Notes

**1. Which of the following are obstacles to data-driven marketing?**
Multiple campaigns and time delay make determining cause and effect difficult.

**2. For data-driven marketing, it is easier to obtain data from B2B customers than from B2C customers.**
FALSE

3. In Chapter 3, the Data-Driven Marketing book discusses a four-stage marketing behavior impact model (i.e., Awareness, Evaluation, Trial, and Loyalty). Connect the stages of the model to the corresponding metrics.
Awareness - Brand Awareness
Evaluation - Test-Drive
**Trail - Take Rate**
Loyalty - Churn

**4. According to the textbook “Data-Driven Marketing”, which of the following metrics are used to measure marketing ROI?**
NONE OF THE ABOVE

5. What percent of companies do not calculate marketing ROI?
About half of all companies do not calculate marketing ROI.

6. Metric #4 Customer satisfaction (Net Promoter Score) is important because:
It predicts the future growth of the company

7. The Net Promoter Score can have negative values.
TRUE

8. From the ROMI framework, which stage is the most challenging and requires the most assumptions:
Upside

9. What is the role of sensitivity analysis in ROMI?
To evaluate how changes in initial assumptions affect the projected financial outcome of a marketing campaign.

10. Experiments conducted within a natural setting of customers, who typically are unaware of the experiments, are called:
Field experiments

**11. Conjoint analysis most often is used:**
To determine the relative importance of product/service attributes

**12. In marketing campaigns, cause and effect can be proved by:**
Experiments

13. Customer Lifetime Value represents:
The expected cash flow from a customer during their tenure with a company.

14. Customer Lifetime Value can be used to value a company.
TRUE

**15. In predictive analytics, the modeling techniques can be classified as:**
Classification and prediction

**16. A company wants to predict which customers will accept an offer. The dependent variable that has to be predicted is in a Yes/No format, and all independent variables are metric. Which predictive models could be used for that purpose?**
Discriminant Analysis
Logistic Regression
Decision Tree

**17. The independent variables in Decision Tree model can be numerical and categorical?**
True

18. In RapidMiner, what does the Apply Model operator do:
It predicts the dependent variable and creates a new variable in the data set with the newly predicted values.

19. In RapidMiner, which of the following is not a variable role?
Numeric

26: What other predictive models can be trained on the insurance dataset?
Discriminant Analysis
Logistic Regression
Neural Net