| Zurich,
Switzerland, May 23 - 24, 2006
World
Trade Center
Leutschenbachstrasse
95
CH-8050 Zurich
Switzerland
Limited Space Available
Discrete
Choice Analysis (DCA) allows
us to understand the behavioural process that leads to
an individual's choice. This is accomplished by statistical
models which describe decision-makers' choices among alternatives.
The decision-makers can be people, households, firms,
or any other decision making unit, and the alternatives
might represent competing products, courses of action,
or any other option or items over which choices must be
made. DCA is based on administering a series of choice-based
experiments, where respondents are shown different products
or services, with each product or service having different
attributes. For example, respondents might be shown four
different computer models, each model having a different
CPU speed, Ram, hard drive capacity, weight, and price.
This choice-based experience is repeated over and over
again, each time the attributes for each product are changed.
In some cases, individuals are given the opportunity to
select none of the products or services on the list, thus
simulating as closely as possible the true purchasing
environment.
Applications:
- Predicting the Sales and market
share penetration of new products
- Predicting the Sales and market
share penetration of competitors' new products
- Predicting the effectiveness
of product marketing (features, packaging/labeling,
pricing, distributing, promoting, packaging) strategies
- Understanding Brand-shifting
Behaviour
- Predicting the effect of sales
promotions
- Market segmentation based
on decision making behaviour
- Measuring and predicting Brand-Switching
Behaviour
- Predicting how changes to a
product's packaging, pricing, advertising, and distribution
will affect Sales and Market Share
- Understanding consumer decision
making behaviour
Advantage over
Conjoint Analysis:
Discrete choice holds a number
of advantages over traditional conjoint analysis including:
It is a more realistic exercise for individuals
to indicate which product they would purchase rather than
rating/ranking since this is what they actually do in the
marketplace. In discrete choice, individuals can be
given the option to select “none” of the products, thus
indicating that they do not find any of the products appealing.
Discrete choice allows for much more complex statistical
modeling to be performed, which often yields better data
(e.g., interactions, alternative–specific effects, cross-effects,
etc., can be accommodated). As with traditional conjoint
analysis, the utilities that come from discrete choice
can be used to develop market simulators and can also
be used to examine whether different segments exist, which
use different decision making rules.
Stated
Preference and Revealed Preference:
Revealed preference relates
to individuals' actual historic choices, such as the car
they own, or brand of coffee they recently purchased. Stated
preference relates to future choices, and is assessed in
an experimental setting, where choices are made in hypothetical
situations. Individuals are repeatedly asked to select among
a group of products or services, while each time each product
attribute, such as price, quality, size, colour, etc, changes.
Advantages / Disadvantages
of Revealed Preferences:
- reflects actual choices
- choices are limited to the products
and services presently available in the marketplace
- no variation may exist in relation
to some attributes, such as the price of electricity
(everyone is charged the same price)
Advantages of Stated Preference:
- assesses consumer behaviour to new
products / services
- allows for the assessment of variation
in product attributes that do not exist in reality,
such as the price of electricity
- what individuals say they will do is
often not the same as what they actually do
Combining revealed preference
and stated preference allows us to capture the advantages
of each approach.
Past
Choices Reveal Future Choices:
The choices a person makes
at one point in his/her life may have an impact on the
options or choices available to him/her in the future.
Taking individuals' previous choices into consideration
allows for more accurate predictions of future choices.
It also allows for the assessment of market share shifts,
and shifts in brand loyalty. |
Discrete Choice Analysis: Predict Consumers'
Future Behaviour
TYPES OF Discrete
Choice Models:
- Logit
- Probit
- GEV
- Mixed Logit
- Latent Class
Logit
- Hierarchical
Bayes Mixed Logit
What you will learn:
- When each type of discrete
choice model is most appropriate
- How to estimate each type
of discrete choice model
- How to interpret the results
of each model
- How to simulate future changes
in consumer preferences, and market share, based on
changes in product attributes
- How to segment consumer choice
into sub-groups, using individual level information
provided by latent class choice models, or probit
/ mixed logit models.
Hands
on learning:
Learn by doing. Attendees
will need to bring a laptop. Attendees will be able to use
discrete choice analysis by the completion of this course.
Predict How Consumers will React in
the Future
Using Discrete Choice Analysis
(DCA), researchers can understand how consumers will react
in the future, before changes to the marketing mix are instituted.
In other words, before a product's packaging is altered,
retail price is changed, advertising campaign is launched,
or new product is developed. DCA can be used to determine
the effect of such changes on sales, and market share. Moreover,
one can also assess the effect of such changes to the marketing
mix on competitors' sales and market share. Almost any "What
if" question can be answered.
Growing sales and profits fundamentally
depends on understanding how your customer thinks. Victory
in marketing warfare belongs to the firm that best understands
how consumers think.
Advantage over other
Workshops / Courses
Unlike other
courses teaching discrete choice analysis, we do not use
only one software package. As this workshop is not a means
of marketing software, it will be taught using all the leading
software packages: LEM, Nlogit, Latent Choice, Guass, and
Winbugs. Each software package has its advantages and disadvantages.
In addition, this course teaches you to use a variety of
discrete choice models, not only models specific to one
type of software package being marketed.
Attendees will
learn by doing. Multiple examples will be used throughout
the workshop.
|
