Marketing Analytics Exam

multiple regression in marketing mix models.

if a company can find an important variable before their competitors find and use it, management can build its strategy around that variable to gain an advantage.

metric goal

dependent variable

marketing input

independent variable

line equation concept term

intercept

concept term given for the number before x. for example, y = 1.42x + 9.9

slope

slope also stands for

regression coefficient

what does R² represent

how accurate the line equation is

if R² is .601 what would that mean

about 60% of the data points are accurately predicted

if R² was 0.0 what would that mean

the line predicts none of the data points

if R² was 1.0 what would that mean

than the line predicts ALL of the points

is your p-value is 0.00 what does that represent

there is a 0% change that intercepts and slope coefficient will substantially change if you added population data or took a different sample from population

dv: superbowl ad liking - whats the iv

storytelling

dv: ad liking - whats the iv

dramatic acts (storytelling)

what was the resource allocation for super bowl ads (30 sec)

8 million

single variable linear regression model

just one iv

when do you want the tension level to be highest in the dramatic structure

act 3

linear regression is also known as

marketing mix modeling

4 p’s of marketing

price, product, promotion, place

prescriptive nature

if I allocate x-amount of time/$$/energy to iv, then i can expect an increase (or decrease) in dv

evaluation of the different slopes associated with all the added ivs

which iv slope coefficient among ivs is the biggest

slope is also known as

effect size

four key rules to determine a causal relationship between two variables

• implementing a marketing activity increases sales

• not implementing a marketing activity causes no increase in sales

• implementing a marketing activity today affects sales in sequent time periods

• there were no other established external/alternative factors affecting sales (MOST DIFFICULT)

dan willards marketing experiment

runs a/b test on package design elements and their effects on sales velocity

a/b test in our in-class questionnaire - condition a

mktg451 section students

a/b test in our in-class questionnaire - condition b

mktg495 section students

marketing experiment is what type

after-only experiment

when, in our questionnaire, is there experimental manipulation

section #3 - in our scenario

iv1 —> dv is what

isolated

marketing experiments

are the most effective scientific research method to test and isolate cause and effect relationships

after-only experiments are

monadic

before-after experiments are

sequential monadic

after-only experiment definition

just like a marketing experiment and a/b test. when you assign customers to either the test group or control group, show the new advertisement to the test group, show the old advertisement to the control group, and then measure the sales for each of the groups

before-after experiment definition

allows you to measure a metric goal of interest both before and after participants exposure to marketing inputs. would expose the test and control groups to the old advertising and the new advertising - measuring both the sales after each exposure

the main isolated relationship in a/b test

iv1

different reasons to using a coupon

iv1: the percentage discount

iv2: how discount is achieved

iv3: time of the year/holiday

iv4: digital/print

iv5: expiration date

iv6: frequency use of product

iv7: coupon usage simplicity

iv8: distribution location

conjoint analysis

manipulates a number of product attribute combinations to come up with an optimized product variation that consumers find most attractive —> measuring each product attributes importance level in purchase decision —> this often gets very interesting for expensive product categories

CLV stands for

customer lifetime value

CLV definition

the long term costs of keeping a customer vs the long term revenue potential of the customer for the company

B2B contracts

The intercept of a regression line equation specifies the value of “y” when “x”

is zero

The best way we have to isolate cause-and-effect relationships is to

run an experiment

Which of the following analytics creates a score that represents either a positive or negative overall emotion

text analytics

field experiments

conducted to test a campaign within a natural setting - rather than in a lab

web experiments

can be executed more quickly and cheaply than offline field experiments

you can manipulate more than one marketing input

natural experiments

a marketer observes the effect of certain naturally occurring incidents on customer behavior and other factors. need to find test and control groups

paid search advertising

technology of online auctions - they allow businesses to bid on the opportunity to put their ads in front of certain types of customer