Your office of the CDO needs a vision. Success won’t sprout from a rock. Leveraging data as a strategic asset can’t be done without defining a strategic approach to that data. Building a data strategy doesn’t prevent you from being agile in your approach. At the beginning, your vision of the organization’s data strategy might be fuzzy. That’s okay.
As you develop your data vision, provide guidance on how to unify business and IT perspectives, and promote value metrics from a data-driven culture, remember that change is welcome. If you’re headed down a path and not garnering the necessary organizational buy-in, change your course. There’s a better path forward. You just need to discover it—without going through the difficult journey that Milton Hershey did.
4 failures to success
The sweetest city in the world is Hershey, Pennsylvania. However, it didn’t start out sweet for Milton Hershey, an American chocolatier, businessman, and philanthropist. Milton didn’t see much use for school and only had a fourth-grade education. At the age of 14, he started an apprenticeship for a local printer. That was short-lived, and he was later fired for dropping his straw hat into a machine. He quickly was paired in another apprenticeship with a confectioner named Joseph Royer, based in Lancaster, Pennsylvania.
During his four years with Royer, Hershey learned everything he could about the candy business. Then, at the age of nineteen, he moved to Philadelphia to start his confectionery company. Unfortunately, he couldn’t make it in Philadelphia because of heavy supplier debts, so he moved to Denver, New York, Chicago, and eventually New Orleans, but never made a success of his business in any of them.
Throughout his journey from location to location and with each failure, Hershey was learning. He discovered that fresh milk is vital to good candy. In 1886, at the age of twenty-nine, he was penniless. Eight years later, he sold Lancaster Caramel Company for $1 million and turned to chocolate, where he soon founded the Hershey Chocolate Company in what would become Hershey, Pennsylvania.
Hershey became a successful businessman. When he died, he signed over all his company shares of the Hershey Chocolate Company, via a trust, to the Hershey Industrial School, which was an orphanage he founded. The shares were valued at $60 million. For reference, that same year, Coca-Cola sold for $25 million.
It’s fascinating what an individual can do with a vision and passion. Milton was famous for saying, “The caramel business is a fad.” At the time, he sold Lancaster Caramel Company, profits were at all-time highs. Yet, he sold his business and went into chocolate. Understand what makes your business successful.
4 pillars of success
To discover your company’s data-strategy vision, build around the four core aspects of the office of the CDO. These require careful set-up to enable your successful data office initiative:
- Talent and culture
Governance defines the process, establishes forums, and promotes strategic communication. Architecture creates the guardrails for reference architecture, identifies common lexicons, and develops an edge for the data platform. Standards elaborates operating procedures, specifies technical standards, and identifies design precepts. Talent and culture span education and training, skills, roles, and responsibilities of agile teams—which are the human aspect of change.
These four areas are the pillars of a successful office of the CDO.
Adjusting how we look at opportunities—including data—doesn’t happen overnight. We’re transforming an organization. To do this, we must lean on these four core pillars of a successful office of the CDO, which we’ll now discuss in detail.
Data governance: discovering better insights
Governance offers accountability for data, business agility, better compliance, IT agility, and stronger insights. Setting up seamless data governance facilitates stakeholder interactions and makes the decision-making process easy. Without this framework, decisions will spin, and participants will become frustrated that engagement isn’t consistent or uniform across business areas or divisions.
There are six areas of interest when we’re talking about data governance and the role of the office of the CDO:
- Enterprise data governance
- Data-quality management
- Master-data management
- Metadata management
- Data-protection management
- Data strategy and diagnostics
Enterprise data governance is the management of business data. This includes the overall management of the availability, usability, integrity, and security of enterprise data. Data quality management is a set of practices that aims at generating and maintaining high-quality data used for decision-making. The quality measures ensure that, throughout its lifecycle—from acquisition through distribution—the data is fit for use. Master-data management is a method for an enterprise to link critical data to a common point of reference. This discipline converges IT and business partners to ensure the uniformity, accuracy, stewardship, semantic consistency, and accountability of master-data assets. Metadata management is the administration of data that describes other data. Metadata involves any information that can be integrated, accessed, shared, linked, analyzed, and maintained. This organizational agreement describes enterprise information assets. Data-protection management enables data security across the enterprise—including automation, orchestration, and document management—to control the many data-protection activities required to run an enterprise. Data strategy and diagnostics is a guide for optimizing data, removing redundant data, and simplifying the lifecycle management of data.
There are many more areas we could add to this mix including data modeling and design, data integration and interoperability, documents and control, data storage and operations… the list goes on. However, we want to start with the basics.
Together, these six elements build the foundation of a robust, data-governance program.
Data standards: sharing data we trust
Data standards communicates enterprise data-sharing frameworks. The objective is to improve trust. We can measure the trust of the data by measuring the credibility, reliability, intimacy, and self-orientation of data.
Addressing trust gets us to the primary value of establishing a standard for data sharing. Who should participate? How transparent is the data? Who can share data? How do we remove misaligned interests?
Producers and consumers of enterprise data must meet baseline standards. Additionally, trust frameworks must be tailored to producer and consumer needs. Combined, this approach raises the level of trust and decreases the risk associated with data production and consumption.
Standards for data sharing describes the ways data can be shared. It also highlights how data can be restricted or access to the data can be increased by removing restrictions in the following sequence:
- No awareness of the data set
- Awareness of the data set
- Awareness of data scope and data dictionary
- Query highly aggregated, obfuscated, or perturbed data
- Query lightly aggregated, obfuscated, or perturbed data
- Access aggregated, obfuscated, or perturbed data
- Access to data
- Ability to share data
No awareness of the data set means the existence of the data set isn’t known. Awareness of the data set makes knowledge of the data set known. Awareness of data scope and data dictionary increases knowledge to include the scope and parameters of a data set—for example, knowledge or access to the data dictionary. Query highly aggregated, obfuscated, or perturbed data enables queries on data sets, but these are highly restricted—in this case, access to a division’s or a department’s data might be removed. Query lightly aggregated, obfuscated, or perturbed data slightly widens the endpoint of data access. For example, in this case, previous access may have been to a state-based population, and this opens access to multi-state searching. Access aggregated, obfuscated, or perturbed data enables the ability to run defined logical operations and pull de-identified data. This level of access provides access to an aggregated data set; however, access to the raw data set isn’t permitted. Access to data unlocks the technical restrictions of operations that may be performed with the data, although specific access rights are usually restricted to certain individuals. The ability to share data may allow sharing to one consumer but may restrict that consumer from sharing the data with another consumer.
Data trust is validated by the enterprise standards in place to share data.
Data architecture: integrating data investments with business strategy
Data architecture guides the integration of enterprise data assets. Architecture is focused on the abstraction of the system, not the system itself. As complexity within your enterprise increases, the need and value provided by data architecture become ever more valuable.
Data architecture has a broad reach and includes many components to build a single version of the corporate “truth.” These ten pieces of enterprise data architecture will ensure the following enterprise data assets are integrated:
- Business entities
- Business relationships
- Data attributes
- Business definitions
- Conceptual and logical views
- Business glossary
- Entity lifecycle and states
- Reference-data values
- Data-quality rules
Business entities refer to the various components of the business. Business relationships identify how those entities interact and ultimately share data. Data attributes tag and identify data elements using known classifications. Business definitions clarify the intent behind the data sets. Taxonomies establish schemes and classification system for groups or attributes with similarities. Conceptual and logical views identify a high-level relationship between entities and how the data is physically represented in the database. The business glossary defines terms across domains and serves as the authoritative source for the data dictionary. The entity lifecycle and states specify where the entity is in its lifecycle from acquisition to destruction. Reference-data values are standards that can be used by other data fields. Data-quality rules are the requirements that the business sets on its data.
It’s also useful to conduct an information value-chain analysis. This process identifies matrix relationships among data, processes, organizations, roles, locations, objectives, applications, projects, and data platforms.
Talent and culture: moving people not data
When we think of data, we often think of the bits and bytes. We should be thinking about people. Talent and culture are the most difficult aspect to get right when developing a successful office of the CDO.
There are seven main areas in driving culture and getting the right folks in the right roles:
- Talent acquisition
- Performance management
- Competency management
- Learning and development
- Leadership development
- Career management
- Succession management
Talent acquisition is the process of finding and onboarding skilled data talent. Performance management ensures that individual activities and outputs meet organizational goals. Competency management is the process of developing the skill sets of individuals. Learning and development attempt to enhance individual performance by tuning and honing skills and knowledge. Leadership development helps to expand an individual’s capability to grow within the organization. Career management is the deliberate planning and coaching of an individual’s activities, engagements, and jobs over a lifetime. Succession management is the systematic process of developing, identifying, and grooming high-potential individuals for more aspirational roles.
By developing a talent management plan, we improve our odds of attracting, developing, motivating, and retaining high-performing employees.
Designing a world-class office of the CDO
Designing a world-class office of the CDO begins with a vision and a data strategy. We establish a strong organizational base for scalability and growth by using the pillars of governance, standards, architecture, and talent and culture.
Take time to understand the intrinsic value of data. Is your data accurate and complete? Determine the cost value of data. If you lost your data, what would it cost to replace it fully? Evaluate the business value of the data. How fit-for-purpose is this data to make data-driven decisions? Measure the performance value of your data. What parts of the data fuel key business drivers?
Designing the office of the CDO is an exciting process. Let’s hope your venture into the data business is less bumpy than Milton Hershey’s entrance into the chocolate business.