Predictive Analytics for Marketers

Using Data Mining for Business Advantage

Barry Leventhal

• Publisher: Kogan Page
• Year: 2018
• ISBN:
  9780749479947
• Edition: 1
• Topic: Business
• Subtopic: Marketing
• Index: Business

Predictive analytics has revolutionized marketing practice. It involves using many techniques from data mining, statistics, modelling, machine learning and artificial intelligence, to analyse current data and make predictions about unknown future events. In business terms, this enables companies to forecast consumer behaviour and much more. Predictive Analytics for Marketers will guide marketing professionals on how to apply predictive analytical tools to streamline business practices. Including comprehensive coverage of an array of predictive analytic tools and techniques, this book enables readers to harness patterns from past data, to make accurate and useful predictions that can be converted to business success. Truly global in its approach, the insights these techniques offer can be used to manage resources more effectively across all industries and sectors. Written in clear, non-technical language, Predictive Analytics for Marketers contains case studies from the author's more than 25 years of experience and articles from guest contributors, demonstrating how predictive analytics can be used to successfully achieve a range of business purposes.

06

Predicting customer behaviour using analytical models

Introduction

Earlier chapters have discussed the process, data, techniques and tools for predictive analytics. From this chapter onwards, we examine how these resources are applied in various industries and domains. The most common application is for predicting customer behaviour, which is the main focus of this chapter.

The main aims of this chapter are to:

• Consider the use of predictive models for helping to manage the customer journey.
• Review critical elements that need to be designed when planning a new model that will deliver value for your business.
• Walk through all the steps required for building and implementing predictive models.
• Show how model performance can be evaluated using lift and gains charts.

Overview – building and deploying predictive models

The use of predictive models goes back to the early days of direct marketing, when companies started to record customers’ responses and transactions, and used the data to target their marketing campaigns.

The data-mining process, reviewed in Chapter 2, translates into a series of main steps for building and deploying targeting models. These steps are summarized in Figure 6.1 and are discussed in the sections that follow.

Figure 6.1 How targeting models are built and deployed

How targeting models are built and deployed

Managing the customer journey

The customer journey is the series of interactions that each customer makes with your business, from becoming aware of your products or services, through to making a first purchase and eventually to leaving. This journey is liable to be long and complex, and will be different for each individual. Many companies find it helpful to map the steps in the journey, in order to help identify how to improve their customer-facing processes. This, in turn, should lead to more satisfied customers, who will stay with your company for longer, purchase more and be more profitable.

Predictive models can help to target resources for managing key stages of the customer journey – for example:

• Customer recruitment may be targeted, in order to maximize your return on investment. This applies irrespective of how your company recruits customers; so, for example, targeting may be used for direct marketing, online ads and newspaper advertising.
• Once a customer has been recruited, ie purchased an initial product or servi...