider include the following:Are the user interface elements the most appropriate ones
for the task?In order to minimize the time it takes for
the user to scan a Web page and make the appropriate
selection or provide required data, it may be neces-
sary to redesign the layout and interface components.
For example, a long block of text with choice buttons
at the bottom will be less effective than a short list of
bullet items including hyperlinks. The former must be
read in its entirety, whereas the latter can be scanned
quickly for the desired choice. Fill-in forms should be
constructed for easy visual scanning from field to field,
and common conventions (e.g., tab key) should be uti-
lized for switching the focus from field to field.
Do the user interface elements support an effective task
flow?If a user task involves decision-making during
a set of sequential steps, it is important not to force
an early decision that might be retracted later based
on the result of a subsequent step. Prototype evalu-
ation should consider what will happen if the user
changes his or her mind about the task at hand. Normalprogress through a task should be streamlined and ef-
fective, but it must also be easy for the user to retract
a choice, cancel a transaction, or start over.
Are suitable defaults provided for menu selections and fill-in
forms?An initial requirements specification may not be
detailed enough to provide information about the de-
fault content or behavior of fill-in fields, search param-
eters, date ranges, etc. which are part of a user input to
the system. Effective defaults are an important part of
streamlining the user’s experience, especially when a
lot of data must be entered and sensible defaults exist.
Does the system check for combinations of data values that
are invalid for the application?It should not be possible
for the user to enter combinations of values in a fill-in
form that result in an invalid request or transaction.
Testing a prototype can be an effective way to uncover
such cases, if they have not been specified in detail in
the initial requirements documentation.
Is it easy to navigate from page to page in the site or
application?Effective Web applications are “training
free”—the user can immediately intuit how to navigate
through the site or application after briefly scanning
the first page or screen. If the user has to stop and think
about how to navigate, or worse, experiment with dif-
ferent mouse clicks to find what he or she needs, the
effectiveness of the site will suffer and the user’s sub-
jective experience will degrade.
Can each part of the site or application be reached with
a minimum number of mouse clicks?Even if intuitive
navigation is provided, a particularly deep or content-
rich site might require lots of user interaction to reach
certain portions of the site content. A prototype eval-
uation can assess whether a reorganization of the site
content or navigation structure might provide better
accessibility. Secondary navigation mechanisms such
as search engines or site maps can be valuable in this
regard.Although by no means exhaustive, this list illustrates
how a prototype review can be used to improve the in-
terface component of a Web application. By discussing
these criteria with the customer and/or evaluating them
in a more formal usability study, valuable information can
be gathered which can trigger design improvements for a
full-scale system.Content Prototypes
The termcontentis used in an e-commerce context to refer
to the parts of the site or application that are not directly
related to navigation or data entry, which are typically
presented as a combination of text, graphics, and/or mul-
timedia. Content must simultaneously satisfy two basic
criteria—it should be consistently presented and it must
be easy for the user to scan and understand. The page
designer determines how to partition and present infor-
mation in a manner that is consistent with the overall vi-
sual design (look and feel) of the site. Choices of typeface,
font size, spacing, color, and page size all have a dramatic
effect on how easy it is for the user to assimilate the con-
tent (Siegel, 1996). Helping the user to complete the task