Chapter 2
Executive Call to Action—How Chief Data Officers and Business Sponsors Can Empower Results
Abstract
This chapter asks readers to define the scope of the data they wish to improve and create an actionable plan that can be passed on to data stewards and executives. This is done through three key areas. First, it requires a clear, crisp, and consistently communicated data leadership message with the vision, scope, and targeted outcomes of the journey (implicitly or explicitly including improved data capabilities and behaviors). Second, users must develop a data leadership team with a chief, where appropriate, and officers who are empowered to lead and execute on your vision. Finally, the user must define a set of measures, methods, and models or templates that demonstrate the commitment to ongoing execution and progress tracking and rewards.
Keywords
Analytics; Big Data; Business sponsors; Data analysis; Data officers; Regulatory compliance
Executive Call to Action—What Do We Need From You?
First, develop a clear sense of data’s importance to you and your enterprise. Much of this importance stems from the risks and rewards that data presents:
1. Risks—Regulatory, Operational, Market, Client and Reputation, and Cost
2. Rewards—Data and Analytics Opportunities—Regulatory, Cost, and Speed or Competitive Advantages, and Sustainable Performance, Efficiency and Market Improvements
We discussed the value of data delivery and the commensurate value and importance of governing data delivery in the Introduction. The most direct path to proactively engaging data leaders in your firm is to establish a message that clearly balances what you expect in terms of risk reduction and performance and financial and competitive rewards. Second, construct a mandate and viewpoint for the way you want your enterprise to handle its data. Establish an expectation that, as part of the journey toward your risk and rewards targets, you expect the enterprise to permanently improve its data-handling behaviors, practices, and skills with appropriate, efficient controls. Establishing your controls perspective communicates a sense of having a “controls conscious management,” which is irreplaceable and critical to the enterprise.
Here are four major drivers behind the movement toward data leadership:
Each driver presents risks and rewards. Rewards are best defined and understood when tied to a committed enterprise strategy and goals in a clearly communicated mandate. For example, a firm commitment to outstanding customer experience or innovative products and services would identify reward areas that bolster those goals. Customer experience goals require tremendous customer insight and a consistently high reputation with customers. The corresponding data reward zones would be comprehensive data gathering, integration, and protection for shared customer experience insight and improvements in experience. These boil down to an emphasis on operational and competitive analytics as well as Big Data integration and exploration. A firm focus on continuous product and service integration would target Internal Reporting and Analysis for deep quality and knowledge testing as well as Operational and Competitive Analytics to define target markets and innovation focus. Increasingly, firms must balance between operational and regulatory issues and the desire to grow and innovate, and this balancing act poses the greatest challenge.
The bottom line for executive messaging and involvement is: articulate your passion for growth, innovation, and excellence. Share, in the same message, your continuing commitment to provide a “controls conscious management team” that supports the efficient use of corporate data and process
governance as a means of reliably reaching your goals and preserving your reputation. Let your organization know that you are not looking for herculean efforts to keep things running in “business as usual” mode while innovation and transformation efforts are underway. Rather, you actively encourage your leaders and teams to provide regular feedback on the pace at which major improvements can be achieved within a controlled and sustainable environment. The next step is a bit easier. Once you’ve laid out a vision and some guardrails for the journey, the next step is to help everyone understand how to get the scope of data and related processes right. Defining what is critical and therefore in scope for this work is a way of managing expectations and building confidence. People will see a clear sense of direction, speed, and course as a real enterprise commitment they can support.
So what happens to companies, government agencies, and nonprofits when the management team does not value controls? What about smaller firms or firms where lesser levels of controls insensitivity are prevalent, how do we improve on this for data?
Fair question. Some fail under the weight of the results of a consistent lack of controls. Others face repeatedly stiff fines, penalties, and other financial impacts from years of control ignorance. We’ve seen a spate of banks and insurance companies recently caught manipulating foreign exchange, interest rate, and related markets, even after a 10-year global recession wrought by uncontrolled speculation and leverage.
The other side of the equation is clearly the benefit story. In firms where controls are not on the agenda, benefits we would ordinarily target become the entire rationale for governance. Escalating the benefits discussion may require tying specific benefits to individual initiatives. Major initiatives that target enterprise, business line, or technology transformations often require multiple cost/benefit mappings. These mappings can provide sponsorship and funding to both “fix the data” needed as part of systems migration and “improve the data handling” to resolve underlying issues and introduce new controls. The table provided previously shows examples of some of the benefits that can be used to acquire sponsorship when control focus is not sufficient.
One additional cost/benefit issue can often be used effectively. Mapping out a data flow to depict the controls problems often shows that a multitude of controls are actually in place. The problem is not so much a lack of controls but a lack of controls effectiveness. Perhaps everyone rereconciles data they receive, even though the provider has already reconciled it to an authoritative source. Perhaps there are import filters for inbound files that reject records or entire files, but there is no profiling of these files before the import and no little or no exception handling for the rejected items. The opportunity to provide a management team that does not value controls with a way to actually reduce the number of controls used by simply improving a few key items often gains sponsorship.
Scope and Focus Area Definition
Defining critical data is the subject of much discussion. The truth behind this definition is so simple it’s a wonder why it takes up so much time and energy. Think about your data in terms of what it represents, as you normally do when discussing reports, analysis, and projections. What reports and numbers do you use to run your business and make strategic decisions? Operations executives, including operational risk executives, use a run book with all their operational and strategic performance indicators and other dashboard level numbers in order to see trends and issues. What is critical to you as a stakeholder, an officer of the corporation, and a leader? The answer we get once we simplify this exercise is remarkably consistent and clear: “I care most about customers, partners, and associates and am equally responsible for our financial and regulatory reporting requirements. These are the things I lose sleep over and must attest to as an officer and steward.” This is a solid, actionable list. The data we need to govern and improve for these responsibilities can be boiled down to three primary areas, as shown.
Many other domains of data may be critical to a firm. Examples include intellectual property, such as analytic models and results; clinical and other studies; analysis; and human capital data, including internal and external resource-protected data. The value of this data is most obvious where the quantity, quality, and timeliness of it are most impaired. Gaps in reporting and analysis, as well as material inconsistencies, indicate where valuable information is not being delivered as required.
Master and reference data are increasingly seen as critical data. Early examples included financial, organizational, and human resource hierarchies. These eventually expanded to include product, supply chain, market and customer hierarchies, and valid values. The master data focus is now centered primarily on customers, products, and suppliers. Financial systems now routinely provide integrated master and reference data. Master data refers to the objects of your business that you can master, including customers, accounts, products, and organizational structures or geographic locations. Reference data is generally specified outside your control, and you need to ensure you are using and reconciling to it consistently. Examples include everything from tax identifiers, to currency codes, to diagnostic codes, and so on.
There is value in the simplicity of a message about scoping critical data and related processes to focus primarily on customer, financial, and regulatory areas. Consider your partners and associates as part of the extended financial record, to keep this simple. Then you can communicate this initial focus so that everyone understands what to target for governance, stewardship, quality improvement, and related efforts. Other data subjects will be added over time, but in the beginning, it is most important that everyone understands where we are starting and why.
Critical Data requires some level of “attestation,” or commitment, from its producer or owner, usually a business executive. The more formal the attestation required, the more critical data controls, quality, and ownership transparency become to both the producer and consumer. Data that carries an implicit or explicit commitment is called “attestable data” and should carry with it some measures of its integrity, timeliness, completeness, and, as of its effective date, accuracy. Attestation submissions to external constituents, especially regulators, should be logged in the system of authority used to produce them along with their quality and integrity level.
Your contribution as an executive, sponsor, and mentor for your data leaders is a clear and consistent mandate. Your data leaders will then bring this mandate down to more detailed program and project planning. The “why” and “what” answers are what everyone needs from their executive leadership. You have a simple approach to providing the “why”—the four basic drivers mapped to your strategic goals and imperatives. The “what” starts with the identification of which data and related processes are critical, as we covered above.
The key is people seeing that an overall executive mandate and sponsorship is present, along with regular attention to progress and support in order to resolve issues as they occur. We know that sustainable competitive advantage requires sustained executive commitment, oversight, and active support. Improving your data leverage to become a data leader is no different.
Some background and context on data governance as well as stewardship and quality approaches and practices will be useful as you help guide and monitor the actions of your data leaders and teams.
Traditional Approach to (Data) Governance
The typical approach to data governance is framed using standard governance approaches. We’re all familiar with the pillar model of governance. Each pillar represents a core domain of governance activity, within which specific controls and behaviors limit the use of certain assets, while managing access and change permissions to ensure asset preservation. Data governance and quality programs typically leverage this model, providing activities for overall program governance, data stewardship, data cataloging facilities, and data-quality testing and improvement. All of these are appropriate domains to focus on for data management; however, they do not function discretely, and most organizations have moved beyond using finite projects and programs to institute them.
The traditional four-pillar approach to governance is also very monolithic. The model presumes that each pillar will stand in a permanent manner independent of the others. It also assumes that, once the roof is installed, the pillar will be standing with the necessary strength and size to hold its share. For these and many other reasons, this model is increasingly outdated. It is however, more prevalent than many of us realize, since newer models have yet to be fully deployed. The other challenge is that, in many organizations, these models are not uniformly defined or deployed at the enterprise level. Acquired companies, now run as and pockets of “legacy” systems and stores, each have their own governance model or activities. Many look more similar to this monolith than they do to newer models and approaches. It is wise to review the various models, efforts, and resources in place, rather than assume they are up to date.
However, we know that improving data and its outcomes takes a much less monolithic and much more iterative set of actions. Therefore it’s important to be able to support different levels of resources, activities, and timings of outcomes across various parts of the enterprise in a staggered, parallel execution model. Some areas will start sooner and progress more quickly because of a business or regulatory imperative. Other areas may delay their start or extend their duration because they have more business process or application changes that must be made. There is also the overarching set of enterprise activities that must occur, even as these domain areas begin to be executed. Thus it is worth pulling out the organizational governance pillar and understanding that it rightly belongs in the center of the cycle. It’s also important to understand that the organizational governance function will mature at its own rate, sometimes letting domain-specific activities that are being conducted lag behind.
A strong movement toward a more sustainable cycle of data management is emerging. This cycle is driven by a growing recognition of the complexity of the corporate data landscape and the regulatory and stakeholder demands it must address. Cyclical data management takes advantage of two trends in business today. The first is the use of agile or rapid iterative execution models to produce faster results and engage key resources more efficiently. The second is a better understanding of how to achieve a sustainable competitive advantage using data and advanced analytics capabilities. There is clearly a growing sense of urgency among executives about the need to leverage data more quickly and with better integrity within operational and strategic analytics areas. This new requirement level has to be met by both legacy operational systems and emerging Big Data sources. In all cases, there is a heightened need for data-quality visibility and improvement, which brings us back to using an iterative approach on a continuous basis.
This expanding set of challenges leads us to speak with clients, vendors, and analysts about the need for an approach that handles the growing wave of Big Data-driven analytics. The new requirements for integrity and attestability are one dimension of the challenges involved; increasing the volume and variety of your data is the other. Much has been written about competitive analytics capabilities. A long period of sustained investment in enterprise resource-planning suites and enterprise performance-management systems has driven a substantial change in the way analytics are used in business. Today, it is safe to say that much of the analytic results being produced are consumed by systems making, decisions about inventory, pricing, channels, customers, and so on in a closed analytic loop. This closed analytic loop raises the bar on data integrity and quality requirements. Analytic engines and algorithms now routinely feed operational systems, including enterprise resource planning and financial management systems, with near real-time updates on pricing, shipping, and customer retention functions. The automation of this process does not allow for human intervention or quality control over the data in use. So the data must be controlled, and, where necessary, improved in quality enough to be attestable and auditable.
We use the term “attestable data” to reflect the commitment corporate officers and executives make when they rely on data and the results of closed analytics loops in their decision-making process. This includes pricing decisions, product decisions, and large-scale acquisition and re-organization decisions. Each of these decisions represents a commitment of resources and reputation. All of them are based on data and analytics flows so must bear the weight of attestability.
The move from “information as an asset” to “trusted data” to “attestable data” is a transition driven by reputational as well as financial risk. “Attestable data” is less apt to vary in meaning across different people than terms like “trust,” “golden,” or other terms currently in use. Attestation requires an executive commitment under specific terms and conditions that can be held against an executive if their data is subsequently shown to be flawed. The level of commitment is more specific; the attestation is that both the data and the controls around its production are adequate to all regulatory requirements. This is a higher-level commitment reflecting a higher level of scrutiny and control.
Taking a more integrated and sustainable approach still requires addressing all four of the core governance domains. In order to ensure that we address each domain, we must first articulate each domain’s challenge so that we know what action is required and can gauge the overall support needed to succeed across domains. Here we see the four domains displayed more in terms of the nature of the challenge than in terms of monolithic success. We recognize that whether you start with a single project, multiple projects, an enterprise program, or a dedicated enterprise function, there is an organizational dynamic that we must have in place to begin.
To stand up any programmatic or functional business capability, you must have executive vision, sponsorship, communication, and ongoing support. In this case, you also require solid guidance from key people in the organization who have a clear understanding of and experience with the data challenges, as well as an understanding of the benefits of resolving those challenges effectively.
It’s important to engage what will become your first-line data officers early in the germination of a data program or data function. There is always time to adjust membership, assignments, and responsibilities as people’s other responsibilities and your needs dictate. Building a small group of data leaders with whom you socialize your strategy and vision is an essential building block for your success. Part of that socialization will be understanding what roles and responsibilities are going to be important both initially and over time. There is always a difference between building something and running something, so your small group of leaders will need to include folks with strengths in each or both areas. This is also where we see the integrative nature of these four core capability areas.
Your data stewards will be doing the work to uncover data and its issues, catalog and resolve its meaning, and ascertain and improve its quality. Understanding what has to be done in data cataloging and data-quality areas is essential to identifying the right data stewardship function and to assigning responsibilities of that function to the best people possible. Also critical here is the notion of cost-effective data stewardship. This concept flows from cost-effective computing, a discipline John Clark and members of Gavroshe have applied globally for two decades. In the context of data stewardship, it is applied to the selective use of part-time resources under the guidance of full-time officers. We do not ask stewards to be full-time, proxied owners of data and take on the underlying problems that come with it alone. Rather, we provide leadership to data officers, part-time data stewards, and subject matter experts in business and operational areas so they can work together to govern and improve their data.
Getting to cost-effective stewardship requires a balance of training, mentorship, and experience. Traditional training approaches have given way to more effective and repeatable methods. Classroom training is still used, but leaders have moved toward using a “Khan Academy” approach with Learning Management Systems (LMS) and other online means of delivering knowledge and supporting collaboration among the stewardship community. Finally, the use of expert mentors from inside the organization as well as from consultant experts is very prevalent. Using a combination of training, experience, and outcomes is an emerging standard. The belt system popularized by Six Sigma and other enterprise programs allows us to distinguish stewards not just in terms of their job title or position but also in terms of their belt status. An example is shown in the following table. A simple approach like this one communicates the value of learning and achieving.
Our first order of business in creating a more sustainable or cycling view of this process is to pull the organizational construct out of the pillars and put it at the top of our framework. This format expresses that the organizational construct is an essential, creative element for our program or function and must be created and managed distinctly from the work streams that it mandates. The result is a framework that describes three key aspects of improving data through governance, all of which are driven under a common executive mandate and organizational construct as shown.
Data Leadership Cycle—Framing Your Mandate
Beginning with the end in mind is a common proposition when dealing with enterprise issues. As you consider the type of controlling and improving data mandate you want to provide your organization with, it’s very useful to consider what your peers are doing across industries and where similar programs and functions have already gleaned substantial benefits. Identifying the benefits and risks you are
targeting with your program or function is a critical first step; then tying those to each of the three core areas becomes fairly simple.
Mapping some of your most challenging data-related initiatives and imperatives against the three core areas often highlights the need to deliver multiple benefits. Anti-Money Laundering (AML), Watch List Monitoring (OFAC), Financial Planning and Analysis (FP&A), and Comprehensive Capital Analysis and Review (CCAR) are all examples of programs required in financial services firms that demand multiple data capabilities and outcomes. Data for your imperatives must be traced, cataloged, and tested for integrity regularly, because the outputs of these programs are regulatory reports and filings requiring executive attestation.
ITAM, or IT Asset Management, Information Security and Privacy (CISO/CPO), and Governance and Regulatory Compliance (GRC) are all ongoing corporate functions with similar data requirements. Legal Entity Identifiers and Relationships (LEI/LER) are emerging requirements for financial services providers enacted to support Counterparty Risk Management (CCR) and will ultimately require coding each transaction with these identifiers.
These are just a few of the growing regulatory requirements being levied in the financial services industry, as regulators establish new guidelines and requirements in the aftermath of the great recession. Leading firms are increasingly tying these programs, functions, and requirements to broader enterprise data services. This is the first data leverage point in new programs. Enterprise data services provide everything from initial data tracing and cataloging to data-quality testing and dashboards. Advanced services can ensure that data sources, quality, and issues are logged with the data so that when reports and filings are prepared, prior to attestation, all the conditions of the data being used are exposed. These enterprise data services are often provided using lean, agile methods and have been shown to save between 2.5× and 4.5× the cost of simply allowing lines of business and individual initiatives to provide these outcomes independently. The expertise, tools, and specialized methods required by these services are far more efficient when housed in a shared or enterprise data service function.
The second leverage point data leaders are using is the integration of multiple programs or functions to share both enterprise data services and the data they each generate. Thus savvy risk management executives have learned to ensure that their GRC systems integrate, in nearly real time, the data feeds from their AML and Identity Theft programs. This process is critical and efficient, since a red-flagged account in Identify Theft requires enhanced due diligence in the antimoney laundering area. The fact that both areas are using attestable data from documented sources with proven controls further streamlines the entire process and provides the highest possible credibility with regulators, as shown.
Risk, Cost & Benefits Examples from Financial Services Industry
Taking a More Sustainable Approach
We see Business Data Governance as a function of business data ownership, production, use, and interchange. Similarly, we see other corporate governance functions over cash or marketable securities as ways to control changes in their location, condition, use, or ownership. Our opening quote from Walter Wriston indicates our simple, direct focus on this issue of data as an asset. In fact, that data is almost as valuable as the asset it describes. Thus data about money, income, expense, or risk is of critical value to the enterprise.
We treat “business data governance” as a noun representing a business function, just as one might treat the term “corporate governance.” “Governing business data” is the verb and we treat that very expansively. We include the many facets often separated into their separate subtopics like stewardship, quality, cataloging, relationship management, and so on. More on this later—this is just a callout to establish some basic operating assumptions, so we can make sure we are speaking the same language.
Data Leadership also seeks to leverage organizational leadership. We’ve seen the rise of dedicated data roles and positions including stewards, officers, and scientists, as well as long-term data programs focused on governance, quality, and analytics. Some firms have fully committed to organization functions and to units dedicated to governing and improving data. These groups often spring up in finance, marketing, and other crossfunctional areas of a large business, since these areas are recipients of all upstream data behaviors and problems of the enterprise.
Data Leadership Roles—Leveraging More Than the CDO—The Power of Data Officers
The Chief Data Officer, or CDO role, has come to the forefront in recent years as the person who is accountable for data outcomes. Recent elevation of the digital experience has resulted in the use of the CDO title to also refer to Chief Digital Officers, so the title can be confusing, and many firms choose to use other titles for this role including “Global Head of Data.” All these titles and positions underscore the importance of data and provide for data leadership positions. It is worth noting that different industries are highlighting this role in different ways. Media, creative arts, and retail industries, with heavy content and user-experience domains, tend to reserve the “D” for “Digital” and use “Global Head of Data” titles. Financial services, on the other hand, emphasize the CDO as one of their key positions to deal with regulatory concerns. Whatever the title, the role and sponsorship are clearly understood to be critical to enterprise data management and governance.
Different styles and approaches are emerging in data programs that reflect the mandate and focus of their CDOs. Some basic patterns we now see in financial services, healthcare, and other industries include:
1. Regulatory and Compliance (Controls Focus)—Enterprise financial and regulatory reporting must satisfy a myriad of complex, multiagency regulatory requirements. Meeting those requirements depends on executing a complete, consistently applied, and continuously monitored set of data controls for all source data used in these reports.
2. Data and Reporting Improvement (Data Transformation Focus)—We need to improve our data management capabilities, behaviors, and outcomes in business-unit and enterprise-level reporting. We have clear and compelling evidence that many of our operational challenges and cost bubbles are occurring because of underlying data-handling issues. We are therefore committing ourselves to changing the way we handle our data and produce our reports.
3. Analytic Competitiveness (Analytics Leadership Focus)—Our business and industry is increasingly driven by analytics, both operationally and strategically.
4. Hybrids—Transformational and Analytic Leadership-based efforts can often merge or evolve to address both of these areas. Control-focused efforts are generally relegated to the core controls and quality issues they start with since their mandate and resources are so constrained.
Chiefs of Data, Governance and Analytics
The CDO is very much in vogue now, even as we simultaneously embrace highly commoditized technologies and methods for data management. Cloud-based computing elevates virtualization and hosting to new levels, while further removing day-to-day management of data and analytics platforms. Look no further than
Salesforce.com and its effect on entrenched competitors like Oracle, Microsoft, and IBM to see the impact of a new class of cloud-service components that grew from the Customer Relationship Management seedling. Pending merger and strategic alliance conversations are increasingly public, sometimes resulting in brief stock exchange trading halts and always driving uncertainty into cloud-computing service contracts. Data executives watch these market forces and actions with a great deal invested in their outcomes. Executives have a growing voice in the contracting and strategy for use of cloud-based analytics environments.
The wide range of interests and responsibilities that executives have has also driven a growing array of chief titles and positions. The CDO role is well established and new positions, such as the Chief Analytics Officer, Chief Data Governance Officer, and Chief Data Risk Officer, are emerging as hiring priorities. The type of organizational construct needed or available to enable these positions is less clear. Each of these titles is really an offshoot of one of the primary responsibilities of the original crop of CDOs. Some CDOs have made names for themselves as governance leaders, most notably in money center banks, where they used self-assessment and top-down approaches to show progress in basic data governance areas. Others have shown a more holistic and business-benefit-driven set of capabilities, embracing the business directives and requirements while working to satisfy compliance and risk issues. These CDOs work hard to engage and support business partners with realistic data improvement efforts that are both measurable in their outcomes and sustainable over time. The rare CDO is able to address both data and business process, sometimes covering business process areas like counterparty risk management and credit-pricing leadership.
Some newer CDOs have come into their positions with two key assets. The first is a direct reporting line to the Chief Operating Officer (COO), ensuring consistent business engagement, sponsorship, and demand for enterprise data services. The second is a keen understanding of and often decades of experience with data integration, warehousing, reporting, and analytics delivery. These CDOs also understand the need to engage their organization with a new structural approach. They are usually asked or given the opportunity to collect key groups of data modelers and architects: data integration, warehousing, and delivery teams; data stewards and quality experts; and analytics and Big Data experts, which they then form into a coherent team with the CDO. These CDOs usually move quickly to identify a line of “data officers” who lead each of their respective areas, while collaborating as a team with the CDO and his or her constituents across the enterprise. In this emerging best practice model, the CDO establishes the Analytics Officer as the leader who provides the “voice of the decision” for the business. This voice provides transparency and control over the analytic algorithms, models, and integration points across the enterprise, giving business partners a clear picture of their analytic capabilities, options, and outcomes. In some places, there is a Chief Analytics Officer (CAO) who reflects a return to the notion of decision support with the new facet of decision/action/outcome traceability. CAOs will move us beyond closed-loop analytics and machine learning and into organizational learning that binds business intelligence with behavioral intelligence. It is worth noting that the “chief” qualifier may not be the most effective way to establish this role.
Demand Management Model
Demands are most often clearly expressed at the tactical, project-oriented level. These include demands for data, data resources, data services, or data outcomes that must be compared with available resources in an open-resource management model. These are often called Demand Management Models, or Offices, and this where the allocation of resources is made. Demand management efforts start at this level through listing demands and are often composed of nothing more than disparate projects and issues people are trying to resolve, as shown.
Once these issues are collected and we have a sense of their size, current resource load, project timing, and priority, the current state of data conditions begins to emerge. This is still only a partial, if imposing, picture. The CDO still needs to collect broader management initiatives that often span years and sometimes business areas. Summarizing those initiatives adds another layer to the picture:
These management data initiatives have usually been created, funded, and resourced based on a single line of business or an operational area’s needs. The initiatives may be “burning platform” projects, transformation programs, system migration, or virtualization programs, and even cloud migrations or IT outsourcing events. It is important to determine the extent to which an overarching business-driven data strategy and planning model was used to approve and fund these initiatives. It is also critical to
identify the level of shared enterprise Program Management Office (PMO) services in place and that are coordinating the execution of these programs, the use of any existing shared services, and the calibration of resource requirements over time. The PMO can also help identify what level of agility, reflected in the current development and deployment methodologies, these initiatives follow. Finally, the PMO should include any shared or common success metrics and updates.
The last layer of the demand management model is the strategic piece: the enterprise vision-driven actions with data and analytics that are planned or already underway, as shown.
When you add these strategic priorities and actions, a complete picture of the enterprise data landscape emerges, which you can improve over time. The first impression is that of an elephant standing on its head: all the weight is teetering at the top on a very narrow bottom. This is very typical, since most, or even all, of the activities and resources previously committed across the enterprise for data-related issues and needs were not guided, directed, or even evaluated by a data executive. Turning this picture right side up is something an enterprise data executive can do when supported by the top enterprise and business executives. Obtaining and maintaining that support requires a clear, transparent approach that puts business executives’ representatives at the table, ensuring prioritization and changes to human, financial, or political capital are made equitably. A more effective, efficient, and sustainable model emerges from this effort over time:
The method behind this reversal and aligning process is detailed in Chapters 3 and 4. In these chapters we focus on the methods used to provide strategic alignment, so you can get the top of this pyramid right. This Demand Management function is a critical component in your data leadership effort and will dovetail with existing portfolio management processes once in place. Providing the strategic mapping
for the top of the triangle is accomplished by the Mission Map, a simple tool that serves to capture and communicate strategic thinking and mapping to measures of success.
The top level of the Mission Map is the mandate, as described at the beginning of this chapter, worked out by the CDO and COO or CEO. Large financial services, health and life sciences, retail, technology, and even federal agencies now routinely enjoy the engagement of the CEO, CFO, and other executives in strategic data planning. This is not the same thing as the typical “data strategy” approach used in IT. Data strategy is critical to building development and deployment roadmaps for data architecture components. Some data strategies even address data services and resources, or data virtualization approaches, as well as Big Data integration efforts. These are all critical activities and outcomes that are increasingly managed under shared leadership with the Chief Data Officer (CDO), Chief Information Officer (CIO), and others.
However, the CDO Mission Map operates at an even higher level, providing key executives with the ability to direct alignment of all priorities, commitments, and targeted outcomes across the enterprise and over long-term horizons. This is the arena of Supply and Demand Management, or the Demand Management Office. CDOs are able, after assembling their data officers and functions, to identify a demand management model for data resources and services that is at once simple enough to communicate across the enterprise and complete enough to ensure a fair and open allocation of capital. The first order of business is to align the DMO planning and commitment process to the strategic plans of the enterprise. The CDO Mission Map can be constructed on one page, capturing the mandate for the CDO from the enterprise and deriving the goals, targets, measures, and requirements for success from that mandate in each of three to four clearly defined functional areas. Savvy CDOs have learned to start with their programmatic requirements in this Mission Map in order to highlight for their executive sponsors what it will take to provide all of the initiative and service-based delivery support required by the enterprise. The typical Mission Map divides a CDO program into two to three functional areas, generally focused on services or outcome-based capabilities. As a result, a well-formed Mission Map becomes an equally effective planning and communication tool. Many clients have cascaded mission maps the way strategy mapping suggests to support an enterprise strategy Mission Map at the top level, followed by the CDO Mission Map, and then maps completed by each data officer who leads a functional area of the CDO. Since each of these maps contains measures and requirements, they can be used for performance and resource management. Communication is a consistent theme in the data arena, so the Mission Map is an excellent way to enhance that process and engage each level of the organization.
Modeling Effective Communication
The best way to communicate vision is through personal and professional experience. Executive stories are often a powerful way to share and model effective communication, especially when we are trying to change behaviors across the enterprise. That behavioral change is key to this messaging and leadership: the recognition that, like so many other challenges, data leadership is an organizational change-management journey. Data is always the result of behavior, usually a combination of automated and manual efforts with manual oversight and adjustment. Changing those behaviors and the design patterns and standards used to automate them requires systemic behavioral change.
We have a couple of stories from executive messaging and leadership experiences to share with you that shed light on what has worked well. Each of these stories is from a real-world client and occurred recently enough to relate practical insights.
Large initiative review and alignment
Storyline: A C-level executive cuts through the noise to deliver a simple, three-part test for large, complex programs and then applies it consistently to enterprise class initiatives:
Mary, an IT and Operations Executive in a large asset management firm, started a review of the current portfolio of initiatives upon joining the firm. She knew how critical it was to get a handle on the committed work she inherited and to get a sense of the health of these initiatives as well as the expectations of her sponsors. Many of these were more than a year old, with large, some even very large, spending and investment levels. She had seen this trend before and was always concerned about inheriting an aging portfolio of initiatives that could not be clearly traced back to a solid demand management process or executive sponsorship and business case foundations. Her experience indicated that she had very little time, perhaps only a few months in which to engage her constituents positively. She also wanted to highlight the process they used to consider and review these major commitments. She had learned, many times over, the importance of these initiatives, especially multiyear programs. Past experience made clear the value of large, complex programs and she wanted to share this with her leads and her constituents: “large, complex programs often highlight organizational capability and leadership issues.” While the focus on capabilities was immediately obvious, the leadership dimension was less clear. Her approach indicated that her thinking about effective leadership was rooted in the quality of the decisions and commitments made by leaders and thus required clear decision-making and communicating. Therefore she decided to provide an incredibly simple and clearly communicated three-part test for these programs. It was something Mary routinely asked when anyone requested her sponsorship of any major program or initiative. True to her values, Mary also used it to discuss supporting major initiatives demanded by her peers. As a result, the approach to enterprise class initiatives was utterly consistent, and, over time, she began to expect it from everyone who approached her for support.
The three questions/mandates were:
1. Who is the business person “pounding the table” for this project, and why?
2. When have we effectively executed something of similar scope and import? Are we including people and processes that served us well before?
3. Does our execution of this initiative take us further down the path we have set for ourselves, or does it take us off that path in some way?
Mary then reviewed her portfolio of business and technology initiatives to determine which satisfied these criteria and which required some remediation. In each case, she was able to explain her reasoning to the program sponsors and how they could work together to align with the program over time. In each case, there were gaps or hurdles, but the sponsors and leads understood the need to raise the bar on such large and impactful commitments. Some programs were found to be outside the three-part criteria and were merged with others. A couple of very large programs, already well underway, required interventions to put them on track with enterprise direction, design, and development patterns. In less than 6 months, the entire portfolio had been subjected to a simple set of tests where everyone recognized the alignment of the remaining programs with all three mandates. Even more impactful, no new initiatives had been approved or proposed without these three critical criteria. This executive and her team successfully changed key behaviors with corresponding improvements in outcomes.
Similar tests from C-level executives follow this pattern and ask more data, controls, and risk management centric questions, including:
• Will we maintain control and ownership of our data when we do this?
• What visibility do we maintain over the custodianship and handling of our data?
• How has this program engaged internal audit and risk management in its business-case and high-level solution design?
• Does this project leverage our enterprise data services or attempt to provide data cataloging, testing, and controls alone?
• Does this program intend to leverage our standard data change and quality controls or create its own from scratch?
In each case both the correct answer and incorrect answer were fairly obvious. Communicating these to sponsors who routinely evaluate requests for new initiatives from their teams conveys these expectations early enough to stop the noise from rising to executive levels. What questions do you routinely use to ensure the quality of major initiatives, and how do you communicate them as expectations at early stages?
Executive experience: Chief Information Officer, Fortune 100 global manufacturing firm
Context: A CIO was recently installed from outside the firm and recognizes the need to galvanize new thinking, behaviors, and outcomes with data. He had already started an IT Transformation program aimed at unifying key IT services and components with a vision of service excellence for business partnership. Then, he saw the opportunity to add data as a service to this vision.
Storyline: We can improve our results with improved data—but must improve our data habits and behavior first.
“I’m getting to know more about our teams every day, and today is no exception!”
The CIO used this opening to engage a large group of IT leaders, managers, and team members assembled for an announcement about the new data program. The participants were selected because of their teams’ involvement in key initiatives that produce, integrate, and rely upon enterprise data. The CIO knew he had a tall order because little attention had been paid to these people and their relationship to enterprise data practices and outcomes. Even worse, enterprise data practices and outcomes had not been addressed, formalized, or even explained by his predecessor. So he used a tactic he knew would resonate with these crossfunctional teams. His IT leadership team had been hard at work formalizing and socializing Enterprise IT Services, so IT would better align with the needs of the business and be recognized for superior performance. His team had developed a solid foundation of critical infrastructure, application, and integration services and was rapidly reforming key teams around these core competency areas. He and his leadership team had already started mapping key data services into this foundation, working to identify key managers and team members who had deep experience addressing data problems from “cradle to grave,” while constantly under the gun to deliver new systems and upgrades linked to business capabilities. The progress his leadership team was making in mapping out data services suggested that certain key people would be needed to lead new data services, while many others would need to better understand the role of data services and improve their “data habits” to rapidly support better outcomes.
He then told the leadership team: “You’ve all been pressed very hard for a very long time to deliver new systems and upgrades that support improved business capabilities. You’ve delivered these while constantly dealing with difficult data issues at every turn. Some of these deliveries were data centered, requiring massive data integration and delivery. Those have taught us that data deliverables are becoming a critical business requirement as part of enhancing their capabilities. We even have business partners from finance, marketing, and design clamoring to partner with us to deliver rapid, scalable, and sustainable data-driven capabilities. So, the business has learned that data capabilities are as core as any other automation-based capability.”
There, now he had the basis for his challenge and sponsorship offer—he was ready to make the ask and offer all in one: “…so now we embark on a shared journey with our business partners to deliver, in tandem, data-driven outcomes using a set of standardized services, skills, and teaming models.” He was prepared to support this claim directly. The CIO had engaged a key business partner from marketing who was respected across the organization. She was known for driving efficiencies in the way of marketing-sourced and used data, especially from outside providers, while simultaneously holding the line on some of her group’s “shadow IT” desires. As the CIO introduced her, everyone began to realize this was no longer just an IT exercise; it was something unlikely to fade into the background like so many previous data projects. She had prepared her remarks very carefully to fully support her CIO and his commitment to partner with her organization at every turn. She knew she had to engage her IT teams as never before.
So, she started with a simple commitment, first to the business goal and second to her IT team: “I am here to ensure we, as partners in data, will deliver best-in-world data driven outcomes for the enterprise. I am also here to share that commitment with you exclusively. That’s right, you’re it, I mean IT! You’re our IT partners, so we are here to collaboratively build a new way forward to leveraging our data.”
She knew she had to continue; she had to tell enough of the story to get some acknowledgment from this critical group of partners.
“We must know more about our customers, markets, competitors, economics, performance, risks, and opportunities,” she continued. “We must know much more about these areas with far greater speed, accuracy and completeness. Our speed to market is in all our hands because it is driven or constrained by our speed to delivery! You’ve all been deeply involved with major programs that enhanced our operational capabilities. In recent years, you’ve delivered on our emerging analytic needs and have helped us understand the challenges and risks inherent in our enterprise data. It’s time to step up our game by a factor of ten; we can’t get by doubling our data flows and technology inventory or manual efforts with data every year or so, as we have been. We must make a leap forward together and fast!”
She paused and held her breath for a long minute before seeing signs of recognition and acknowledgment. Her IT partners did understand the critical role of data to the business and the challenges in meeting her requirements. They also seemed to care about helping, as many team leads and members were shaking their heads and nodding assent toward one another. Now for the one-two punch: “I am here to offer you a new level of partnership…IF you will craft and deliver the best data in the world and the services to scale that delivery across the enterprise then…I will commit to our exclusive use of those services as they become available.” She went all in based on her trust in her CIO, his organization, as well as her desperate need for competitive data and analysis to help grow the business. She didn’t share what everyone already knew, at least to some extent.
Now the CIO needed to expand on his intent. He stepped forward, standing shoulder-to-shoulder with his business partner. He began by explaining his intent so he could showcase her willingness to collaborate: “I asked our key business partner to share her time and her story with us so that we might all begin a sea of change in our thinking about data. I really thought she was going to share the nature of her challenges and maybe ask for some help or offer to engage us once we organized our thinking. She has far exceeded my highest hopes and provided us an opportunity to drive enterprise impact that everyone sees, benefits from, and builds interest in leveraging for themselves. We could not hope for a better first and best client partner!”
He reiterated the challenge, sharpening the request she had started to make of his teams: “What our marketing partner offered is a direct, exclusive partnership. What she requires from us to make that work is a dedication to provide best-in-world data and ongoing services that scale to meet all the needs of her and her peers over time.”
Now he knew to stop and let that sink in. He saw his leaders surveying the assembled crowd and felt the pride of a leader with strong teams and partners. He felt certain that this approach to open, dedicated partnership would engage his IT teams and his partner’s marketing teams in a sustainable and scalable way. He also recognized that he and his marketing partner would need to demonstrate and reiterate this commitment across the enterprise constantly and over a period of years to change the game permanently. He knew they were committed to doing just that.
This firm went on to globally compete and beat every major competitor for over a decade, surviving the global “great recession” with a solid balance sheet, growing domestic market share and steady margins and financial performance. This firm also held on to its key leadership talent for an extended period, well beyond its peers and competitors. Did data do that? No—they simply treated data the same way they did everything. They treated their data as a “performing asset” that merited measurement and scalable, dedicated services to produce the required returns.
Everyone says executive support is critical for effective business data governance and management. We agree but find that the path to gaining executive support is not always the same, nor is it always direct.
In addition to these executive stories, we have interviewed several midlevel managers to see what their current perspective is on three simple questions:
1. How do you see data governance expanding its scope and value as it matures into a standard practice?
2. How do you see the value of standard, repeatable methods and processes as companies deploy data governance?
3. What is the typical level of adoption for your data playbook or operating model? Are you embedding them into data governance solutions from vendors like Collibra, Adaptive, IBM, and SAS? Are you using e-learning, wiki, or other knowledge-building or -sharing solutions to build enterprise knowledge and practice?
The answers surprised us in some cases and reflected an overall deepening awareness of the enormity of the challenge embodied by data and analytics governance.
Managers from firms with repeated iterations of data programs indicated marginal improvements amid repeated strategy, roadmapping, and planning efforts. There was a general sense that standard methods were important and valuable, but there was little commitment to exposing them through data governance or knowledge coordination tools. Some of these firms are experiencing pockets of data governance advancement based on regulatory or risk management issues they need to resolve. Current examples include antimoney laundering regulations, HIPAA compliance, and information security overhauls. None of these soloed efforts were accepted or sponsored by executives as the enterprise standard or mandated approach. Technology managers with proxied business stewardship rights were handling most of the efforts.
Firms engaged in deploying data governance programs with executive support and specific business goals are fairing better; they are also leveraging either data governance or knowledge coordination solutions to expose and expand standard methods and practices.
Firms that have an executive sponsor, dedicated data officer, and standardized methods, practices and measures are achieving enterprise results with cost-effective resource leverage. They also exhibit a technology-driven commitment to provide enterprise data services that are best-of-class and operate at efficiency levels of three or four times what local technology or business area activities can attain. These are the award winners enjoying consistent data improvements and business impacts.
We reached out to data governance solution vendors to understand how they are helping clients and what challenges clients want help addressing. These vendors were willing to give us even greater insight into how some of their clients are becoming data leaders. We have been fortunate to be engaged by clients to help with each of these and many other solutions in use in data governance programs.
Our dialogue occurred with vendors such as Collibra, Oracle, and GlobalIDs. Stijn (Stan) Christiaens, Collibra cofounder, was quite clear about the level of benefits his clients were experiencing as they adopted and deployed a standard operating model for data governance. We confirmed this with two different clients and found two additional prospective clients engaged in contracting with software vendors, who identified the acquisition of data governance software as their turning point. The specific benefits these clients were targeting were all requirements of a sustainably successful data governance program: identifying, defining, and taking accountability for critical data and providing a common operating model for its control.
Our work with Oracle product management, sales, and engineering teams consistently points to two key benefits their clients expect and generate. The first is the ability to manage complex master and reference data hierarchies for use in analytic and operational systems, which presents an advanced data change control challenge. The second is much more broadly impactful: the sense of confidence that multiple business areas gained from having a clear and simple set of data governance workflows to follow.
Dr. Arka Mukherjee, founder and president of GlobalIDs, shared a set of benefits his clients continue to realize several years after implementation. His approach provides ways to automate massive data discovery and relationship mapping, a critical set of functions for data governance programs to trace their data assets and lineage. Arka says that automating discovery and mapping dramatically accelerates the process of identifying “authoritative sources” of data and consolidating lesser copies to improve attestation and reporting outcomes.
Two basic themes emerge from all of this client and vendor feedback. The first is that there are now multiple capability areas that data governance and quality programs must deliver. Cataloging data, definitions, and stewardship assignments is a critical first step but the addition of data-quality requirements and testing or profiling outcomes is also a requirement. The second is the increasing sense of value large companies are finding in the use of data governance software-based solutions to support a robust and continuing success in their programs.
Summing It Up
Your data leaders, teams, and the entire enterprise need your leadership in three key areas:
1. A clear, crisp, and consistently communicated Data Leadership message with the vision, scope, and targeted outcomes of the journey (implicitly or explicitly including improved data capabilities and behaviors).
2. A Data Leadership Team with a chief where appropriate and officers who are empowered to lead and execute on your vision.
3. A set of measures, methods, and models or templates that demonstrate the commitment to ongoing execution and progress tracking and rewards. In other words, a Data Leadership Playbook.
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.