how the 9/11 commission report and the decline of ... · blockbuster had the opportunity to...

19
A FRAMEWORK FOR ANALYTICAL ECOSYSTEM GOVERNANCE How the 9/11 Commission Report and the Decline of Blockbuster Serve as a Warning of Failing Analytical Systems

Upload: others

Post on 10-Aug-2020

2 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: How the 9/11 Commission Report and the Decline of ... · Blockbuster had the opportunity to purchase Netflix in 2000 for $50 million (Netflix is currently valued at $40.6 billion)

A FRAMEWORK FOR ANALYTICAL ECOSYSTEM GOVERNANCE

How the 9/11 Commission Report and the Decline of Blockbuster Serve as a Warning of Failing Analytical Systems

Page 2: How the 9/11 Commission Report and the Decline of ... · Blockbuster had the opportunity to purchase Netflix in 2000 for $50 million (Netflix is currently valued at $40.6 billion)

2 alithya.com

Table of Contents

03 Foreword

03 Reporting vs. Analytics

04 The Four Systemic Success or Failure Factors

04 Imagination

05 Policy

06 Capabilities

07 Management

08 IT Governance and the Systemic Factors

10 Data Source Components

11 Semantic Components

12 User Interface Components

13 User Personas

18 Analytic Components and Enterprise Ownership

19 Final Framework Considerations

Page 3: How the 9/11 Commission Report and the Decline of ... · Blockbuster had the opportunity to purchase Netflix in 2000 for $50 million (Netflix is currently valued at $40.6 billion)

3 alithya.com

ForewordAlithya blog posts, such as Oracle Business Intelligence EPM and Relational Federation – A Strategic Approach, Upgrading Oracle Business Intelligence to 12c, and The Role of Oracle Data Visualizer in the Modern Enterprise, have provided ancillary thoughts on governance. In this white paper, we want to explicitly state some governance strategies as they relate to analytical systems. For the purposes of this framework, governance is defined as the processes put in place by enterprises that manage and control changes. A well thought out governance process is essential to create an analytical ecosystem that responds to, or even anticipates, changing competitive landscapes and opportunities. As we have learned from the failures of the U.S. intelligence organizations and Blockbuster Video, applying the same governance processes for transactional systems to analytical systems will result in a degraded return on investment, and perhaps, unprecedented failure. Furthermore, within the analytical system, broad strokes of governance should not be applied. As a side note, governance should not be confused with security that determines user access to the system or particular data elements. While security is an essential organizational consideration, security processes should be designed to complement governance processes and not the other way around.

To help develop this framework, we will start by defining reporting versus analytic systems. The systemic factors that determine the success or failure of analytic systems will be defined. A generic analytic system architecture will be explained with correlation to each of the systemic factors. Finally, the role of the IT support team along with functional business teams in the governance process will be explored. We will attempt to outline the convergence of people, processes, and technologies throughout this governance discussion.

Reporting vs. AnalyticsWhile organizations often create analytic systems by evolving reporting systems over time, it is important to establish a difference between these systems. Reporting involves the process of organizing data into information to create the ability to monitor the performance of different business departments, functions, or processes. Reporting typically creates more questions than answers: Why was revenue down year over year this quarter? Why are margins so low in this particular geographic region?

Analytics should be capable of answering these questions and create opportunities for action: Revenue is down because Widget B has comparable products competing in the market that are forcing the price down. Margins are low in this particular region because of a substantial tax increase in this state. Analytics may create additional questions while answering others (i.e. we introduced Widget C in anticipation of this market pressure; why have we not seen an increase in our revenue figures?), but there is data to either answer the new questions or the organization has the desire to gather supporting data. The ability to gather this data in a cost effective and timely manner is beyond the scope of this paper.

Page 4: How the 9/11 Commission Report and the Decline of ... · Blockbuster had the opportunity to purchase Netflix in 2000 for $50 million (Netflix is currently valued at $40.6 billion)

4 alithya.com

The Four Systemic Success or Failure FactorsWith these definitions in mind, what factors contribute to an effective analytical system? After the tragic events of September 11, 2001, President George W. Bush and Congress requested an official report of the circumstances 2 leading up to the attack. The result was the 9/11 Commission Report (hereafter referred to as The Commission) which identifies four systemic failures of the analysis capabilities of the intelligence agencies that created the opportunity for the attack: imagination, policy, capabilities, and management.* While an enterprise operating in even the most competitive industry will never face the levity of the consequences of these failures, these are helpful factors to evaluate and determine analytical governance approaches.

ImaginationDespite the difficulty of explicitly defining the role of imagination within an enterprise, this factor directly contributes to the agility of the analytic system. In our opinion, agility is the single most important factor in governing an analytical system. This correlates with The Commission’s view that imagination is the most important factor that created the opportunity for the attack: “Imagination is not a gift usually associated with bureaucracies... It is therefore crucial to find a way of routinizing, even bureaucratizing, the exercise of imagination” (emphasis added). Had intelligence analysts been able to imagine the nature of the attack, they may have been able to explore or create channels for other actionable intelligence. These channels may have produced the kind of data a team of skilled analysts may have been able to assemble into actionable insight.

Again, the levity of the failure of imagination does not have such extreme consequences in a business setting. Regardless, if Blockbuster had had the imagination to recognize the threat of Redbox and Netflix, and the data to determine how to react to the changing market, those little blue membership cards might still be dangling from our key chains.

The Commission identified four common elements of understanding and forestalling changing conditions that may require an enterprise to change its approach to analysis and be prepared to react. These elements, paraphrased from The Commission’s report, serve as a method to institutionalize imagination:

1 Think about how the competitive landscape is changing

2 Identify telltale indicators (or KPIs) connected to the most dangerous possibilities

3 Where feasible, collect data on these indicators

4 Adopt strategies to deflect the most dangerous possibilities or at least trigger an earlier warning

* National Commission on Terrorist Attacks, The 9/11 Commission Report: Final Report of the National Commission on Terrorist Attacks upon the United States (New York: Norton Trade E-Books, 2011), Kindle edition), Chapter 11: Foresight—And Hindsight.

Page 5: How the 9/11 Commission Report and the Decline of ... · Blockbuster had the opportunity to purchase Netflix in 2000 for $50 million (Netflix is currently valued at $40.6 billion)

5 alithya.com

In an analytics ecosystem, this corresponds to:

1 Identification and prioritization of competitive threats and opportunities by strategic leadership

2 Working with experienced analysts to determine data driven indicators

3 Integrating data and creating solutions for predefined analytical content, such as reports, dashboards, and ad hoc analysis

4 Making sure the appropriate group in the organization uses this information to take action

Creating the exercise of imagination within an enterprise seems to be an impossible task; however, standard practices of organizational change management apply. In most organizations, there are individuals who are already asking hard questions and assembling (or attempting to assemble) the data to gain insight. Leadership should enlist, enable, and champion these individuals to the rest of the organization.

PolicyPrior to 9/11, the United States had not confronted an organization like al-Qaeda through concentrated, prolonged military action. As a nation, we were well versed on how to deploy highly organized and trained military against the militaries of hostile nations. The numerous terrorist attacks in the two decades prior to 9/11 certainly bordered on acts of war, but against whom? What sort of action can be taken against these individuals in sovereign nations? As The Commission stated, “These policy challenges are linked to the problem of imagination [already] discussed.”

Returning to the Blockbuster conundrum, in 2004, as Netflix gained market share, one of the policy issues that this organization faced was franchise agreements. Franchise owners were resistant to participating in mail delivery exchange programs due to the lack of adequate compensation. Data had to strongly indicate that Blockbuster was losing market share to Netflix (perhaps too late to take corrective action). Despite this, Blockbuster was unable to create a sufficiently competitive product to counteract the market influence. The organization had to have the data to indicate that future sustainability was questionable, but was unable to take corrective action due to rigid policies.

For departments supporting analytics systems, we have seen analytics systems failing to reach their true potential due to the result of policies rooted in obsolete information technology approaches. This concept will be explored more thoroughly in the system architecture section, but this amounts to two basic best practices. First, users should be trained, empowered, and encouraged to perform ad hoc analysis. Second, different architectural components of an analytical system require different degrees of oversight when changes are proposed. Treating the entire solution with a homogeneous governance approach will reduce the agility of the solution by imposing unnecessary restrictions. As The Commission noted, the role of imagination and policy go hand in hand; the interdependence of these factors is no different for an analytical ecosystem.

Page 6: How the 9/11 Commission Report and the Decline of ... · Blockbuster had the opportunity to purchase Netflix in 2000 for $50 million (Netflix is currently valued at $40.6 billion)

6 alithya.com

CapabilitiesIn addition to policies hampering the efforts of the U.S. combating terrorist groups before 9/11, “the United States tried to solve the al-Qaeda problem with the same government institutions and capabilities it had used in the last stages of the Cold War and its immediate aftermath.” For instance, in the latter part of the 1990s, the United States had naval vessels stationed in the Afghanistan region that were prepared to launch cruise missiles. Cruise missiles are expensive ($830,000 per very large explosion) and are typically used to destroy high-value targets such as ships and command bunkers. Deploying these weapons against individuals or even small groups of terrorists was not only costly, but often ineffective.

Military personnel faced similar capability shortfalls in respect to their perception of their role on the battlefield. For instance, special forces prior to 9/11 were determined to be self-reliant and “were reluctant to seek the support of outside [fire support] in order to keep operations small and light.” Task Force Dagger, the first special operations task force deployed in Afghanistan, found they had difficulty effectively deploying air support from non-special forces units.* Many governmental organizations (including the military) “are often passive, accepting what are viewed as givens, including the efforts to identify and fix glaring vulnerabilities to dangerous threats would be too costly, too controversial, or too disruptive.” Reluctance and unwillingness to confront the status quo briefly limited the capabilities of the United States' most well trained military assets. U.S. special operations have since become some of the most capable units, able to operate effectively in both independent missions and as part of larger, more integrated operations.

Blockbuster’s capabilities fell under a similar cadence. Blockbuster had the opportunity to purchase Netflix in 2000 for $50 million (Netflix is currently valued at $40.6 billion). Blockbuster CEO John Antioco declined to do so because he viewed Netflix as a niche market. In 2007, despite mounting pressure from Netflix as well as Redbox, Blockbuster surprisingly doubled down on the in-store, retail-oriented model. Blockbuster reduced marketing efforts of the Total Access online service, purchased a media retailer and distributor, and contemplated purchasing the struggling Circuit City. During this period, the number of chain and independent media rental stores was in steady decline. Blockbuster, founded in 1985, had developed capability when The Breakfast Club was hitting the shelves in VHS format. Despite data to the contrary, Blockbuster kept trying to capitalize on brick-and-mortar store capability long after Vice Principal Vernon let those five misfits out of detention.

Especially in this era of Big Data, legacy business intelligence systems will struggle to satisfy analytic requirements. This problem can be resolved through hardware and software and it is perhaps the easiest to remedy. The more difficult challenge is developing and continuously refining new data models to provide the capability to address a changing competitive landscape. Analysts must be capable of leveraging these new tools and data models or they risk having the same issues as Task Force Dagger - a staggering amount of firepower without the capability to use it. This does not require a one-time training effort, but instead involves a team willing to continuously learn and create organizational capability.

* Braganca, Eric. "The Evolution of Special Operations Joint Fires." Joint Force Quarterly 35 (2004): 64-68. Web. 15 July 2016.

Page 7: How the 9/11 Commission Report and the Decline of ... · Blockbuster had the opportunity to purchase Netflix in 2000 for $50 million (Netflix is currently valued at $40.6 billion)

7 alithya.com

ManagementThe Commission likened the various federal law enforcement and intelligence agencies (e.g. FBI, CIA, and NSA) to specialists in a hospital. Each was certainly adhering to its agency’s mission and directives, but “[w]hat [was] missing is the attending physician who makes sure they work as a team.” Each agency had data and even had terrorist suspects in custody that could have stopped the plot had the intelligence been coordinated across agencies. This glaring lack of a unifying management structure directly contributed to the failure to predict the 9/11 attacks.

In all fairness to Blockbuster, Netflix’s innovative approach to home entertainment did appear to be a niche market in 2000. Many customers seemed to be reluctant to abandon the spontaneity of selecting their Friday night movie entertainment directly from the shelf. Netflix’s model, which required some planning on behalf of its customers, did not seem logical. However, the data out there indicated that Blockbuster’s mode was in for some mortal blows.

First, 16% of Blockbuster’s revenue was based on customers paying late fees for rentals, a policy hated by many (ironically, it was this very policy that prompted Reed Hastings to found Netflix).* After briefly abandoning this policy, Blockbuster reinstated late fees; the type of policy flip-flop generally appreciated by customers. Second, disruptive, Internet-based models were already becoming viable. For instance, while the ill-fated Napster was not the first music file sharing network, it was the most user friendly and eventually attracted 80 million users. Music companies, initially perceiving online music sharing as little threat due to the popularity of portable compact discs, initially ignored the service. However, musicians and distributors quickly realized the impact on revenue these types of file sharing services were having. In an unusual trifecta, Metallica, Dr. Dre, and a few major record labels joined forces to effectively cause the demise of the original Napster in 2001.

In 2004, Blockbuster was struggling, but was still considered a powerhouse in the home entertainment market. It commanded the data from millions of customers, including rental trends, customer complaints, and market analysis. It had highly respected industry leaders poised throughout the organization and had already trounced many other home entertainment competitors. Despite all this data and the lessons learned from past successes, Blockbuster was unable to find that attending physician to orchestrate these capable teams.

The end goal of analytics is corrective or proactive action. Analysts who are secretly (or not so secretly) data junkies may be content with sifting through the numbers and discovering relationships. However, if this work does not seem to be of value to the organization or failing to spurn action, either corrective or opportunistic, these analysts may become discouraged. Likewise, competent leadership will look across the organization and start questioning why there is such an investment in analysts and technology without measurable returns. As with the various government agencies, an attending physician is essential to provide overall direction to the analytics program and help the various departments leverage this data.

* Anderson, M., & Liedtke, M. (2010, September 23). Hubris - and late fees - doomed Blockbuster. Retrieved September 14, 2016, from http://www.nbcnews.com/id/39332696/

Page 8: How the 9/11 Commission Report and the Decline of ... · Blockbuster had the opportunity to purchase Netflix in 2000 for $50 million (Netflix is currently valued at $40.6 billion)

8 alithya.com

IT Governance and the Systemic FactorsAn analytical system architecture is more than the technology components, such as a data warehouse and guided user interface (GUI) tools. The people and processes that interact and intersect (respectively) with these solutions are perhaps more important than these technology components. This intersection of technology, people, and process creates an analytical ecosystem. Governance, in the form of imagination, process, capability, and management, plays a role in the entire analytic ecosystem. Below is a generic diagram of a typical analytical ecosystem along with how the systemic factor impacts each component. Each component and the ranking of the systemic factors will be explained in subsequent subsections.

FIGURE 1: ANALYTICAL

ECOSYSTEM AND SYSTEMIC FAILURES

IImagination

PPolicy

CCapability

MManagement

SEMANTIC (LOGICAL) COMPONENT(S)

3 3

3 3

USER INTERFACE (PRESENTATION)COMPONENT(S)

3 3

3 4

Data Repository(Data Warehouse,

Data Mart orOLAP Cube

1 1

4 3

Extract, Transformand Load

(ETL) Process

DATA SOURCE COMPONENTS

USER PERSONAS

2 2

3 4

Execs and Functional Leaders

3 3

3 2

BusinessConsumers

4 4

4 3

BusinessAnalysts

1 1

2 3

Customers

Page 9: How the 9/11 Commission Report and the Decline of ... · Blockbuster had the opportunity to purchase Netflix in 2000 for $50 million (Netflix is currently valued at $40.6 billion)

9 alithya.com

The ranking of each systemic factor is relative to the components of the analytical ecosystem and represents a spectrum of approaches.

An imagination rank of one indicates the component should be fairly stable and not changed at the whim of a few users or market pressures. An imagination rank of four indicates the organization should be relatively free to adjust or alter the component to address changes or discover new information. It is worth noting that a ranking of one does not mean that component should be unalterable. Instead, these components should gradually and slowly change unless there is a major shift in the organization or market situation.

The Commission noted that imagination and policy are closely linked, and a complete correlation can be seen in the analytical ecosystem diagram. To compliment the ranking of imagination, components with a ranking of one should have policies that support a gradual changing while those with a ranking of four should support a flexible approach. A ranking of four does not mean that the enterprise should have a freewheeling approach to governance; however, approaches that limit users' agility and speed of response should be avoided. While the Policy rankings are relative to the analytical ecosystem, it is helpful to view the role of Policy within the organization’s broader information technology policy. Oftentimes, enterprise systems such as ERPs and CRMs have fairly stable and rigid governance approaches since they are designed to support the day-to-day running of the organization’s business processes. When compared with these enterprise systems, adjusting analytical ecosystem components with a ranking of four should entail significantly less enterprise governance.

Capability is divided into two general categories: technological and user. Technological includes performance speed, capability to aggregate and calculate data (including statistical calculations), and ability to store and transfer large amounts of data. These are the Data Source, Semantic Layer, and User Interface components. User capabilities includes the ability to understand how analytics contribute to their decision making and manipulate data further to tease out additional insights. These are the four personas grouped at the bottom of the diagram.

Finally, management is how these various components are leveraged to address the competitive landscape. Ranking a persona with a four indicates responsibility to drive change within the organization using data. A ranking of one for a persona indicates the responsibility to leverage data to compliment the execution of business processes and decision making. Ranking a technology component with a four reflects the responsibility of the support team and analytics leadership to ensure these tools are as simple and easy to use as possible to facilitate this enterprise-wide engagement. A ranking of one indicates these tools are typically hidden from end users; however, given the interdependence of the technology components, there is typically a cascading effect. Anticipating how design changes of lower ranking components affect higher ranking components should be the focus of the analytics support team.

Page 10: How the 9/11 Commission Report and the Decline of ... · Blockbuster had the opportunity to purchase Netflix in 2000 for $50 million (Netflix is currently valued at $40.6 billion)

10 alithya.com

Data Source ComponentsThe Data Source Components are comprised of the Extract, Transform, and Load (ETL) tools (sometimes ELT tools) as well as the data repository. Typically, changes in one results in changes to the other; therefore, the governance policy and systemic factors typically align closely. ETL tools are the means by which data is pulled from various source systems, cleansed, integrated, and logically extended before loading into the data repository. There are a number of tools on the market such as Oracle Data Integrator (ODI), Informatica, and Clover ETL. Data repositories are comprised of data warehouses or marts, online analytical processing (OLAP) cubes, and data lakes (typically for Big Data solutions). Many times, data are loaded into dimensional models that are optimized and conformed to enable fast querying and aggregations. Data repositories typically have highly integrated data organized into conformed data models (except for data lakes).

Recently, Oracle has introduced Visual Analyzer and Data Visualization (VA/DV) which provide the ability for users to combine local data with repository data through data mashups. Because this process does not formally conform or integrate data, the VA/DV tools that produce mashup data are not considered part of the data source.

Since data sources serve as the backbone of the entire analytical system, they receive relatively low rankings in imagination and policy. These data models should not be adjusted or changed at the whim of end users, especially for complex data sources. Rapidly altering these data sources can result in unexpected changes across the entire analytical ecosystem. Note that imagination is an essential component during the design and implementation phases so that the organization can establish a data repository that satisfies the identifiable concerns of the organization. Policy and imagination receive rankings of one so that the data sources experience only gradual and controlled change. As the target of all analytical queries, the capability of this component receives a ranking of four. Data sources should have the hardware and software to support the organizational demand. In addition, highly skilled implementation and support teams are essential to establishing optimized data models as well as hardware and software components. Finally, a management ranking of three ensures a team that can look forward through the analytical ecosystem and be able to determine how data source changes, configurations, and performance enhancements to hardware, software, and models can continuously improve the experience for end users.

Page 11: How the 9/11 Commission Report and the Decline of ... · Blockbuster had the opportunity to purchase Netflix in 2000 for $50 million (Netflix is currently valued at $40.6 billion)

11 alithya.com

Semantic ComponentsThe Semantic (or Logical) components typically lay between the data repository and the tools individuals use to interact with the data. The semantic layer allows administrators to organize the data so that users can more naturally interact with it. For instance, there may be a model to support sales pipeline analysis along with a dimension for customer and general ledger accounts. A sales rep in the field is not typically interested in the allotment of his customer’s accounts across the general ledger. Likewise, the financial analyst may not need insight into the specific customers interested in purchasing the company’s widgets. The semantic layer allows the administrator to present relevant and useful attributes to end users. As an extension of this, the semantic layer typically has a large role in the solution security. Examples of the semantic layer are the BI Administration tool for Oracle Business Intelligence, Data Modeler for Business Intelligence Cloud Service, and Enterprise Performance Management Architect for many Oracle Hyperion products.

In many cases, the semantic layer can be used to extend the logic of the data source and, in some cases, add capabilities that are not possible with the data source alone. For example, a correlation coefficient (a ranking of the dependence between two measurable factors) can be calculated in the semantic layer in a way that is reusable, scalable, and flexible. While this measurement could be calculated in the data source at a particular data source at a particular data intersection, it would be very difficult to design data structures that support the myriad of queries users will inevitably require. Whether organizing or adding value to the data source to the users, the semantic layer has no impact on the structure of the data source.

As a side note, there are tools such as Oracle BI Publisher that allow users to interact directly with the data repository without a semantic layer. Oftentimes, the semantic and user interface layers for these tools are tightly coupled; however, it is essential to identify the delimitation to ensure proper governance. These tools typically require users to have varying degrees of skill with querying languages such as SQL and MDX. These tools also typically lend themselves more naturally to reporting rather than analysis; however, a skilled analyst can certainly leverage these tools for analytics.

Due to the value that can be found in the semantic layer in providing additional capability with relative ease, all of the systemic factors receive a rank of three. Imagination and policy, which correspond to how quickly changes to the semantic layer occur, should result in a moderate amount of review and, once approved, these changes should be quickly exposed to the users. The capability of the semantic layer is ranked at three since many analytical systems offer more abilities within the logical layer than enterprises use. However, it is important to not have a tool so rigid that the analytic system support team has to go to great lengths to satisfy requirements. Finally, given the semantic layer’s role in organizing and extending the logic of the data (as well as security), the management aspect receives higher than average rankings.

Page 12: How the 9/11 Commission Report and the Decline of ... · Blockbuster had the opportunity to purchase Netflix in 2000 for $50 million (Netflix is currently valued at $40.6 billion)

12 alithya.com

User Interface ComponentsThe user interface (UI), or presentation layer, components are the mechanisms with which users interact with their data. These are often tools for creating and consuming dashboards, reports, graphs and charts, and alerts. Oftentimes, these same tools will allow users to perform ad hoc analysis on their own when standard, enterprise content was unable to anticipate user questions. The two most important features of a useful user interface are “analytics at a glance” and “analytics at the speed of thought.” A well designed dashboard solution will provide insight into how business processes are performing at a glance. Drill down/drill across capability will allow users to navigate down to more granular details and determine where problems or opportunities lie and take action. Ad hoc analysis coupled with standard, enterprise content should enable users to answer questions at the speed of thought. Analytics at a glance and at the speed of thought are largely influenced by the functional teams providing input and the skill of the implementation team with providing guidance and developing content. A good analytics software package will be able to facilitate that vision. An example of analytical tools with a proven track record of providing this capability is the Oracle Business Intelligence family (Enterprise Edition, Standard Edition, and Cloud Service) which offers Answers as the primary method of consuming data. Recently, the Oracle’s Visual Analyzer and Data Visualization tools have provided the capability for users to extend enterprise data through data mashups and create dashboards and reports with this combined data. An increasingly popular option is mobile technologies so that users can consume data through smart phones and tablet devices. When creating content for mobile devices, analytics at-a-glance becomes especially important.

FIGURE 2: ANALYTICS AT

A GLANCE—PROMINENT

FEATURING OF KEY PERFORMANCE

INDICATORS (KPI), VISUALIZATION

ARRANGEMENT, AND USE OF

COLORS SHOULD BE USED TO

QUICKLY DRAW THE ATTENTION OF

CONSUMERS TO KEY INFORMATION

Revenue

9BMargin %

6.47%Margin

592MExpenses

9B

Provider Margin % Key

FINANCIAL PERFORMANCE TILES

TOP 5 BUSINESS UNITS BY REVENUE, EXPENSE AND MARGIN

AVERAGE EXPENSES AND MARGIN PER ACTIVE CUSTOMER

Revenue Expense Margin

All Others 6.50%

Wachusett Rivers Medical Center 3.71%

Vista Healthcare 6.58%

First New England Hospital 6.42%

West Community Hospital 14.00%

Worcestershire Medical Group 6.48%

30K 18.00%

Rev

enue

(Avg

.) M

arg

in (A

vg.)

Marg

in %

All Others Wachusett RiversMedical Center

Vista Healthcare First NewEngland Hospital

West CommunityHospital

WorcestershireMedical Group

25K 15.00%

20K 12.00%

15K 9.00%

10K 6.00%

5K 3.00%

0K 0.00%

Page 13: How the 9/11 Commission Report and the Decline of ... · Blockbuster had the opportunity to purchase Netflix in 2000 for $50 million (Netflix is currently valued at $40.6 billion)

13 alithya.com

In the case of the user interface component, imagination, policy, and capability are tightly correlated at three. The ability to satisfy the imagination of functional leaders and implementation experts plays a large role in creating analytics at-a-glance and at the speed-of-thought. Enterprises need the ability to create rich visualizations, create and extend data calculations, drill down from a macro to a micro level, perform intuitive ad hoc analysis, and tie all these capabilities together to tell a compelling story through data. Without the capability to satisfy the imagination of the users in a prompt manner, the time and effort spent on curating and populating the data source will never satisfy the return on investment. Enterprise policy means that users need to be free to create ad hoc analysis to answer the unanticipated questions. As users gain more experience with the tool and available data, they will inevitably begin to ask more questions and develop an appetite for viewing their data through different facets (or dimensions). To support continuously evolving use cases, enterprise content should never be static, it should be relatively easy to update, add to, and refine enterprise content.

Of course, absentee oversight means unmitigated disaster. Management receives a ranking of four for the user interface component since this “wild west” needs a sheriff. Unregulated and lacking in strategic and enterprise level oversight, dashboards and other analytic content can quickly become disorganized and disjointed. It is the responsibility of management to be sure new and updated content satisfies strategic and operational goals of the organization, business process flows continue to be enhanced by analytics, and, most importantly, decisions continue to be made with complete, factual, and agreed upon data and definitions. Without capable management over the user interface, the entire analytical ecosystem will break down and become another IT tool bleeding money from the cash flow statement. Keep in mind that this is a balancing act - unmitigated oversight also means unmitigated disaster.

User PersonasUser personas are the broad classifications of data consumers. These are the individuals gaining insight and leveraging insight. With the exception of customers, these individuals should be interested in leveraging data for strategic and operational insight and continuously trained on how to do so. Customers, on the other hand, should be provided data that allows them to gain insights that are mutually beneficial. Of course, not all users have the same goals, responsibilities, and time; therefore, a slightly different approach is needed for each persona.

EXECUTIVE AND FUNCTIONAL LEADERS

Executive and functional leaders are the individuals within the organization making decisions with widespread, cross functional impact. Despite being responsible for asking questions and leveraging data to make decisions, these users receive a relatively low ranking in imagination and policy. These users should have analysts who know how to curate, visualize, and present the data in a meaningful manner (or at a glance) for these users to consume. Enterprise or curated content should be the life-blood of these consumers; in only the rarest of occasions will these executive and functional leaders be creating their own content. Keep in mind that these rankings are relative to the components of the analytical ecosystem; executive leadership has the responsibility of making competitively differentiating decisions with the insights provided by analysts.

Page 14: How the 9/11 Commission Report and the Decline of ... · Blockbuster had the opportunity to purchase Netflix in 2000 for $50 million (Netflix is currently valued at $40.6 billion)

14 alithya.com

In the case of these users, capability manifests itself in the ability to frame questions to analysts that provide enough information to provide direction, but are not so constraining as to result in the expected or desired answer. To help understand this distinction, consider the story of Abraham Wald, a Jewish mathematician during World War 2. Wald was asked to examine Allied bombers that were being shot down and help identify where additional aircraft armor was needed. Initially, the Allied military was examining where planes had been frequently shot (for instance, the wings and tail) and increasing the armor for those commonly damaged components. The problem? The Allies were examining bombers that had made it back to base and not those that were lost. Wald questioned if armor should be applied where the returned planes had been shot; after all, the pilots and crew had made it home. Instead, armor would be best applied where these planes had not been shot.* Wald was asked to help reduce the number of planes shot down over Europe and the Pacific, not to determine where the most bullet holes occurred on a plane. Making the distinction between these inquiries,and trusting in the analyst saved untold numbers of pilot lives.

FIGURE 3: AGGREGATION

OF BOMBER BULLET HOLES OF RETURNED

PLANES. THE COCKPIT, LANDING

GEAR, ENGINES, AND TAIL (BUT NOT THE TAILFINS) ARE UNDAMAGED AND

ARE THE MOST VULNERABLE.

* Mangel, Marc, and Francisco J. Samaniego. "Abraham Wald's Work on Aircraft Survivability." Journal of the American Statistical Association79.386(1984):259.Web.

Page 15: How the 9/11 Commission Report and the Decline of ... · Blockbuster had the opportunity to purchase Netflix in 2000 for $50 million (Netflix is currently valued at $40.6 billion)

15 alithya.com

The most important role of executive and functional leaders is management. In the case of this persona, this does not mean oversight, but example. Organizational users should observe leadership leveraging data through the analytical tools to make informed decisions and ask increasingly insightful and complex questions. This behavior will drive the organization to better data-driven decision making. Complimenting, rather than supplementing or stifling, technical and functional analytical leadership should be the goal of executive and functional leaders.

BUSINESS CONSUMERS

Business consumers cover a broad range of roles within an organization that are tasked with the tactical or dayto- day execution of the enterprise’s strategic goals. These may be store managers, call center representatives, or client relationship managers. These users should be able to perform casual ad hoc analysis to answer unexpected questions that occur during the course of fulfilling their responsibilities. Often (but not always), these questions will be answered through a one-time ad hoc analysis that will be discarded shortly after the information has contributed to a decision. As such, these users are ranked with a three in both imagination and policy. It is essential that they are empowered to use analytic tools to quickly answer questions.

Likewise, capability is ranked three. Business consumer users are not typically tasked with quickly creating complex dashboards or analysis to answer questions that can impact the business. These users should be well versed in quickly putting together visualizations from well understood data sources that can resolve immediate problems and keep tactical and day-to-day operations moving forward.

Management receives a ranking of two since business consumer users should naturally have access to enterprise content and data that aligns and compliments their roles within the organization. Thoroughly considering the applicable data needed by each type of business consumer during implementation will result in a solution that satisfies the depth and breadth of each user’s responsibility. As with other components, management should not be static or unwilling to provide users additional data assets. For instance, users that display unique analytical capability and roles that begin to cross into other functional areas should be provided access to corresponding analytical assets. Despite this, after the initial implementation and stabilization of the analytical ecosystem, these roles should be relatively stable.

Page 16: How the 9/11 Commission Report and the Decline of ... · Blockbuster had the opportunity to purchase Netflix in 2000 for $50 million (Netflix is currently valued at $40.6 billion)

16 alithya.com

BUSINESS ANALYSTS

Business analysts are the powerhouse users of the analytical ecosystem. If the executive leader is responsible for asking where to apply armor on the Allied bomber, the business analyst is responsible for clearly stating why the wings should not be encumbered by additional layers of steel despite the organization’s contrary inclination.

As such, imagination, policy, and capability are all ranked four. Wald, who had a doctorate in mathematics at the age of 29, was still unproductive due to Jewish persecution in Europe leading up to World War 2. It was not until he immigrated to the United States and began to work for the Center of Naval Analyses that he was able to save countless pilot lives and father the field of operational research. This unique combination of personal life experience, education, and opportunity provided Wald with the mental agility (or imagination), receptiveness of the U.S. Navy (or policy), and education (or capability) to directly contribute to the defeat of the Axis powers.

As with the attacks of 9/11, the gravitas of defeating the Axis powers far outweighs the role any business analysts contribute to moving an organizational objective forward. Regardless, business analysts should be enabled and trusted to derive analytics and provide direction into an organization’s strategic direction. While education in mathematics, data modeling, and visualization should be considered key components to an analyst’s capabilities, relevant and broad experience should also be considered. After all, Wald was invited to the United States to work on economics research and ended up advising mechanics on where to apply steel plating.

A great business analyst is able to bring imagination and capability to the role along with an enthusiasm for data and analytics. These very traits can make deliberate management of business analysts essential. Some business analysts will be content with sifting through data and putting together fantastic visualizations all day. Given no direction, they may produce some great insights, but are these insights contributing to the strategic and operational goals and/or producing actionable insight? A common story told about the value of analytics is the discovery by Wal-Mart that there was a correlation between fathers being dispatched to purchase diapers and “accidentally” dropping a six pack of beer in the cart. Because of this insight, store managers were directed to place diapers and six packs of beer in proximity in the store. The first part of this tale is true (although it was Osco Drug that discovered the correlation); however, a concentrated effort was never taken to implement the beer and diapers sales tactic. It was a curious insight (and tells you something about fathers shopping unsupervised), but it produced no actionable insight. The second case for strong management is that analysts that do not feel that they are making contributions will often become bored and seek other opportunities. With the scarcity of good analytical talent, it is becoming essential for organizations to retain, continuously develop, and (most of all) challenge these analysts. A good management structure can make this possible.*

* Davenport, Thomas H., Jeanne G. Harris, and Robert Morison. "Chapter 6: Analysts." Analytics at Work: Smarter Decisions, Better Results. Boston, MA: Harvard Business, 2010. 91-116. Print.

Page 17: How the 9/11 Commission Report and the Decline of ... · Blockbuster had the opportunity to purchase Netflix in 2000 for $50 million (Netflix is currently valued at $40.6 billion)

17 alithya.com

CUSTOMERS

Enterprises that open up analytical resources to customers create opportunities as well as risk. Banks have been early adopters of customer facing analytics; by providing their customers with spending habits and trends, they are able to encourage better saving habits:

FIGURE 4: BANKING ANALYTICS FOR

CUSTOMERS

Money In

$15,375.63Money Out

($7,455.09) $7,920.54

FINANCIAL PERFORMANCE TILES

Categories Amount %

Uncategorized 49.27%

Checks 28.78%

Loans 8.57%

Travel 8.20%

Utilities 2.86%

Uncategorized 2.26%

ATM/Cash Withdrawals

$4,400.00

$2,570.00

$765.00

$732.01

$255.05

$202.25

$5.20 0.06%

Besides the obvious security to prevent customers from seeing other customers' sensitive information, organizations should prevent customers from gaining an excessive amount of insight into their own profiles. Guided analytics provides customers with the ability to gain specific understanding that hopefully leads to better behavior while preventing unnecessary insight. Giving customers the ability to create their own analytics can attract customers (especially business- to-business customers). Very careful planning and design is needed before opening these analytic portals since creative and insightful customers can leverage this capability to gain insight into aspects such as segmentation or profitability.

Since customer facing analytics needs to be very tightly controlled, imagination and policy receive a ranking of one. Content should be tightly controlled and policy should ensure a thorough vetting of new content or self service capability. Most self-service customers will be content with standard data visualizations such as pivot tables and charting capabilities; therefore, capability receives a ranking of two. Finally, management needs to ensure that customer facing analytics provides the kind of insight to help customers make informed decisions that encourage behavior that benefits both the customer and the enterprise. Returning to the banking analytics, users able to make informed decisions about potentially excessive spending may cut back on spending in certain categories. The customers may begin to place more money in savings which provides the bank with more capital to lend.

Page 18: How the 9/11 Commission Report and the Decline of ... · Blockbuster had the opportunity to purchase Netflix in 2000 for $50 million (Netflix is currently valued at $40.6 billion)

18 alithya.com

Analytic Components and Enterprise OwnershipPerhaps because analytics systems converge data, servers, hardware, and networks, many organizations place the ownership of the entire system under their information technology support team. This ownership model guarantees that the governance policies suggested in the previous sections will fail. Analytical systems should be a balance and partnership between the IT department and functional users as illustrated below:

FIGURE 5: OWNERSHIP AND

INPUT BALANCE

SEMANTIC (LOGICAL) COMPONENT(S)

USER INTERFACE(PRESENTATION) COMPONENT(S)

Data Repository(Data Warehouse,

Data Mart orOLAP Cube

Extract, Transfortand Load

(ETL) ProcessIT Ownership

IT Input

FunctionalInput

FunctionalOwnership

DATA SOURCE COMPONENTS

Data source components are entirely in the domain of the IT support team while functional users own the user interface. Ownership does not mean exclusion. Functional users should have input into data source components while IT should be providing guidance to users during content development. This ownership and input balance enables the systemic factors of imagination and policy.

As the bridge between the data source and user interface, IT and functional users should also be meeting in the middle at the semantic layer. In most organizations, the semantic layer will be owned and maintained by the IT support organization, but this ownership should be done in close collaboration with functional users. Emerging analytical tools are greatly simplifying the semantic layer. Oracle Business Intelligence Cloud Service (BICS) features a sematic layer with a web-based interface that is intuitive and easy to use. Simultaneously, many organizations are employing functional talent with a moderate understanding of data modeling concepts. Organizations leveraging tools such as BICS with skilled analysts may find the functional users owning the semantic layer with oversight by IT support teams. This model is particularly useful for enterprises with an overworked IT department.

Page 19: How the 9/11 Commission Report and the Decline of ... · Blockbuster had the opportunity to purchase Netflix in 2000 for $50 million (Netflix is currently valued at $40.6 billion)

JASON L. HODSON IS A PRINCIPAL ARCHITECT WITH ALITHYA. HE FOCUSES ON THE ORACLE BUSINESS INTELLIGENCE PLATFORM, WITH PARTICULAR EMPHASIS ON THE FEDERATION OF EPM AND RELATIONAL DATA SOURCE, BUSINESS INTELLIGENCE CLOUD SERVICE (BICS), AS WELL AS DATA GOVERNANCE WITH HYPERION DRM. HE HAS EXPERIENCE WITH CLIENTS IN THE INSURANCE, PUBLIC UTILITIES, MANUFACTURING DISTRIBUTION, AND HEALTHCARE INDUSTRIES. A FORMER U.S. MARINE, JASON HAS AN UNDERGRADUATE DEGREE IN MATHEMATICS/PHYSICS FROM BALL STATE UNIVERSITY, AN MBA AND MS-INFORMATION SYSTEMS FROM THE UNIVERSITY OF CINCINNATI, AND A MS-INFORMATION AND KNOWLEDGE STRATEGY FROM COLUMBIA UNIVERSITY. HE CURRENTLY RESIDES IN DENVER, CO AND ENJOYS HIKING, SNOWSHOEING, AND THE LOCAL CRAFT BEER INDUSTRY.

ABOUT THE AUTHOR

JASON HODSON, PRINCIPAL ARCHITECT

About Alithya

ALITHYA GROUP INC. IS A LEADER IN STRATEGY AND DIGITAL TRANSFORMATION IN NORTH AMERICA. Founded in 1992, the Company counts on 2,000 professionals in Canada, the United States and Europe. Alithya's integrated offering is based on four pillars of expertise: strategy services, application services, enterprise solutions and data and analytics. Alithya deploys solutions, services, and skillsets to craft tools tailored to its clients’ unique business needs in the Financial Services, Manufacturing, Energy, Telecommunications, Transportation and Logistics, Professional Services, Healthcare, and Government sectors.

alithya.com | [email protected] | 914-253-6600 | 514-285-5552

C O N TA C T U S

Final Framework ConsiderationsAs with any strategic framework, organizational strategy, culture, and processes as well as regulatory factors should be considered and the systemic factor rankings evaluated accordingly. For instance, organizations in finance, insurance, and healthcare industries often have heavy government regulations that require them to provide, model, and audit their data in a more controlled manner. For these organizations, it might be essential to reduce the systemic factors ranking of Imagination and Policy in the user interface. Enterprises with a strong emphasis on research and development might find it advantageous to increase the ranking of imagination and policy in the data sources. Regardless of the ranking of each factor within the various components, remember that the analytical ecosystem should not be static. Implemented properly, system evolution should occur as a reaction or in anticipation of changing strategies and market conditions.

Recall that reporting is not analytics. Even so, it is easier to degrade an analytical solution to a reporting solution than it is to go the other way. A proper assessment of a governance strategy using the systemic factors of imagination, policy, capability, and management helps ensure the analytical solution continues to answer questions and create actionable insight. A key factor of assessment and process implementation is to remember that each technical component and individual requires a different approach to governance. Despite the seemingly myriad application of systemic factor rankings, proper application of this framework can lead to a holistic solution that sustainably satisfies an organization’s data needs.