Logistic
Regression
An introduction
Prof. Bhuwandeep
Almost all of us are familiar with odds.
What are the chances one thing will
happen versus another?
What are the chances you will succeed
Introduction at work today?
What are the chances your favorite
game-show contestant will win today
versus the chances he or she will lose?
• What you might not be familiar with is how
odds can be applied to marketing analytics.
• What are the chances a customer will buy
your product versus the chances he or she
Examples won’t?
• What are the chances you will retain a
customer versus the chances you will lose
him or her?
• When you are using odds, you are examining
two opposing outcomes. Any such unknown
(one that can only be one thing or another)
So… is known as a dummy variable.
• But if you know how to examine dummy
variables properly, the results are anything
but dumb.
• Think about an important metric in marketing:
customer retention. If Keep-money Bank wants to
use a regression analysis to examine whether it will
retain a customer, it will set retention as its
dependent variable.
• Rather than being normally distributed in a bell curve
in the manner of continuous variables ( Figure 13-1 ),
however, a 1 will be assigned to represent customer
retention and a 0 will represent customer loss. Only
those two outcomes are possible. Again, this is a
dummy variable, wherein what you are trying to
predict is one of two options
• Studies have shown that logistic regression is the
best model for examining dummy variables such
as customer retention.
• But why can’t Keepmoney use its trusty linear
regression to determine the likelihood of
customer retention given a set of independent
variables?
• Again, linear regressions assume a bell-curve
distribution of outcomes (what is known as a
normal distribution) from negative infinity to
infinity. Most things in life follow this sort of
distribution.
Choice Behaviour
• The objective of logistic regression in this
example is to represent consumers’ choice
behavior as accurately as possible. When
individual consumers choose products, the
value they place on the product does not
typically increase linearly with increases in a
preferred feature of the product. Instead,
research indicates consumer valuation of a
product typically follows an S-shaped curve
with increases in the levels of a preferred
attribute.
• You can test whether the S-shaped curve represents
consumers’ choice behavior with a simple exercise.
Imagine that on the x-axis you have the level of discount
on a $300 plane ticket from Charlottesville, Virginia, to
New York. Ask a group of your friends how many of them
would purchase the flight. Then offer a discount of $20.
How many additional people said they would buy the
ticket? Probably not many. Increase the discount to $40.
Maybe one person half-heartedly jumps in. At $60, you are
likely to see a spike in purchasers. And from $60 to $100,
the number of purchasers should increase at every level;
however, at about $100, the number of
additionalpurchasers will taper off, as you have reached
the upper threshold.
• In most real-life situations, this S-shaped
curve represents how people make
decisions.
• As a discount (or promotion) increases,
the odds that people will make the choice
to buy will increase. In this example, at a
$60 discount, 2 in 10 people are likely to
purchase the flight to New York; 8 in 10
are unlikely to purchase the flight.
