A Step Ahead: Data Strategy for Complex Systems — Landscape Analysis and Visioning

An enterprise data strategy can be an effective means to articulate an organization’s long-term vision of exploiting enterprise-wide data to facilitate accurate and comprehensive data analysis. The data strategy can:

• Identify the capabilities needed to maximize the value of enterprise data
• Describe current gaps in capabilities
• Define a path forward to address the gaps and embrace new approaches

A powerful, three-step methodology for developing a data strategy combines landscape analyses for all relevant areas of the enterprise (Step 1), a Strengths, Weaknesses, Opportunities, and Threats (SWOT) analysis for each of the areas (Step 2), and subsequent visioning sessions to review the landscape and SWOT analyses and develop a collective understanding of commonalities and differences across the enterprise (Step 3). These activities lead to a shared vision, which becomes the foundation for the data strategy recommendations. An effective data strategy can drive the evolution and maturation of a data ecosystem as the core intelligence layer of a complex enterprise.

Step 1: Landscape Analysis is a systematic way to explore an ecosystem’s challenges and technical resources by identifying the current state of data practices and planned or potential practices for each component of the enterprise. The analysis (1) provides essential context about a component’s position and capabilities relative to mission needs, and (2) helps an enterprise understand its current technical posture and how to scale to a future state to meet rapidly evolving data engineering requirements. The analysis is usually conducted through a series of environmental scans and interviews with leaders, users/consumers, implementers/producers, and mission partners, and involves asking the following types of questions:

1. Define the scope of the analysis: In this area of the ecosystem, what is the universe of current, planned, or potential strategic activities, programs, initiatives, and mission data that relate to the data strategy?
2. Conduct an environmental scan: When looking at developments in multiple operational spheres (private sector, governmental sphere), what potential initiatives should the organization consider? How have similar enterprises with the same challenges addressed them? It should be noted that using Large Language Models (e.g., Llama) to assist with the scan has the potential to provide a rapid domain overview, which can be enhanced with iterative prompt engineering and question refinement for optimal results. [1]
3. Identify the mission-critical challenges and gaps: What are the mission-critical challenges that the enterprise faces in executing its programs well, especially as they relate to data, technology, business processes, governance structures, and human capital? What challenges or gaps does the enterprise foresee for executing planned or potential initiatives?
4. Identify relevant data sets: What federal, state, local, and private sector data sets are instrumental for informing the enterprise’s decision-making? What are the opportunities for and challenges with working with these data sets?
5. Assess evolving data management processes: What are the opportunities for and blockers to evolving data approaches for future initiatives?

A Framework for Organizing the Landscape Analysis Findings

A helpful approach to organizing the Landscape Analysis findings in a consistent way is to categorize them and then conduct the SWOT analysis on each category. The suggested grouping involves aspects of the data lifecycle as well as a human capital component that is necessary for managing data in a complex enterprise. Potential questions to ask leaders, users, implementers, and stakeholders are listed for each category.

• Human Capital: Does the enterprise have enough of the right staff with training or experience in the needed disciplines to execute the planned initiatives? What additional skills and resources would help? Is there a need for a workforce program to train and help build a data-literate workforce with the skills and the knowledge to address consequent data challenges? How can the burden on individuals to manage data properly be eased? How can staff be incentivized to ensure compliance with data management best practices? Are there funding challenges to address (for staffing, training, technical resources)?
• Data Access: Do the stakeholders/users have easy, low-friction access to the system and needed data services (e.g., tools, data in interoperable formats, data catalogs) in both the enterprise and in partner organizations? How can the organization eliminate roadblocks that inhibit accessing/sharing data or collaborating within the enterprise and with other enterprises (e.g., regulatory constraints that prohibit being able to easily transfer data, history of data silos, and organizational autonomy)? Is data discoverability (e.g., query capabilities, data inventories) an issue? Does duplication of infrastructure across the organization waste resources? Are messaging activities sufficient to publicize to potential users what data and data services are available?
• Data Collection and Analysis: Are there barriers to collaboration (e.g., policy issues, technical issues such as incompatible formats or inconsistent field names, staff time constraints) for sharing or linking data between organizations? Are there business processes in place to support the needed activities? Are there user and stakeholder engagement plans for determining data usage needs (e.g., specific data formats) for specific groups? If a huge amount of data is being collected, is there a threshold at which point managing the data becomes impossible? If artificial intelligence (AI) is used to go through that data, can the data and resulting analysis be trusted?
• Data Quality: Are there sufficient processes in place to assure the required level of data quality, including data timeliness? What are the data management activities the enterprise needs to add to its current processes to ensure data quality to make data AI-ready? Is there duplication of data across the enterprise? Are the processes to ensure data quality manual, labor-intensive, and error prone?  Does the organization work with industry and academia to assess the data quality of its data products and their fitness-for-purpose to continuously improve them? Is the data comprehensive enough for valid analysis? Where is the enterprise on the Data Management Maturity Model [2]?
• Data Governance: Are there problematic dependencies on other enterprises’ technologies and data libraries (e.g., because of inconsistent data standards for formats and tagging)? Are standard classification codes for the relevant domains up-to-date? Is there a baseline data inventory (e.g., a data catalog) across the enterprise? Is there a data governance structure in place to ensure there are authorities, organizational structures, and budgets in place to (1) support the management and use of data assets and (2) to guide the development of enterprise-wide efforts over time to meet emergent needs?

Step 2: Strengths, Weaknesses, Opportunities, and Threats (SWOT) Analysis

A SWOT analysis of the ecosystem’s position and capabilities (determined by the Landscape Analysis outlined above) relative to mission needs for each category can help the enterprise to identify (1) the strategic objectives to be achieved, (2) the challenges to be overcome, (3) the areas where current and proposed initiatives can be leveraged, and (4) the key areas to be addressed through the development and implementation of a data strategy, including clearly defined ownership, governance, metrics, and funding.

Step 3: Visioning sessions with subject matter experts and leaders then draw on the landscape and SWOT analyses to develop a cross-enterprise understanding of commonalties and differences across the enterprise to shape a shared vision for the data ecosystem that helps highlight the art of the possible. In these sessions, the participants identify common themes, connections, and new ideas and build out the shared vision by (1) consolidating objectives into communal thematic outcomes, (2) proposing initiatives to achieve those outcomes, and then (3) prioritizing the initiatives by considering their impact and feasibility.

Then the appropriate personnel develop data strategy, which presents a path forward for the organization by articulating defined outcomes, needed enterprise capabilities, specific strategies, and indicators of progress over time. The strategy encompasses four dimensions:

• People: What kinds of data talent and skills are needed to actualize the vision?
• Data: How do we prepare our data to meet the broad range of user needs?
• Process: What type of data governance enables us to work together in service of our shared business goals? What are the processes that we cannot do today but will be able to do in the future?
• Technology: What technology do we need to enable what we cannot do now?

Conclusion

The enterprise data strategy begins with defining the decisions and mission outcomes that determine the capabilities required to support them. Landscape Analyses, SWOT analyses of the findings, and visioning sessions together form a three-step process that provides a systematic approach to developing a data strategy an enterprise can use effectively over time to evolve, mature, and enhance the utility of the data assets that exist across mission ecosystems.

References

1. “Enhancing Landscape Analysis with Llama 2 and mChat: Lessons Learned from the Abacus AI in Oncology Project,” D. Donato, T. Adamusiak. MITRE Corporation Internal Briefing, May 2024.

2. Department of Labor Data Strategy (n.d.). Data Management Maturity Model dol.gov/agencies/odg/data-management-maturity-model

About the Author

Dr. Marcia Kerchner is a Principal Information Systems Engineer at the MITRE Corporation with extensive experience supporting government CDOs develop and implement data strategies. She has expertise in data management and reporting analytics, operational requirements analysis and development, and information retrieval and search technologies in many domains.  She earned a PhD in computer science from Cornell University.

Approved for Public Release; Distribution Unlimited. Public Release Case Number 26-1952

The author’s affiliation with The MITRE Corporation is provided for identification purposes only, and is not intended to convey or imply MITRE’s concurrence with, or support for, the positions, opinions, or viewpoints expressed by the author.

©2026 The MITRE Corporation. ALL RIGHTS RESERVED.