CMMI Institute

Newsroom

The latest information for media, analysts, and others interested in the CMMI® Institute and process improvement.

Data Professional Introspective: Organizational Factors Impacting Enterprise Data Management

By Melanie Mecca, CMMI Institute's Director of Data Management Products & Services

“Money won’t save you, no, no, no, no, no.” – Reggae artist Jimmy Cliff, 1974.

In fact, money, insofar as it underwrites the complex data environment that well-heeled organizations typically possess, can be a challenge to data management success. On the one hand, the EDM program needs continuous, non-discretionary funding to centralize and integrate sound, repeatable data management practices across the organization. It takes funds to establish data standards, develop a business glossary, populate a metadata repository, provide compliance capabilities, and implement data quality rules and tools.

On the other hand, easy access to funds can lead to proliferation of operational data stores, repositories, toolsets, numerous best-of-breed data management technologies that are difficult to integrate, projects that overlap in data scope, etc. These consequences are even more likely when lines of business control their own funding for data and IT efforts. But – (fooled ya!) – we’re not going to talk about money in this exploration of issues that organizations may face in evolving successful EDM programs.

In working with many organizations to evaluate their data management capabilities against the reference model framework of the Data Management Maturity (DMM) Model, I’ve encountered numerous organizational themes/scenarios – both positive and negative – that impact organization-wide data management progress. These factors affect organizational operations and the likelihood of success for data management. Positive themes can be characterized according to these three qualities:
  • Aspiration – in an organization, the degree to which staff and leadership share a sense of mission, purpose, and ambition to achieve goals.
  • Judiciousness – while ‘strategy’ is a plan for achieving a specific goal (or goals), judiciousness refers to good judgement, which is the ability to form opinions and undertake actions through evaluation and discrimination (e.g., what is both valuable and achievable).
  • Lucidity – clarity and ease of understanding in thought and communication, increasing focus, and minimizing ambiguity. With those positive qualities in mind, let’s look at three representative scenarios that organizations frequently experience, where these qualities are not strongly developed.
We’ll describe them, and provide suggestions for organizational activities and changes that will give the EDM program a better chance for success.

Discouragement About Resolving Core Data Issues –Failure of Aspiration

The attitude of discouragement – an expressed opinion on the part of staff that “it can’t get better” – is often encountered in a variety of scenarios, for example:
  • The organization is highly distributed, either through a diverse business model, multiple mergers and acquisitions, or multiple geographic locations – staff feel that there is no clear path to resolving complexity. 
  • The organization has recently undergone a reorganization and the lines of influence and authority are not clear – staff members don’t know who is responsible for what data. 
  • Senior management has not focused on improving the organization’s data layer. For example, there may be many overlapping operational systems containing the same data, but there has been no real effort to integrate them – staff members believe they have to live with the existing business problems. 
Needless to say, this situation, whatever specific conditions combine to cause it, is inimical to effective data governance and data management initiatives in general.  The antidote for discouragement – Failure of Aspiration – is three-fold:
  • Creating a data management strategy (mission, goals, objectives) aligned with the business strategy and communicating it across the organization.
  • Active engagement on the part of executives (commitment).
  • Establishment of governance bodies with a clear roadmap for what they need to achieve (effective collaboration).
The tendency is to believe that the discouragement is a technology problem. “If we just had the right tools and platforms, everything would straighten out.” As we know, data management issues are primarily people problems. Technology can assist and support, but real improvements must originate in the lines of business, and each individual involved must be sufficiently convinced of the value of their efforts.

Informal Processes - Failure of Lucidity

When data management processes are being performed in various projects and programs but they are not institutionalized, promoted or shared – this is quite common, regardless of the organization’s size or financial condition. This is a failure of knowledge management, which an organization needs to care about.

Typically, organizations are good at hiring smart, skilled staff, and these individuals have throughout their careers learned effective methods and techniques to get the job done. However, the sound practices they apply may not be shared across the organization due to many factors, such as:
  • The organization does not have a general requirement to document processes. This results in a project by project approach (also called ‘ad hoc’), which means knowledge transfer is lacking. If staff retire or leave, it creates varying degrees of risk. There are extreme examples of this situation, when there is only one individual who understands a complex legacy data structure, or a contract is canceled and the interfaces were never documented, necessitating extensive re-work.
  • The organization has no centralized data management organization (DMO). Since a key primary purpose of a DMO is to mine successful practices, formalize them, and make them available to the organization at large – policies, processes, standards, templates, guidelines methods, etc. – no one is ‘minding the store.’
  • The organization has not established data governance. This equates to lack of accountability and responsibility for the data, and therefore confusion and ad hoc processes.
  • Business process modifications – e.g., managing customer accounts from country to country – may be used as an excuse not to document the underlying similarities, including the data created or used in common.
From the standpoint of efficiency and cost avoidance, as well as to achieve the objectives of improving data quality, access, and the underlying data architecture, organizations need appropriate documentation, encouragement, and a compliance process for data management policies and practices. This is true not only for data, but for business processes, privacy considerations, regulatory mandates, and other considerations.

For example, while authoritative sources for shared data may be well known among frequent users, lack of documentation and accountability (i.e., designated data owners) creates confusion and inefficiencies for occasional use by consumers. In addition, data is often aggregated data for management decision-making and predictive analytics. If the sources are not reliable, the conclusions will not be correct.

Formalizing and standardizing data management processes takes effort. It involves organization change, stakeholder education, and management commitment. With data management, as the saying goes, ‘if you do what you’ve always done, you’ll get what you’ve always gotten.’

The antidote for Failure of Lucidity can be generalized as follows:
  • Create data management policies (one document with many segments, or multiple documents).
  • Stand up a centralized Data Management Organization and task it with creating processes and standards that can be applied organization-wide.
  • Task data governance groups to ensure that all lines of business understand and apply the policies, processes, and standards.
Characteristics of well-managed knowledge about data assets include:
  • Documentation
  • Publication and Communication
  • Available to all relevant stakeholders
  • Managed as an organizational standard policy, process,or work product

Failure to Plan = Planning to Fail – Failure of Judiciousness

To navigate a path through the wilderness, you need a map. To counteract the decades-long neglect of the data layer and lack of sound management practices for this critical asset, the organization needs a data management strategy and a reasonable sequence plan for achieving its goals.

It often seems to me that virtually all organizations are caught up in the desire for a quick fix, both with regards to their business plans as well as the corresponding data and technologies to realize their goals. My personal shorthand term for this short-term solution orientation is ‘Quarterly Profits.’

The time, cost, and effort to develop a strong data management program varies for each organization. And since education and an evolution of culture is required to commit to and sustain improvements, it takes patience and determination on the part of the organization to succeed. In the desire for “once and done,” another symptom of the ‘Quarterly Profits’ mentality, some organizations make a big move and throw funds at a problem.

For example, the organization facing pressure to demonstrate the lineage of data underlying its regulatory reporting may suddenly decide to buy an industry-leading metadata repository and engage large teams to populate it completely in a short period of time. This ‘big bang’ approach incurs risks – it is not judicious (AKA, not a smart decision). This organization may encounter many traps, for example:
  • It may have neglected to determine exactly what metadata should be captured to meet the requirements.
  • The legacy environment is likely to be significantly undocumented, or the completeness of documentation may differ greatly by business line and project.
  • There may be no solid understanding of the existing gaps in knowledge about the data.
  • Participants gathering the metadata may not have training or experience.
  • There may be no priorities or sequence set for key subject areas, etc.
If instead, the organization had engaged data governance to explore with the business lines and internal regulatory compliance scoping elements – such as how many data elements are affected, where they reside, what the documentation status is, level of effort estimates, and a criticality determination – it would be able to create a strategy and sequence plan for metadata management, with clear (Lucid) and manageable (judicious) tasks.

The judicious approach to progress in building EDM capabilities is to ensure success in small increments. It is best to establish the finest practices for one project, preferably an implementation effort that is already approved, which has a clearly bounded scope, and which will impact multiple business line suppliers and consumers – for instance, a master data project.

If tasks are added to the pilot project with the intent of reusability – such as creating a business glossary, creating metadata standards, creating quality rules, training Data Stewards, specifying data requirements thoroughly, etc. – the organization can re-apply the knowledge to other initiatives, eventually institutionalizing the practices for enterprise-wide application. In most organizations, this doesn’t happen naturally, because projects are often siloed – data is captured and managed for a specific set of business processes, with different project teams and different stakeholders.

The general antidote for Failure of Judiciousness consists of several elements:
  • One hour of strategic thinking – what are our goals and desired outcomes – can save months of extraneous effort. Or as the saying goes: ‘Put the big rocks in first.’ The organization needs to create a strategy for data management, based on the business strategy. It needs to be written, approved, and broadly communicated.
  • The sequence plan accompanying the strategy should reflect analysis of both the achievements that will be most valuable and also practical (the intersection of the two). It should address both the new data management capabilities and the data scope to which they will be applied. For example, ‘This fiscal year we will implement data quality processes for customer accounts.’ Balancing incremental capability building with business benefits will serve to focus the organization’s efforts.
  • Stakeholder engagement is critical – an organization needs an executive level governance body and a standing data stewards group, at a minimum. Informed minds working together with shared Aspirations (vision and perspective) will prevent unwise data decisions.
  • Data management training is highly recommended, for governance participants, business data experts, and their information technology partners. Knowledge of fundamental data management practices and a common language will equip many individuals with the skills to make Judicious plans and decisions.