Introduction: The Last Analytics Mile

THE LAST MILE TO ANALYTICS SUCCESS

Analytics became widely known and accepted as a competitive imperative in 2006 when Thomas Davenport published his landmark article, "Competing on Analytics," which soon became one of the top-10 must-read articles in the history of Harvard Business Review.1 Analytics had always been helpful, but as long as your competitors were not using it, you had a chance to survive without it, too.

Now that analytics has become affordable and practical to do at scale, everyone is doing it, and so must you if you wish to survive in the new age of Big Data, and the intense competitive pressure brought on by those who know how to "compete on analytics" well. Ahmer Inam, chief data and AI officer at Relanto, painted a stark picture of the competitive landscape when he said businesses in today's world must "do analytics or die."2

Most businesses know the importance of using analytics, prompting them to invest more than $100 billion by 2018,3 which shows what was collectively spent by organizations to take advantage of the power of analytics. Unfortunately, close to $90 billion of those funds missed the mark by failing to generate the expected return on investment (ROI). That is right: nearly 90% of all analytics projects failed to generate "significant financial benefit," according to a report that MIT released in 2020.4 No matter how good we analytics professionals are at building models, if they are not adopted and integrated into the organization in a way that creates value, they will be perceived to be failures.

In some ways this failure rate is to be expected, because all young disciplines fail in the beginning as they learn to walk. Analytics is, in fact, still a very young discipline, and one that is changing more rapidly than nearly any other, so it is understandable that the failure rate is so high. Understandable, yes, but it is still unacceptable, and is a tragic waste of resources that could be used much more effectively if analysts and business leaders were able to more effectively work together to manage and avoid the preventable, the manageable, causes of analytics failure.

Yes, some projects will always fail, just like some planes still crash and some bridges will still collapse. However, these rates are much less than they were in the beginning. The failure rates in other professions have been dramatically reduced as they have matured and developed a set of professional best practices that taught them and their sponsors how to work together to succeed. This can be true of analytics as well.

Indeed, this must be true for analytics. We must make it true. Analytics is not just about making more money, or even about firm survival in a competitive landscape, as important as those things are. It is also about doing things better, doing them more efficiently, sustainably, and intelligently. It is about, or can be about, with the right moral and ethical practices, building a better society that uses analytics to compete, but also uses it to make our world a better place to live and work.

We know that this concept can be true because a handful of companies such as Capital One, Amazon, and Intel have shown that it can be done well. Even so, in that same MIT report showing only an 11% average success rate of adding significant value, they also reported that some firms had achieved as much as a 73% success rate. That is nearly a sevenfold increase in success and something to which all of us can aspire. Even more, I believe the value created by these successes is far greater than the cost of all of the failures put together, as shown in Figure 0.1.

The firms that have succeeded in achieving these astonishing levels of success with analytics had a few things in common. The most important of which is that they achieved a high level of organizational learning for, with, and on behalf of AI and analytics. They learned from analytics how to change because of it, and they were willing, even eager, to change. Not just among the analytics professionals, but across the entire organization, along with their partners who were able to learn, adapt, and grow to apply and support analytics and did so at an industrial scale across the organization.5

What is the secret to this level of organizational learning about analytics? Certainly investment in people, tools, and technologies. Certainly commitment to do it. Certainly competitive pressure pushing for it. But none of these reasons are enough on their own. It also takes many more analytics supporters and enablers than analytics doers, maybe on the order of 10 to 1, for the organization to learn, grow, and succeed at this much higher level.

To be the type of organization that truly takes advantage of the potential of analytics, one that succeeds much more often than not, the organizational leadership as a whole, not just the analytics professionals, need to know their role in adopting and supporting analytics. They need to develop the skills to work together with the analytics team, and vice versa, to understand the business value of analytics and the critical nontechnical role they have to play in analytics success. They need to become an adaptable learning organization that uses, and is continuously and skillfully driven by, analytics to compete, grow, and improve in their efforts.

Figure 0.1 Many Failures, Some Astonishing Successes

Nontechnical employees need something more than data literacy, but less than coding, to succeed in analytics. They must become fluent in the best practices of analytics, at least the ones that interact with their role and function in and around the organization. Yes, analytics professionals need to also learn to better interact with the business, as has been identified many times in many places, but it will never be enough. Analytics professionals are not capable of doing it on their own. They must be guided and supported by at least an order-of-magnitude times as many skillful analytics supporters and enablers to succeed.

This organizational learning should not be confined to the analytics team, though that is part of it, too. Indeed, it cannot be confined to the analytics team if we want analytics to move from the lab into production. There are three core groups of people in the organization who must work together skillfully for analytics success. This is what Dr. Rudi Pleines, head of business transformation at ABB Robotics, calls the minimal viable team needed for analytics to succeed.

This minimal team includes (1) an executive champion to sponsor the project, (2) a business process owner who is able to integrate a tool and the related analytics into the processes of an organization to ensure they are used and the generated value is maximized, and (3) the technical analytics person or team to do the analysis. Notice the technical and analytics teams are necessary, but they cannot succeed on their own. Even working together, it still takes skillful collaboration aimed at overcoming common causes of failure to succeed.

Most of the causes of analytics failure are manageable, wrote Dr. Karl Kempf,6 head of analytics at Intel, whose team is responsible for documented savings exceeding $55 billion from analytics. Manageable means preventable, if you are smart and skilled enough to manage the causes of failure correctly. The analyst has direct control over only one of the five manageable tasks Kempf identified as necessary for analytics success.

This means that the entire organization, including many more non-analytics professionals than analytics professionals, need to learn how to engage with and support analytics effectively for long-term analytics growth and success. This community effort is how analytics becomes a profession: by growing beyond a few innovative pioneers into a standard, repeatable, organization-wide process that can consistently add value.

This is the last mile of analytics: learning to work together to get more projects into production successfully. We, as analytics professionals, cannot walk that last mile alone. It takes all of us, working together, skillfully, to dramatically increase the success rate of analytics and provide value to the organizations we work for and with.

This book is about how you, whoever you are and whatever your function, can more effectively lead, guide, support, and integrate with analytics to build the kind of mature analytics organization that succeeds, not just on a few projects, but on most of them. It is a collection of best practices addressing each of the manageable tasks in analytics, the preventable causes of failure that destroy projects, and how you can use them to compete on analytics as Thomas Davenport advised in his 2006 Harvard Business Review article. It is about how you can help analytics cross that last mile to analytics success and maturity as a profession into a practice of success.

EXPERT CONTRIBUTORS

The knowledge and experience of more than three dozen highly successful experts is condensed and shared as best practices in this book, along with their many stories, illustrating what analytics can do and how to use it. I am grateful they agreed to share some of their time and wisdom in an effort to help build analytics as a practice and profession and increase the analytics success rate.

These experts were all handpicked for this book because they have something unique to offer...