Chapter 7: Introducing the Solution Space: A Framework for Business Hypotheses and Actioned Insights

Probably to nobody’s surprise, the solution space to address our persistent data and analytics problems with “insight leakage” entails a three-pronged approach consisting of robust technology and a high-impact workflow working in concert with significant consideration for human and organizational factors.

Copyright ? 2019 Alan R. Simon. All Rights Reserved.

Introduction

           Thus far, we’ve mostly focused on the problem space when it comes to business hypotheses in particular, and data-driven actionED insights overall. We’ve examined two high-profile failures (9/11 and the General Motors ignition switch recall), followed by a look at the root causes for why business hypotheses continue to present such difficulty in some areas of our organizations.

           In between, though, we took a deep dive into the very nature and composition of business hypotheses, and also examined the characteristics of use cases in which we actually do a fairly good job overall at immutably shepherding business hypotheses towards decisions and actions.

           This chapter brings together our Part I topics to introduce the solution space for business hypotheses. This solution space is comprised of not only powerful technologies but also a robust workflow, and is rounded out by a portfolio of strategies and tactics that address human and organizational factors.

Philosophy

           Figure 7.1 illustrates the underlying philosophy behind our solution space.

Figure 7.1 – A Three-Factor Model Impacting Insight Leakage

           We saw in Chapter 6 how the outer ring of our three-factor decomposition of business hypotheses, in which we see the combination of:

1. high personal accountability,
2. strong guidance, and
3. high familiarity

tends to result in greater success with shepherding business hypotheses…and thus yields significantly less “insight leakage” along the way towards data-driven insights and actions.

           Likewise, the three-factor model from Chapter 5 illustrated and described how the inner ring of that model – more straightforward predictive analytics rather than their “fuzzier” exploratory counterparts; departmental or single-function rather than cross- or pan-enterprise breadth; and something we are actually anticipating – represents the “sweet spot” of greater analytical successes.

           We find a direct relationship between the outer-ring factors of Chapter 6’s model and the inner-ring successes of Chapter 5. Therefore, our approach follows the philosophy stated below and illustrated in Figure 7.2:

 Apply the outer-ring success factors from Chapter 6 to the entire spectrum of analytical use cases covered in the Chapter 5 model, all the way out to the “danger zone, with the objective of reducing insight leakage and increasing actioned insight throughput.”

Figure 7.2 – Bringing Both Three-Factor Models Together

           Philosophies, of course, tend to be easier to espouse than they are to actually implement. How, then, can we “walk the talk” of the above prescription to our analytical woes?

           As introduced earlier and described in the sections that follow, the combination of 1) powerful technologies, 2) a robust workflow, and 3) attention to human and organizational factors gives us the right recipe for success (Figure 7.3). 

Figure 7.3 – Our Three-Pronged Solution Space to Address Business Hypotheses and Actioned Insights

           Each of the following sections serves as a brief introduction to the three prongs of our solution space. The book sections that follow address each in much more detail: Technology in Part II, Workflow in Part III, and Human/Organizational Factors in Part IV.

Technology

           Since the 1990-1991 time frame, data-driven insights have been powered by the highly synergistic combination of data integration technologies and powerful information delivery tools. Classic data warehousing and business intelligence held the stage for close to two decades, and both still play a prominent role in the overall analytics space. Their logical successors – big data and “advanced” analytics (e.g. machine learning), respectively – have played a critical role in moving data-driven insights into an era hallmarked by predictive and discovery analytics.

           However…

           The title of Chapter 10 says a great deal: Big Data By Itself is Not Enough. Without a doubt, platforms that enable the original “3 V’s” of big data – volume, velocity, and variety – are an order-of-magnitude game-changer when compared to the technology used for classic data warehousing. We will explore the “paradox” of big data in Chapter 10, but as a preview, consider the following assertion:

The most robust, high-performance, and reduced-friction integration of data does not guarantee insights produced from those data will immutably drive downstream decisions and actions.

           We’ll explore the role of big data in our business hypothesis/actioned insights solution space in Chapter 10.

           Similarly, advanced analytics applied against that big data-managed content have opened up an entirely new era of data-driven insights. However, the entire continuum of analytics, including classic reporting and business intelligence, has a role to play in converting those initial insights into what we need to drive decisions and actions. Chapter 8 explores the analytics continuum, while Chapter 9 calls out the revival of discipline that actually predates the business intelligence era…specifically, decision support systems.

           Finally, artificial intelligence (AI) definitely has a role to play, as discussed in Chapter 11. AI is a foundational game changer – just like big data and advanced analytics – but without the accompanying process/workflow and human/organizational components of the solution space, AI by itself is not enough to immutably bring about actioned insights.

Processes and Workflow

           We introduced aspects of the actioned insights workflow in Chapter 5, when comparing our “sweet spots” of success with corresponding analytical use cases where we typically see significant insight leakage. Figure 7.4 depicts this workflow in graphical form.

Figure 7.4 – The Business Hypothesis/Actioned Insights Workflow

           A key element of our actioned insights workflow is that while we definitely leverage and make use of advanced technologies, we also find elements of “classic” business process management converging with our analytical and data management capabilities. Figure 7.5 illustrates this philosophy of fusing these disciplines, while Figure 7.6 goes a step further.

Figure 7.5 – The Fusion of Analytics and Business Process Management (BPM)

           Specifically, as illustrated in Figure 7.6, every major step or phase along our actioned insights workflow (Figure 7.4) is hallmarked by three characteristics:

1. the degree of analytical complexity (low to high)
2. the variability of our analytical workflow (low to high)
3. the degree of human involvement vs. “lights out” process automation (low to high)

Figure 7.6 – Three-Factor Model for the Business Hypothesis/Actioned Insights Workflow

           As discussed in detail in Part III of this book, any given instance or use case of an analytical workflow might find itself hallmarked by the combination of values anywhere along each of those three axes. For example, we might drive and shepherd insights all the way through to decisions and actions in one situation through:

• very complex analytics, in combination with
• a highly variable workflow, along with
• minimal human involvement (e.g., heavily or fully automated end-to-end processes.)

           In a second use case, however, we might find:

• moderately complex analytics,
• a workflow with very low variability, and
• significant human involvement along the end-to-end process.

           To complicate matters, as we’ll see in Part III, each step along the workflow will feature its own combination of values for analytical complexity, workflow variability, and degree of human involvement. For example, the manner in which a critical business hypothesis is produced might involve moderately complex analytics, low-variability workflow, and little human involvement; but the steps to prove or disprove that hypothesis require low analytical complexity, moderate-to-high workflow variability, and significant human involvement.

           Don’t worry too much about the details, since we’ll deep-dive into them in Part III. For now, the key points about our actioned insights workflow are that 1) we are fusing best practices from the world of business process management with modern advanced analytics and data management, and 2) any given actioned insights workflow might vary from any other with regards to its combination of analytical complexity, variability, and degree of human involvement.

Human and Organizational Factors

           Part IV of the book details the “finishing touches” that are required to catalyze our technologies workflow into finally realizing the goal of broad actioned insights that we’ve pursued for so many years. Figure 7.7 graphically depicts the “4 E’s” that we’ll discuss:

1. Evangelizing the message of actioned insights and business hypotheses to your organization’s executives and stakeholders;
2. Educating the broad workforce across your organization through a series of powerful, tailored workshops;
3. Evaluating how well your organization currently does with actioned insights…your own “sweet spots” and danger zones; and
4. Experimenting with newly built capabilities that a) adhere to the actioned insights workflow to minimize insight leakage, and b) deliver the insight-to-action “final mile” that in turn energizes your organization to adopt these principles and techniques.

Figure 7.7 – The “4 E’s” for Human/Organizational Factors

________________________________

Alan Simon is the Managing Principal of Thinking Helmet, Inc., a boutique management and technology strategy consultancy specializing in analytical business process management, business intelligence/analytics, and enterprise-scale data management.

Alan is the author or co-author of 31 business and technology books, dating back to 1985. He is also the author of five LinkedIn Learning/Lynda.com courses, the most recent being EDGE ANALYTICS: IoT AND DATA SCIENCE.