Chief Data Office Corner: Data Strategy

The outcome of a Data Strategy is 'consumed' by a vast majority, if not all stakeholders of a company: think of business internal (from product development, customer service and operations to management information dashboards), business partners, technology departments, regulator. The data strategy must include inputs and balance out the demands of each user or creator/owner of data.

Data Strategy: Business Direction, Data Principles, Enablers and Capabilities

A data strategy would ultimately be uniquely crafted for each company, the industry peculiarities it is in and markets it targets, the local, regional or global reach or ambitions, financial strength and so forth. However, commonalities exist, company, market or industry agnostic.

In May 2020, the United Nations produced a paper, Data Strategy of the Secretary-General
for Action by Everyone, Everywhere with Insight, Impact and Integrity which offers a comprehensive framework to establish a data strategy.

This framework proposes a set of data principles and advocates a use case driven approach. This would be based on strategic priorities of a company, identifying the and nurturing capabilities required to meet those priorities, or goals, and then fostering enablers that can make it all happen.

The Australian government published their 2018-20 Data Strategy paper in which they portray a data vision based on a set of data principles that contribute to the wellbeing of Australians. It is based on a set of data capabilities, existing and planned, and on identifying and developing enablers for the capabilities to grow and solidify. It is interesting to see that the 2021-24 Data Strategy builds logically onto the earlier stages, amplifying focus on non-tech dimensions: people, skills, culture, analytics capabilities.

We can see similar approaches of plotting a data strategy and the execution thereof, not driven by tech, but with tech as an (important) enabler at company level as well, for example Deutsche Telekom but also solution providers like PwC are explicit about the moving parts in establishing and executing a data strategy: tech remains an important party but does not drive.

The reason for repeatedly mentioning that data strategy is not an IT strategy, and should not be driven with a tech mindset, is because in companies this is exactly what often happens. A CDO mindset, area and scope of work and goals, are not those of a CTO or CIO. In this entry the CDO vs. CIO and CTO is further explored.

A way to translate the data dimension of a business direction is through a visual that combines the strategy, enablers and capabilities: the Data Strategy House.

The Data Strategy House is a powerful visualization of the context in which a data strategy plays out. In the example here the fictive company is strategizing a way to become data driven in terms of visualization and analysis. The tactics would be the required capabilities to fulfill the strategic objective and the enablers, on the bottom are both organizational (tech, people etc.) and data specific (core and enriched data).

After setting the strategic direction, communicated with or without a Data Strategy House, establishing a set of Data Principles is the next priority.

Data Principles harmonize opinions, personal preferences or other subjective notions on how to treat data. It is the glue that holds together business accountabilities of data ownership and data use as well as the technical constraints and requirements to ensure data in storage and in transition remains of highest quality possible.

The list here is a typical set of Data Principles.

1 / Data is an Asset

In a financial setting, like with your bank account, you would want to know at any given time what your balance is, your stock portfolio, gold holding, credit cards spent - and you want this to be all correct, protected and secure. Data must be treated with similar rigor as other recognized assets in terms of following data quality dimensions:

Safety, Security and Privacy (stored and in transition)
Reliability, Accuracy, Timeliness and Unambiguity
Compliancy (regulatory and other)

2 / Data Requires Governance

Data, that is not archived, has no steady-state. From definition of individual data fields, to storing, extracting, combining, encrypting, ownership changes, producing new data from models etc., data use conditions and restrictions need to be managed. The description of data, meta-data, and clarification of data roles and responsibilities are crucial to an effective ecosystem. Required are:

Data Catalogue for meta-data management
A (Target) Operating Model (TOM), RASCI and processes

3 / Data is Company Owned

Data Owners are accountable for the data in their domain but data itself is not owned by any individual user; it is a company asset and therefore needs to be: