HBR Guide to Data Analytics Basics for Managers

Data

Analytics

Basics for

Managers

Don’t let a fear of numbers

hold you back.

Understand the numbers

Make better decisions

Present and persuade

SMARTER THAN THE AVERAGE GUIDE

HBR

Guide to

Data Analytics Basics for Managers HBR Guide to

Stay informed. Join the discussion.

Visit hbr.org

Follow @HarvardBiz on Twitter

Find us on Facebook and LinkedIn

US$19.95 Management

Today’s business environment brings with it

an onslaught of data. Now more than ever,

managers must know how to tease insight from

data—to understand where the numbers come

from, make sense of them, and use them to

inform tough decisions. How do you get started?

Whether you’re working with data experts or

running your own tests, you’ll find answers in

the HBR Guide to Data Analytics Basics for

Managers. This book describes three key steps

in the data analysis process, so you can get

the information you need, study the data, and

communicate your findings to others.

You’ll learn how to:

• Identify the metrics you need to measure

• Run experiments and A/B tests

• Ask the right questions of your data experts

• Understand statistical terms and concepts

• Create effective charts and visualizations

• Avoid common mistakes

BONUS

ARTICLE

Data Scientist:

The Sexiest Job

of the 21st

Century

ISBN-13: 978-1-63369-428-6

9 7 8 1 6 3 3 6 9 4 2 8 6

9 0 0 0 0

HBR-Guide-DataAnalytics10185_Mechanical.indd 1 1/22/18 1:56 PM

HBR Guide to

Data Analytics

Basics for

Managers

H7353_Guide-DataAnalytics_2ndREV.indb i 1/17/18 10:47 AM

Harvard Business Review Guides

Arm yourself with the advice you need to succeed on the

job, from the most trusted brand in business. Packed

with how-to essentials from leading experts, the HBR

Guides provide smart answers to your most pressing

work challenges.

The titles include:

HBR Guide to Being More Productive

HBR Guide to Better Business Writing

HBR Guide to Building Your Business Case

HBR Guide to Buying a Small Business

HBR Guide to Coaching Employees

HBR Guide to Data Analytics Basics for Managers

HBR Guide to Delivering Effective Feedback

HBR Guide to Emotional Intelligence

HBR Guide to Finance Basics for Managers

HBR Guide to Getting the Right Work Done

HBR Guide to Leading Teams

HBR Guide to Making Every Meeting Matter

HBR Guide to Managing Stress at Work

HBR Guide to Managing Up and Across

HBR Guide to Negotiating

HBR Guide to Offi ce Politics

HBR Guide to Performance Management

HBR Guide to Persuasive Presentations

HBR Guide to Project Management

H7353_Guide-DataAnalytics_2ndREV.indb ii 1/17/18 10:47 AM

HBR Guide to

Data Analytics

Basics for

Managers

HARVARD BUSINESS REVIEW PRESS

Boston, Massachusetts

H7353_Guide-DataAnalytics_2ndREV.indb iii 1/17/18 10:47 AM

Copyright 2018 Harvard Business School Publishing Corporation

All rights reserved

No part of this publication may be reproduced, stored in or introduced

into a retrieval system, or transmitted, in any form, or by any means

(electronic, mechanical, photocopying, recording, or otherwise),

without the prior permission of the publisher. Requests for permission

should be directed to [email protected], or mailed to

Permissions, Harvard Business School Publishing, 60 Harvard Way,

Boston, Massachusetts 02163.

The web addresses referenced in this book were live and correct at the

time of the book’s publication but may be subject to change.

Cataloging-in-Publication data is forthcoming

eISBN: 9781633694293

HBR Press Quantity Sales Discounts

Harvard Business Review Press titles are available at signifi cant

quantity discounts when purchased in bulk for client gifts, sales

promotions, and premiums. Special editions, including books with

corporate logos, customized covers, and letters from the company

or CEO printed in the front matter, as well as excerpts of existing

books, can also be created in large quantities for special needs.

For details and discount information for both print and ebook for￾mats, contact [email protected], tel. 800-988-0886,

or www.hbr.org/bulksales.

H7353_Guide-DataAnalytics_2ndREV.indb iv 1/17/18 10:47 AM

What You’ll Learn

The vast amounts of data that companies accumulate to￾day can help you understand the past, make predictions

about the future, and guide your decision making. But

how do you use all this data effectively? How do you as￾sess whether your fi ndings are accurate or signifi cant?

How do you distinguish between causation and correla￾tion? And how do you present your results in a way that

will persuade others?

Understanding data analytics is an essential skill for

every manager. It’s no longer enough to hand this re￾sponsibility off to data experts. To be able to rely on the

evidence your analysts give you, you need to know where

it comes from and how it was generated—and what it

can and can’t teach you.

Using quantitative analysis as part of your decision

making helps you uncover new information and pro￾vides you with more confi dence in your choices—and

you don’t need to be deeply profi cient in statistics to

do it. This guide gives you the basics so you can better

understand how to use data and analytics as you make

tough choices in your daily work. It walks you through

H7353_Guide-DataAnalytics_2ndREV.indb v 1/17/18 10:47 AM

vi

What You’ll Learn

three fundamental steps of data analysis: gathering the

information you need, making sense of the numbers,

and communicating those fi ndings to get buy-in and

spur others to action.

You’ll learn to:

• Ask the right questions to get the information

you need

• Work more effectively with data scientists

• Run business experiments and A/B tests

• Choose the right metrics to evaluate predictions

and performance

• Assess whether you can trust your data

• Understand the basics of regression analysis and

statistical signifi cance

• Distinguish between correlation and causation

• Sidestep cognitive biases when making decisions

• Identify when to invest in machine learning—and

how to proceed

• Communicate and defend your fi ndings to

stakeholders

• Visualize your data clearly and powerfully

H7353_Guide-DataAnalytics_2ndREV.indb vi 1/17/18 10:47 AM

Contents

Introduction 1

Why you need to understand data analytics.

SECTION ONE

Getting Started

1. Keep Up with Your Quants 13

An innumerate’s guide to navigating

big data.

BY THOMAS H. DAVENPORT

2. A Simple Exercise to Help You Think

Like a Data Scientist 25

An easy way to learn the process of data

analytics.

BY THOMAS C. REDMAN

SECTION TWO

Gather the Right Information

3. Do You Need All That Data? 33

Questions to ask for a focused search.

BY RON ASHKENAS

H7353_Guide-DataAnalytics_2ndREV.indb vii 1/17/18 10:47 AM

Contents

viii

4. How to Ask Your Data Scientists for

Data and Analytics 37

Factors to keep in mind to get the

information you need.

BY MICHAEL LI, MADINA KASSENGALIYEVA,

AND RAYMOND PERKINS

5. How to Design a Business Experiment 45

Seven tips for using the scientifi c method.

BY OLIVER HAUSER AND MICHAEL LUCA

6. Know the Diff erence Between Your Data

and Your Metrics 51

Understand what you’re measuring.

BY JEFF BLADT AND BOB FILBIN

7. The Fundamentals of A/B Testing 59

How it works—and mistakes to avoid.

BY AMY GALLO

8. Can Your Data Be Trusted? 71

Gauge whether your data is safe to use.

BY THOMAS C. REDMAN

SECTION THREE

Analyze the Data

9. A Predictive Analytics Primer 81

Look to the future by looking at the past.

BY THOMAS H. DAVENPORT

10. Understanding Regression Analysis 87

Evaluate the relationship between variables.

BY AMY GALLO

H7353_Guide-DataAnalytics_2ndREV.indb viii 1/17/18 10:47 AM

Contents

ix

11. When to Act On a Correlation,

and When Not To 103

Assess your confi dence in your fi ndings and

the risk of being wrong.

BY DAVID RITTER

12. Can Machine Learning Solve Your

Business Problem? 111

Steps to take before investing in artifi cial

intelligence.

BY ANASTASSIA FEDYK

13. A Refresher on Statistical Signifi cance 121

Check if your results are real or just luck.

BY AMY GALLO

14. Linear Thinking in a Nonlinear World 131

A common mistake that leads to errors

in judgment.

BY BART DE LANGHE, STEFANO PUNTONI,

AND RICHARD LARRICK

15. Pitfalls of Data-Driven Decisions 155

The cognitive traps to avoid.

BY MEGAN MacGARVIE AND KRISTINA McELHERAN

16. Don’t Let Your Analytics Cheat the Truth 165

Pay close attention to the outliers.

BY MICHAEL SCHRAGE

SECTION FOUR

Communicate Your Findings

17. Data Is Worthless If You Don’t Communicate It 173

Tell people what it means.

BY THOMAS H. DAVENPORT

H7353_Guide-DataAnalytics_2ndREV.indb ix 1/17/18 10:47 AM

Contents

x

18. When Data Visualization Works—and

When It Doesn’t 177

Not all data is worth the eff ort.

BY JIM STIKELEATHER

19. How to Make Charts That Pop and Persuade 183

Five questions to help give your numbers

meaning.

BY NANCY DUARTE

20. Why It’s So Hard for Us to Communicate

Uncertainty 191

Illustrating—and understanding—the

likelihood of events.

AN INTERVIEW WITH SCOTT BERINATO

BY NICOLE TORRES

21. Responding to Someone Who Challenges

Your Data 199

Ensure the data is thorough, then make

them an ally.

BY JON M. JACHIMOWICZ

22. Decisions Don’t Start with Data 205

Infl uence others through story and emotion.

BY NICK MORGAN

Appendix: Data Scientist: The Sexiest Job

of the 21st Century 209

BY THOMAS H. DAVENPORT AND D.J. PATIL

Index 225

H7353_Guide-DataAnalytics_2ndREV.indb x 1/17/18 10:47 AM

1

Introduction

Data is coming into companies at remarkable speed and

volume. From small, manageable data sets to big data

that is recorded every time a consumer buys a product or

likes a social media post, this information offers a range

of opportunities to managers.

Data allows you to make better predictions about the

future—whether a new retail location is likely to suc￾ceed, for example, or what a reasonable budget for the

next fi scal year might look like. It helps you identify the

causes of certain events—a failed advertising campaign,

a bad quarter, or even poor employee performance—so

you can adjust course if necessary. It allows you to iso￾late variables so that you can identify your customers’

wants or needs or assess the chances an initiative will

succeed. Data gives you insight on factors affecting your

industry or marketplace and can inform your decisions

about anything from new product development to hiring

choices.

H7353_Guide-DataAnalytics_2ndREV.indb 1 1/17/18 10:47 AM

Introduction

2

But with so much information coming in, how do you

sort through it all and make sense of everything? It’s

tempting to hand that role off to your experts and ana￾lysts. But even if you have the brightest minds handling

your data, it won’t make a difference if you don’t know

what they’re doing or what it means. Unless you know

how to use that data to inform your decisions, all you

have is a set of numbers.

It’s quickly becoming a requirement that every deci￾sion maker have a basic understanding of data analyt￾ics. But if the thought of statistical analysis makes you

sweat, have no fear. You don’t need to become a data

scientist or statistician to understand what the numbers

mean (even if data scientists have the “sexiest job of the

21st century”—see the bonus article we’ve included in the

appendix). Instead, you as a manager need a clear un￾derstanding of how these experts reach their results and

how to best use that information to guide your own deci￾sions. You must know where their fi ndings come from,

ask the right questions of data sets, and translate the re￾sults to your colleagues and other stakeholders in a way

that convinces and persuades.

This book is not for analytics experts—the data sci￾entists, analysts, and other specialists who do this work

day in, day out. Instead, it’s meant for managers who

may not have a background in statistical analysis but still

want to improve their decisions using data. This book

will not give you a detailed course in statistics. Rather,

it will help you better use data, so you can understand

what the numbers are telling you, identify where the re￾sults of those calculations may be falling short, and make

stronger choices about how to run your business.

H7353_Guide-DataAnalytics_2ndREV.indb 2 1/17/18 10:47 AM

Introduction

3

What This Book Will Do

This guide walks you through three key areas of the data

analytics process: gathering the information you need,

analyzing it, and communicating your fi ndings to oth￾ers. These three steps form the core of managerial data

analytics.

To fully understand these steps, you need to see the

process of data analytics and your role within it at a high

level. Section 1, “Getting Started,” provides two pieces

to help you digest the process from start to fi nish. First,

Thomas Davenport outlines your role in data analysis

and describes how you can work more effectively with

your data scientist and become a better consumer of

analytics. Then, you’ll fi nd an easy exercise you can do

yourself to gather your own data, analyze it, and identify

what to do next in light of what you’ve discovered.

Once you have this basic understanding of the pro￾cess, you can move on to learn the specifi cs about each

step, starting with the data search.

Gather the right information

For any analysis, you need data—that’s obvious. But

what data you need and how to get it can be less clear

and can vary, depending on the problem to be solved.

Section 2 begins by providing a list of questions to ask

for a targeted data search.

There are two ways to get the information you need:

by asking others for existing data and analysis or by run￾ning your own experiment to gather new data. We ex￾plore both of these approaches in turn, covering how to

request information from your data experts (taking into

H7353_Guide-DataAnalytics_2ndREV.indb 3 1/17/18 10:47 AM

Introduction

4

account their needs and concerns) and using the scien￾tifi c method and A/B testing for well-thought-out tests.

But any data search won’t matter if you don’t measure

useful things. Defi ning the right metrics ensures that

your results align with your needs. Jeff Bladt, chief data

offi cer at DoSomething.org, and Bob Filbin, chief data

scientist at Crisis Text Line, use the example of their own

social media campaign to explain how to identify and

work toward metrics that matter.

We end this section with a helpful process by data ex￾pert and company adviser Thomas C. Redman. Before

you can move forward with any analysis, you must know

if the information you have can be trusted. By following

his advice, you can assess the quality of your data, make

corrections as necessary, and move forward accordingly,

even if the data isn’t perfect.

Analyze the data

You have the numbers—now what? It’s usually at this

point in the process that managers fl ash back to their

college statistics courses and nervously leave the analy￾sis to an expert or a computer algorithm. Certainly, the

data scientists on your team are there to help. But you

can learn the basics of analysis without needing to un￾derstand every mathematical calculation. By focusing on

how data experts and companies use these equations (in￾stead of how they run them), we help you ask the right

questions and inform your decisions in real-world man￾agerial situations.

We begin section 3 by describing some basic terms

and processes. We defi ne predictive analytics and how to

H7353_Guide-DataAnalytics_2ndREV.indb 4 1/17/18 10:47 AM