New Book Review: Disruptive Analytics
New book review for Disruptive Analytics: Charting Your Strategy for Next-Generation Business Analytics, by Thomas W. Dinsmore, Apress, 2016:
The author mentions at the outset that analytics industry leaders have been struggling, for reasons apparently not associated with product quality, but with disruption. Powerful forces are rearranging the industry: rapidly declining storage costs, an exploding number of data sources, decentralized data ownership, the open source software business model, the elastic business model made possible by cloud computing, widely available and inexpensive computing power, and displacement of conventional data warehousing and commercial analytic software by the Hadoop ecosystem, Python and R.
This book discusses two types of disruption: disruptive innovation within the analytics value chain, and industry disruption by innovations in analytics. From the perspective of startups and analytics practitioners, success is enabled by disrupting their industries, because using analytics to differentiate a product is a way to create a disruptive business model or to create new markets. From the perspective of investing in analytics technology for their organizations, taking a wait-and-see approach might make sense because technologies at risk of disruption are risky investments due to abbreviated useful lifespans.
After an initial walk-through of disruptive analytics innovation, and the analytics value chain, Dinsmore takes a step back to present a short history of the last 50 years of innovation in analytics, followed by discussions on the open source business model, the Hadoop ecosystem, the rapidly declining cost of computer memory and the corresponding rise of large-scale in-memory computing, a survey of streaming analytics, cloud computing and the elastic business model, recent innovations in machine learning (ML), self-service analytics, and a self-described "manager's handbook for disruptive analytics".
Chapters 2 and 3 on the history of analytics and open source analytics are especially well done, and key for anyone not familiar with these areas. As a consultant in this space who began their programming career with SPSS, I recall a diagram that someone posted this past year, a bit of a spin on the theoretical evolutionary chain of humans, albeit beginning with SPSS, with subsequent stages of SAS, R, and Python. In a similar manner, the author walks through the themes characterizing analytics in the years prior to the introduction of the first data warehouse in 1984, the data warehouse era, and key trends today.
As explained in the summary to chapter 2, much is to be learned from a review of the history of modern analytics. "Statistics, machine learning, and data mining technologies developed separately from data warehousing and business intelligence technology. While data warehousing theorists argued that data mining 'belonged' in the data warehouse, and leading database vendors delivered 'in-database' data mining, actual users disagreed with the theorists. On the whole, they preferred separate tools based on servers or desktops, which had much richer functionality than the 'in-database' data mining tools."
"The vision of the enterprise data warehouse as the single source of truth was never realized, even at the peak of the hype cycle. No large organization ever successfully integrated all of its data into a single centralized repository. If the vision of a unified enterprise data warehouse was unattainable in the 1990s, it is certainly unattainable today. There is simply too much data, the data is too diverse, and it moves too fast. The enterprise data warehouse idea may be dead, but the data warehouses themselves will survive, like cathedrals in modern cities."
"Organizations that built them will not decommission them, and some may build anew. Every organization needs to measure its performance, and data warehouses are very good at providing broad access to consistent transactional metrics. But performance measurement isn't disruptive, and it isn't strategic. It is simply a cost of doing business, like pencils and office space. Consequently, the executives who manage data warehouses will be under constant pressure to deliver metrics at the lowest possible cost." And "meanwhile, the most interesting, strategic, and disruptive analytics will be built outside the scope of the data warehouse."
While this book itself could be considered disruptive from the perspective of the field of available texts on this subject matter, as very few exist, it is also important to consider the fact that this book was published in 2016, and the field continues to evolve quickly, demanding a desperately needed update when it comes to specific technologies covered in chapters 3 through 9, keeping in mind there are only 10 chapters. Experienced consultants in this space will not have a hard time finding numerous examples of where this is the case, but as this book arguably targets managers it is important that this aspect be pointed out.
In addition to refreshing this text from the perspective of cited technologies, I would also like to see the author spin off an entirely separate book from the last chapter, labeled "handbook for managers: how to profit from disruption". As someone who just built a data analytics platform with his team this past year, I know how challenging it can be to hire data professionals, so I would especially like to see Dinsmore expand on his "hire the right people" section of this chapter. And as a longtime agile practitioner and open source advocate, I would like to see expansions of the "develop a lean data strategy" and "build an open source stack" sections.