The Average is Always Wrong

A real-world guide to putting data at the heart of your business

Ian Shepherd

• Publisher: Harriman House
• Year: 2020
• ISBN:
  9780857198136
• Topic: Negocios y empresa
• Subtopic: Comunicación empresarial

Everywhere you look people are talking about data. Buzzwords abound – 'data science', 'machine learning', 'artificial intelligence'. But what does any of it really mean, and most importantly what does it mean for your business? Long-established businesses in many industries find themselves competing with new entrants built entirely on data and analytics. This ground-breaking new book levels the playing field in dramatic fashion.The Average is Always Wrong is a completely pragmatic and hands-on guide to harnessing data to transform your business for the better.Experienced CEO and CMO Ian Shepherd takes you behind the jargon and puts together a powerful change programme anyone can enact in their business right now, to reap the rewards of simple but sophisticated uses of data.Filled with practical examples and case studies, readers will come away with a powerful understanding of the real value of data and the analytical techniques that can drive profit growth.

Part One: Data analysis that drives profit

Chapter 1: A few useful concepts

In this chapter, we will explore more about how to get beyond simple averages and see more examples of the business benefits that accrue when we do. Along the way, we’ll pick up some terminology that will help us on our journey in later chapters.

The language of data

Before we dive headlong into the business benefits that can be derived from data, we need to agree some language, so that we can be sure exactly what we are talking about.

Like many topics clamouring for the attention of business leaders, data is awash with terminology. Let’s begin, then, by defining some straightforward and practical terms:

• By data, we mean (perhaps obviously) the set of things you can know about your business. That might be a fact about a customer (their address), something a customer has done (a purchase they have made from you), a fact about your business (the number of units of a particular product in an individual store) or any number of other things. Each of these facts is a data point and together they represent a tiny insight into your business, your customers and your wider network of partners and suppliers. Businesses today are swimming in data – if only they can harness it, understand it and extract value from it.
• A database is where you put that data in order to be able to ask questions of it. There are obviously a variety of different technical solutions and platforms available to do that. We won’t concern ourselves with much of that technical detail in this book, but we will explore some of the ways that a database can be used and the kinds of questions that we can ask it. When you have a very large amount of data, some of the underlying technologies to handle and process the data are different, which is when you’ll start to hear people talk about big data – again, from our point of view, this is underlying technology that we don’t need to worry too much about
• Data science is a grand term for the suite of analytical tools that you could apply to extract knowledge and insight from your data – we’ll see lots of practical examples in this book. You’ll also see the phrase data mining used in a similar context. There are many techniques that fit under the general banner of data science. One key split which we will explore in this book is between supervised approaches, where you are using data science to try to answer a specific question (predict future behaviours, for example), and unsupervised approaches, where you are simply trying to understand the data by breaking it, for instance, into segments or clusters. We’ll see real-world examples of both.
• Machine learning is a specialist area of data science that uses computers to fit models to your data, building prediction engines that attempt to tell you (as an example) which of your products a specific customer might be most interested in buying. You can think of machine learning as a subset of artificial intelligence (AI), which is the broad topic of building computer programs that can learn from data. You’ll hear people at conferences talk a lot about AI, but it is also an often misused and over-hyped term. Machine learning is really the subset of AI that has most relevance to your journey to becoming a data-centric business (though there are other aspects of AI that are being used to automate warehouses and add value to businesses in other ways too).
• A neural network is a specific machine-learning model which is also used in analysing data, particularly where the data set and the question you are trying to answer are quite complex. A neural network is so-called because it uses an algorithm that resembles the way neurons in our brains are interconnected. Unfortunately, that gives a powerful data-science technique a name that is far too sexy, meaning it gets over-used at those business conferences a lot. In fact, it is just one among many analytic techniques that a data scientist might apply to your data, depending on the problem you are trying to solve, and often will not be the best one to use.
• And finally, data-centric business is a term I’ve made up for this book to represent our end goal. Those businesses which are really brilliant at mining insights out of their data have a set of common characteristics and capabilities and we will discover them as we go through the coming chapters.

Buzzword bingo is a championship sport in this area, as you can see. As we go through the book we’ll try to stick to a small set of technical terms and instead use the old-fashioned technique of describing things in English. If someone tries to bamboozle you about how they are using ‘algorithms deployed on a cloud-based data lake’, then challenge them to do the same (it means they are doing some data science stuff on a big database).

Why the average is always wrong

Over the years, I’ve seen businesses waste huge amounts of time and money chasing opportunities that simply don’t exist, while missing out completely on ones that do.

For example, remember the marketing presentation from the introduction that told you your customers visit on average 2.3 times a month? We’ve already explored the ways in which that figure might be misleading. But that’s not the end of the story. Presentations like that often move on to describe your customers demographically. You might hear that your customers tend to be older than average, live in the north west and have pets.

But how true is that really? If 10% of the population live in the north west and 15% of your customers do then it is true that your customers are more likely to live there, but it is also true that 85% of them don’t.

It’s too easy, though, when a table of numbers or a graph is presented to take it at face value. Before you know it, the management conversation is dominated by a pen-sketch of dog-owning Cumbrian pensioners even though they are a minority of the customer base.

We need a way to see through these summary (and misleading) statistics and get properly under the skin of what our data is really telling us about our businesses.

Meet the blobs

We are going to analyse a hypothetical retail business. It has customers, of c...

Table of contents

• Contents
• Acknowledgements
• Introduction
• Part One: Data analysis that drives profit
• Part Two: Valuable data and where to find it
• Part Three: Building the data-centric business
• Conclusion
• Publishing details