Why is thinking without data the key to making better decisions?

Professor Bart De Langhe presents his new book 'Decision-driven analytics'

Say you want to reduce churn. You have lots of data on your customers’ spending history and you’ve invested in a sophisticated churn prediction model that generates a shortlist of customers most likely to churn. You decide to target only these customers with promotional incentives. That’s what best practice tells you to do. After all, why waste money on customers who aren’t going anywhere, right? Wrong. In Decision-Driven Analytics, professors Bart De Langhe and Stefano Puntoni explain why. What’s more, they challenge the conventional wisdom of data-driven decision-making in a world where big data is king.

web-insights-bart-de-langhe-decision-driven-analytics

We are on the brink of a new revolution powered by AI. Never before have businesses had so much data at their disposal. They have invested in increasingly sophisticated tools to analyse that data. “But many executives will tell you they are disappointed with their investments”, says Bart. Why is that?

video-thumbnail-bart-de-langhe-decision-driven-analytics

Leveraging human intelligence to unlock the power of data

Professor Bart De Langhe presents his new book 'Decision-driven analytics'. The book which he wrote together with Professor Stefano Puntoni proposes a radically different approach to data analytics, starting with the decisions and then working backwards to the data. Instead of focusing on data, technology and statistics, their book focuses on decisions, people and psychology.

Divers aren’t runners

Many organisations fail to extract value from their data because there is a gap between those who analyse it and those who make decisions. Bart explains: “To simplify things, you can think of organisations as having two types of people: divers and runners. Divers love a good data set and are always eager to dive in, unleash their statistical models and try to extract insights. Runners, on the other hand, feel the pressure of time and need quick insights to make decisions.”

The problem, he points out, is the divide between these groups, which means that despite significant investment in data and tools, many analytics efforts fail to effectively influence decision-making. “In other words, even though organisations are investing heavily in data analytics, decisions are still being made without taking full advantage of the insights that data can provide”, he concludes.

Decision-driven instead of data-driven

Data-driven decision-making doesn’t seem to be delivering on its promises. So Bart and Stefano suggest that the data should take a back seat, which may seem counterintuitive in a book about analytics. “We propose a radically different approach, decision-driven analytics, starting with the decisions and then working backwards to the data. Instead of focusing on data, technology and statistics, our book focuses on decisions, people and psychology.”

Given the authors’ backgrounds, this is not surprising: Bart trained as a psychologist and later turned to statistics, while Stefano started out as a data scientist before switching to psychology. Their paths crossed halfway when they both became business school academics. This book draws on their extensive experience of teaching in executive programmes at prestigious institutions and advising companies around the world. As to why they decided to write it, Bart says: “Over the years, we’ve refined our approach and now feel it’s the right time to share our framework with a broader audience. Our goal in writing this book is simple: we want organisations to look at data analytics differently. We hope people will pick it up, read it, and use it as a catalyst for positive change in how organisations use data to make decisions.”

From decisions to answers

Decision-driven analytics strategically combines human intelligence and judgement with advanced data analytics to meet an organisation’s decision-making needs. It relies on four sequential steps, each of which is discussed in a separate chapter of the book.

  • Decisions: Decision-driven analytics starts with identifying decision alternatives. What are the different options or courses of action? “And it’s not about casting as wide a net as possible,” says Bart, “it’s about focusing on those options that are feasible and impactful."
  • Questions: The second step is to formulate questions. What are the questions you need to answer in order to make a more informed choice between the different options? Bart’s advice is simple: “Ask the right questions and be as precise as possible, because ambiguous questions can lead to misunderstandings and bad decisions.”  Why you need to ask the right questions is perfectly illustrated by the churn example mentioned earlier. The churn prediction model tells you how likely a particular customer is to churn. But in order to spend your promotions wisely, the question you should be answering instead is: What is the effect of our intervention on the likelihood of a particular customer churning? “And you won’t find the answer to that question in historical data on their spend”, says Bart. So what data should you collect? He smiles. “I don't want to spoil it for those who want to read the book. But I’ll tell you this: best practices might actually make likely churners run even faster.”
  • Data: Only in a third step should you focus on data. What data do you need to collect and analyse to answer the questions you identified in the previous step? Bart warns that decision-driven analytics doesn’t worship big data, it requires relevant data. “You’ll often find that the data you have is not the data you need”, he says. “How your data is produced is as important, if not more important, than the data itself.”
  • Answers: The final step in a decision-driven analytics approach is answers. And unlike the questions, which should be unambiguous, here you need to acknowledge uncertainty. As Bart stresses: “Don’t make these answers too precise, because precise answers are often precisely wrong.”

Means to an end

The benefits for organisations using the decision-driven analytics framework are twofold. “First, it increases the effectiveness of decision-making”, explains Bart. “By using our framework, organisations will analyse the right data and make better decisions as a result. Second, it increases the efficiency of decision-making. As I’ve said before, many organisations invest heavily in data analytics technology, but often fail to see that translate into effective decisions. Our framework reduces this inefficiency.”

If there is one message readers should take away from this book, it is this: data analytics is a means to an end. That end? Making better decisions. “And to get there”, he says, “you need to find the right data. Nowadays, many organisations have a lot of data and all too often they try to find a purpose for it. We think that’s a mistake. Instead, you should try to find data for a purpose.”

Food for thought

While the book presents a fresh perspective on data analytics, it is far from theoretical. Packed with real-world examples from industries as diverse as financial services, technology, pharmaceuticals, consumer goods, e-commerce, and automotive, it also covers various business functions such as customer relationship management, company and brand valuation, and digital advertising. It includes examples from non-business areas such as vaccine development and political campaigns as well. And that’s not all: it’s teeming with delightful anecdotes and little nuggets of wisdom. Still, at just over 100 pages, it’s a short, concise read in accessible language, free of jargon. “You can read it anywhere in just a few hours”, assures Bart, “but we hope it will leave you thinking for many more.”

For novices and experts

Bart and Stefan have written the book for everyone who makes decisions with data. From the novice who wants to better understand and use data to make better decisions, to the expert who is disappointed with how their analysis is being used and wants to improve communication with decision makers, this is the book for you.

So, what does the London Whale Trade have to do with data-driven analytics? Why is Meta's sweeping claim about the effectiveness of personalised advertising fundamentally flawed? What are the dangers of categorical thinking? And why is the fact that vaccinated people are ending up in hospital in the middle of the Covid-19 pandemic not necessarily a sign that vaccines are losing their effectiveness? There is only one way to find out...

Want to read more?
Decision-Driven Analytics – Leveraging Human Intelligence to Unlock the Power of Data is published by Wharton School Press an imprint of University Of Pennsylvania Press. You can also order the book from Amazon.

About the authors

  • Bart De Langhe is Professor of Marketing at Vlerick Business School and KU Leuven, Belgium. He is also the founder of Behavioral Economics and Data Analytics for Business (BEDAB), a consultancy that helps companies use behavioural science and data analytics to make better decisions and improve performance.
  • Stefano Puntoni is the Sebastian S. Kresge Professor of Marketing at the Wharton School, University of Pennsylvania, USA. He is also co-director of AI at Wharton, an initiative to foster, coordinate and promote research and teaching on artificial intelligence across the Wharton School.

Get in touch!

Bart De Langhe

Bart De Langhe

Professor of Marketing