Marketing Analytics

A Practical Guide to Improving Consumer Insights Using Data Techniques

Mike Grigsby

• Éditeur: Kogan Page
• Année: 2018
• ISBN:
  9780749482176
• Édition: 2
• Sujet: Business
• Sous-sujet: Marketing Research

Who is most likely to buy and what is the best way to target them? How can businesses improve strategy without identifying the key influencing factors? The second edition of Marketing Analytics enables marketers and business analysts to leverage predictive techniques to measure and improve marketing performance. By exploring real-world marketing challenges, it provides clear, jargon-free explanations on how to apply different analytical models for each purpose. From targeted list creation and data segmentation, to testing campaign effectiveness, pricing structures and forecasting demand, this book offers a welcome handbook on how statistics, consumer analytics and modelling can be put to optimal use. The fully revised second edition of Marketing Analytics includes three new chapters on big data analytics, insights and panel regression, including how to collect, separate and analyze big data. All of the advanced tools and techniques for predictive analytics have been updated, translating models such as tobit analysis for customer lifetime value into everyday use. Whether an experienced practitioner or having no prior knowledge, methodologies are simplified to ensure the more complex aspects of data and analytics are fully accessible for any level of application. Complete with downloadable data sets and test bank resources, this book supplies a concrete foundation to optimize marketing analytics for day-to-day business advantage.

PART ONE

Overview

How can marketing analytics help you?

01

A brief statistics review

Measures of central tendency 9

Measures of dispersion 11

The normal distribution 14

Confidence intervals 15

Relations among two variables: covariance and correlation 16

Probability and the sampling distribution 18

Conclusion 18

Checklist: You’ll be the smartest person in the room if you … 19

You knew we had to do this, have a general review of basic statistics. I promise, it’ll be mostly conceptual, a gentle reminder of what we learned in Introductory Statistics. Also note the definition boxes helping to describe key terms, point out jargon, etc.

Measures of central tendency

First we’ll deal with simple descriptive statistics, confined to one variable. We’ll start with measures of central tendency.

Measures of central tendency include the mean, median and mode.

Mode: the number that appears most often.

Average is the most representative number. Of course it doesn’t help this argument that Excel uses =AVERAGE() as the function to calculate the mean instead of =MEAN(). I’ve tried asking Bill about it but he’s not returned my calls, so far.

Measures of dispersion

Measures of central tendency alone do not adequately describe the variable (a variable is a thing that varies, like home sales prices). The other dimension of a variable is dispersion, or spread.

There are three measures of dispersion: range, variance and standard deviation.

Range: a measure of dispersion or spread, calculated as the maximum value less the minimum value.

Range is easy. It’s simply the minimum (smallest value) observation subtracted from the maximum (largest value). It’s not particularly useful, especially in a marketing context.

Variance is another measure of dispersion or spread.

Variance: a measure of spread, calculated as the summed square of each observation less the mean, divided by the count of observations less one.

Conceptually it takes each observation and subtracts the mean of all the observations from it, then squares each observation and adds up the squares. That quantity is divided by n–1, the total number of observations, less one. The formula is below. Note this is the sample formula, not the formula for the population.

(Note that X-bar is the symbol for sample mean, while µ would be the symbol to use for population mean; s would be the symbol to use for sample standard deviation and σ would be the symbol to use for population standard deviation.)

Now, what does variance tell us? Unfortunately, not much. It says that (from Table 1.1) this variable of 18 observations has a mean of 25 and a variance, or spread, of 173.6. But variance gets us to the standard deviation, which DOES mean something.

Table 1.1 Variance

X	X-mean	Squared
2	–23	529.3
5	–20	400.3
8	–17	289.2
10.9	–14.1	199.3
13.9	–11.1	123.6
16.9	–8.1	65.9
19.9	–5.1	26.2
22.9	–2.1	4.5
25.9	0.9	0.8
28.9	3.9	15.1
31.9	6.9	47.4
33	8	63.9
34	9	80.9
35	10	99.9
36	11	120.9
39	14	195.8
42	17	288.8
45	20	399.7
Mean = 25.0		Sum = 2,951.5
Count = 18		Variance = 173.6

Standard deviation: the square root of variance.

Standard deviation is calculated by taking the square root of variance. In this case the square root of 173.6 is 13.17. Now, what does 13.17 mean? It describes spread or dispersion in a way that removes the scale of the variable. That is, there are known qualities of a standard deviation. In a fairly normal distribution dispersion is spread around the mean (which equals the mode which equals the median). That is, there is a symmetrical spread around the mean of 25. In this case the spread is 25 +/– 13.17. That means that, in general, one standard deviation (+/– 13.17) from the mean will contain 68% of...