Analytics the Right Way

Book Overview

This book argues that most organizations get little value from their data not because of weak tools but because of muddled thinking about what data is for. Tim Wilson, an analytics practitioner since the early 2000s, and Joe Sutherland, an executive, educator and former White House staffer, set out a business-first approach: start from the decision or question, decide how much evidence it deserves, and only then reach for data and technology.

Published by Wiley in January 2025 as a 256-page softcover, the book is written for business leaders and managers, not analysts writing code. It treats data work as having three distinct uses: measuring performance against targets, validating hypotheses about what will improve results, and enabling operations through automated data processes such as machine learning. Much of the book explains why these uses get confused and how to keep them apart.

Key Topics Covered

• Common misconceptions about data that lead organizations astray
• Making decisions under uncertainty, and what separates a good decision from a lucky outcome
• An introduction to causality through the potential outcomes idea
• A structured framework for the different ways data creates value
• Performance measurement: choosing KPIs and setting targets for them
• Hypothesis validation: articulating testable ideas and keeping a record of them
• Matching the strength of evidence (anecdotal, descriptive, scientific) to the stakes of a decision
• Pitfalls of historical data analysis and of statistical inference
• Operational enablement, including where machine learning and AI fit
• Communication, and the role of technology once the thinking is right

How to Use This Book

This is a first read, best taken in order. The early chapters build the case for thinking about decisions and uncertainty; the framework in the middle chapters only makes sense once that foundation is in place. The argument gains momentum in the chapters on performance measurement and hypothesis validation, where the ideas become directly usable in meetings and planning. The later chapters on evidence and pitfalls are the ones you are most likely to revisit.

Who Should Read This

The book suits business leaders, managers, marketers and product owners who commission analytics and want better results from it, and analysts who need language and structure for pushing back on vague requests. No technical background is required.

It is not a how-to book for building dashboards, writing SQL or running statistical tests. Readers looking for technical skills should pair it with a hands-on course.

Pros and Cons

Pros

• Clear, decision-first framework that applies across industries
• Separates performance measurement, hypothesis testing and operational data work in a practical way
• Explains causality and uncertainty for non-technical readers
• Short enough to finish quickly

Cons

• Little technical detail, so analysts looking for methods will want more depth
• Some ideas are repeated across chapters
• Examples lean toward marketing and digital analytics
• Readers already familiar with experimentation and causal inference will find the statistics chapters basic

Frequently Asked Questions

Is this book for analysts or managers? Mainly for business leaders and managers, though analysts will find its framework useful when working with stakeholders.

Does it teach statistics? It explains the ideas behind uncertainty, evidence and causality, but not calculations or code.

How does it compare with Storytelling with Data? Storytelling with Data focuses on presenting results. This book focuses on deciding what to analyze and how much evidence a decision needs.

Do I need a technical background? No. It is written for a general business audience.

See the current formats and price on the publisher's page if this decision-first approach matches what your team needs.