AI with Python Lecture for Data Science domain

AI with Python

1. Marketing Analytics | 3 rd Edition | Mike Grigsby LECTURE SLIDES FOR CHAPTERS 1 AND 2 A BRIEF INTRODUCTION TO: CONSUMER BEHAVIOUR MARKETING STRATEGY STATISTICS

2. In marketing the consumer is central . This is the marketing concept : Show the consumer's problem What are the consumer’s preferences (in terms of goods / services)? What are the consumer’s constraints (allocating limited budgets)? Given limited resources, what are consumer’s choices? Therefore the consumer’s problem is: “How to maximize satisfaction given limited budget.” Assumptions based on above: Preferences are complete, consumers can rank and compare all products Preferences are transitive Products are desirable (a good is good, or of value), more is better A brief introduction to Consumer Behaviour

3. Show the consumer decision process Extended problem solving vs. limited problem solving. How does marketing tie into each stage? A brief introduction to Consumer Behaviour

4. Classifying Consumer Behaviour Models Stage Model Type Need Recognition Discrete Binary Choice Search for Information Awareness / Consideration Information Processing Perceptual Mapping Pre-Purchase Alternative Evaluation Attitude Models / Touchpoints Conjoint Purchase Discrete Binary Choice Post Purchase Evaluation Satisfaction

5. Show general marketing strategy Michael Porter Four factors determining competitive intensity => rivalry among firms: Bargaining power of suppliers Bargaining power of buyers Threat of new entrants Threat of substitute products Three generic strategies: Be the low cost provider to the whole market Differentiate and focus on high end products for the whole market Segment and target profitable / fast growing parts of the market Treacy / Weirsema Three generic strategies: Operational excellence Product leadership Customer intimacy A brief introduction to Marketing Strategy

6. Sorger’s offensive / defensive actions, tied back to consumer behaviour Defensive Reactions to Competitive Moves Bypass Attack (the attacking firm expands into one of our product areas) Correct counter is for us to constantly explore new areas, aiming at customer’s market basket, “Who will buy wider?” Encirclement Attack (the attacking firm tries to overpower us with larger forces) Correct counter is for us to message how our products are superior / unique providing more value Flank Attack (the attacking firm tries to exploit our weaknesses) Correct counter is to not have any weaknesses! Monitor / research customer’s via early warning systems / satisfaction trackers, etc. Frontal Attack (the attacking firm aims at our strength) Correct counter is to attack back in the firm’s territory A Brief Introduction to Marketing Strategy

7. Sorger’s offensive / defensive actions, tied back to consumer behaviour Offensive Actions Against Competitors New market segments Using behavioural segmentation incents prospects for a win-win Go-to-market approaches Gains insights about prospects preferences in bundling, channels, price sensitivity, etc. Differentiating functionality Ascertains prospects needs by offering product / purchase combinations most compelling A brief introduction to Marketing Strategy

8. Three uses of data/analysis: Observations and variables Two kinds of thinking: inductive and deductive Descriptive statistics—one variable Three measures of central tendency: Mean Median Mode The average is not necessarily the mean Three measures of spread / dispersion Range Variance Standard deviation A brief introduction to Statistics Descriptive statistics - two variables Covariance Correlation R 2

9. The normal distribution Characteristics of a normal distribution 1) Mean = median = mode and 2) distribution is symmetrical (bell shaped) Percentages under the curve Z-score : tells the number of standard deviations an observation is Z = (Xi – X-bar) / s We will Always be At 95% confidence , 1.96 z-score Confidence interval Standard deviation estimated by standard error CI = mean +/- [confidence level * (standard deviation / SQ ROOT(count)] SAMPLE POPULATION Mean X-bar μ Standard Deviation s σ A brief introduction to Statistics

10. Consider home sales prices via the figure below. The mean is 141k but the median is 110k. Which number is most representative?

11. 95% confidence is 1.96 SD

12. SAS has five windows: 1. The Editor window 2. the Log window 3. the Output window 4. the Results window 5. the Explorer window The contents of the log and output windows cannot be edited. SAS statements fall into four broad categories: 1. Data Step Statements 2. Proc (procedure) step statements 3. Statements that can be used in both proc and data steps 4. Global statements that apply to all following steps. All SAS statements end with a semicolon. A run statement should appear (but is not generally required) at the end of every step. Variables in SAS have names which must be defined. The colours have meaning. A brief introduction to SAS Two-level names (LOCATION.FILE). LIBNAME tells the location.

13. PROC CONTENTS PROC MEANS PROC FREQ PROC UNIVARIATE PROC CORREL Merge datasets (xx and yy ) by cust_id PROC SORT zz.xx by cust_id ; PROC SORT zz.yy by cust_id ; DATA zz.zz ; MERGE zz.xx zz.yy ; BY cust_id ; RUN;