Why Data Science Projects Fail: The Harsh Realities of Implementing AI and Analytics, without the Hype

by Douglas Gray, Evan Shellshear

The field of artificial intelligence, data science, and analytics is crippling itself. Exaggerated promises of unrealistic technologies, simplifications of complex projects, and marketing hype are leading to an erosion of trust in one of our most critical approaches to making data driven.

This book aims to fix this by countering the AI hype with a dose of realism. Written by two experts in the field, the authors firmly believe in the power of mathematics, computing, and analytics, but if false expectations are set and practitioners and leaders don’t fully understand everything that really goes into data science projects, then a stunning 80% (or more) of analytics projects will continue to fail, costing enterprises and society hundreds of billions of dollars, and leading to non-experts abandoning one of the most important data-driven decision-making capabilities altogether.

For the first time, business leaders, practitioners, students, and interested laypeople will learn what really makes a data science project successful. By illustrating with many personal stories, the authors reveal the harsh realities of implementing AI and analytics.

Kindle Edition

Published September 5, 2024

Reviews

[Allen G. · Rating 5 out of 5]

I very much enjoyed reading this book for several reasons. First of all, the authors define failure and then highlight a dozen tangible, easily understandable, well-defined reasons why data science, analytics, and AI projects fail, and then back up those well-researched reasons with real-world case study examples from corporations, academic projects, and consulting. Second, this is not a technology book, or a math/computer science book in that the main reasons projects fail are due to "human factors" and "organizational behaviors" underpinning and overarching how the tools are applied, utilized, and deployed inside large, complex organizations. Lastly, the book cuts through the hype associated with Analytics, Data Science and AI by simultaneously a) providing tangible, measurable examples of monumental success in terms of economic impact and business value of what can be achieved and then b) providing a treasure map for managing risk and error-prone behaviors of both leaders and practitioners. If you are a faculty member running capstone projects, a student/recent graduate embarking on a career in this domain, a practioner (veteran or otherwise) or a leader of a team or an executive seeking advice on how to best leverage Data Science & AI, then this book is a must read for you. The fact that Tom Davenport believes that this book is important (in the Foreword), given his prolific and prodigious book production in the domain, means it actually is!

[Ali · Rating 5 out of 5]

I just finished reading "Why Data Science Projects Fail" and it’s an absolute eye-opener. As an analytics professional, I’ve often been frustrated by the hype and overblown promises surrounding AI and machine learning. This book stands out by cutting through the noise and offering a clear, realistic view of what’s required to make data science projects successful—and, crucially, how to avoid becoming part of the staggering 80% failure rate.

What I love most about this book is its focus on failure as a learning tool. It avoids the flashy promises and success stories that so many other books lean on, instead breaking down the real-world challenges you’ll face. You can tell it’s written by people who have actually implemented data science projects. The chapters are well-structured, starting with background and scene-setting before diving into the core reasons for failure—strategic, process, people, and technology. The final chapter's exploration of how even analytically mature companies falter is especially enlightening.

For anyone working in or around data science, this book is a must-read. It’s honest, insightful, and devoid of the typical hype, offering valuable lessons for organizations of all sizes. *Why Data Science Projects Fail* is a 5-star read, leaving you with the tools and knowledge to succeed where others stumble. Highly recommended!

[Tahir Latif · Rating 5 out of 5]

"Why Data Science Projects Fail" by Evan Shellshear and Doug Gray offers an insightful exploration into the complexities behind the failure of data science projects, a critical issue for organizations aiming to leverage data for innovation and growth. The authors meticulously dissect real-world cases and provide a framework for understanding common pitfalls, from poor problem definition and lack of stakeholder engagement to unrealistic expectations about AI capabilities.

Shellshear and Grey’s writing strikes a balance between technical detail and practical advice, making it accessible to both data professionals and business leaders. They stress the importance of aligning data projects with business goals, fostering cross-functional collaboration, and adopting a long-term perspective on AI and data science investments. Their discussion on the "hype cycle" and how it can distort expectations is particularly relevant, especially for organizations venturing into AI without a solid foundation.

What sets this book apart is its actionable insights—it's not just a critique of why projects fail, but a guide for how to avoid these failures. The authors emphasize the need for strong governance, clear objectives, and an agile mindset in managing data projects. Overall, this book serves as a valuable resource for any professional involved in data-driven initiatives.

[Jordan · Rating 5 out of 5]

Why Data Science Projects Fail is written by real professionals who truly understand the intricacies of data science. This book is game changing, cutting through the hype surrounding data science, analytics and AI, and getting to the core of why 80% of data science projects fail - be it due to strategy, process, people, or technology.

The authors comprehensively blend real-world examples with deep insights into the common pitfalls that derail data science projects and offer actionable strategies to ensure success. They emphasise the importance of cultivating a strong analytics culture, which is often overlooked but critical to long-term project success.

Whether you're a data scientist, manager, or executive, this book equips you with the tools and mindset needed to navigate complex projects, mitigate risks, and deliver value. A must-read for anyone serious about transforming data into actions into results!"

[John Gizzler · Rating 5 out of 5]

This is a book that has been sorely needed. With all the hype surrounding AI and Data Science, it is good the authors choose to address the elephant in the room, the fact that most AI and Data Science projects fail.

What this book does brilliantly is it takes what was clearly significant research and provides it with context by telling stories around the reasons for failure and helping the reading to better contextualise the reasons for failure and so more easily apply it to their own situations.

The style is engaging and interesting and definitely worth reading for anyone in the AI and Data Science industry.

[Kah Wee · Rating 5 out of 5]

I just finished reading "Why Data Science Projects Fail" and I must say, it's a game-changer. As someone who's worked in the field of data science and analytics for years, I've seen firsthand the hype and exaggeration that surrounds AI and machine learning. This book is a breath of fresh air, cutting through the noise and providing a realistic view of what it takes to make data science projects successful.

Overall, I highly recommend "Why Data Science Projects Fail" to anyone looking for a honest and insightful look at the world of data science. It's a 5-star read that will leave you with a deeper understanding of what it takes to succeed in this field.

[Raymond Kamb]

With so much hype around AI and Data Science this book provides clarity around how to practically implement AI and Analytics projects in an organization. The authors both seasoned experts in the field of AI and data science draw on their extensive experiences and research to provide valuable insights from both successfully projects as well as failed projects and the lessons learned from those failed projects. The book is clearly written in a non-technical way that anyone can understand including university students and non-experts. A must read for anyone interested in the fast evolving field of Data Science and AI.

[Nick Holland · Rating 5 out of 5]

I loved this book. Both authors have extensive experience delivering complex, high budget projects and they leverage that experience to bring data science and AI to life. They use detailed case studies to highlight common pitfalls, but also to offer meaningful advice. The reality on the ground is very often different to what is presented in the media - data science is difficult, messy and often fails. It made me realise "my organisation isn't such a mess after all - this data science stuff is hard". I would recommend this book to anyone (technical or non-technical) who works in an organisation contemplating a data project (or any tech project for that matter).

[Juna Langforf]

Overdue. This book is a must read for all to understand what is happening in this realm. It is accessible and interesting to not only the student and business manager but also the layman. This is a result of an example rich format accompanied by the expert analysis of the two highly experienced authors. They are to be congratulated. And in addition for realising that despite formidible expertise, we will never succeed in any undertaking if we lose our humanity.

[Sara · Rating 5 out of 5]

Why Data Science Projects Fail provides a clear and insightful look into the common reasons AI and analytics projects don't succeed, cutting through the hype to offer practical advice. The authors use real-world examples and straightforward guidance, making this book an essential read for anyone involved in data science projects.

[Isabella · Rating 5 out of 5]

This book takes a much-needed pragmatic approach, cutting through the marketing hype and exaggerated promises that surround AI. It provides a refreshing perspective on what truly goes into successful data science projects, making it an essential read for anyone serious about understanding AI and its future.

[Davin Shellshear · Rating 5 out of 5]

This book is very timely with all the hysteria and hype over AI. The book provides a balanced approach to AI and how an organisation can plan to develop their own AI system consistent with their strategy, capability, governance, risk management and operational needs

[Michael Langford · Rating 5 out of 5]

Thoroughly enjoyed this book. A combination of pragmatic learnings melded with an enjoyable to read writing style. Strongly recommend for anyone interested in trials and tribulations of data science projects and AI.

[Hannah · Rating 5 out of 5]

Cutting through the hype with real examples, this book is a valuable resource for any data leader. It seems to be a realistic guide, focusing on the complexities rather than overpromising on AIs potential.

[Debbie Hacking · Rating 5 out of 5]

Finally a realistic evaluation of AI, its uses and limitations. Fabulous book and well written.

[Reader · Rating 5 out of 5]

Entertaining and insightful book