Unlock the Power of Data Stewardship for Data Citizens

Executive Summary

This webinar highlights key elements essential for effective data management and community engagement in various sectors, including transit, oil and gas, and finance. Howard Diesel emphasises the significance of data authenticity, stewardship, and business architecture. It outlines the necessity of rigorous Data Governance practices and the implementation of data management artefacts.

Webinar Details

Title: Unlock the Power of Data Stewardship for Data Citizens
Date: 12 June 2025
Presenter: Howard Diesel
Meetup Group: African Data Management Community
Write-up Author: Howard Diesel

Data Stewardship and Business Architecture

Howard Diesel opens the webinar and recaps his recent course delivery, aimed at Data Stewards, which took place on Monday and Tuesday. The course was designed to make the content more practical and accessible for those unfamiliar with complex data management concepts. Howard added a thought-provoking question to participants from the Development Bank of South Africa about their initial responses to receiving data and machine learning models, emphasising the importance of assessing the credibility and reliability of data.

An attendee shared their perspective, highlighting the need for a thorough understanding of data origins and the necessity of accompanying documentation such as data models and business glossaries. Despite the presence of numerous artefacts within enterprise architecture at various organisations, the challenge remains in identifying which ones are credible and useful. Additionally, there is often a lack of polished, well-defined business architecture and capability models to support data strategy alignment.

Figure 1 How-to Steward Data

Data Stewardship in Engineering Rigorousness of the Oil and Gas Industry

The attendee goes on to highlight the significant engineering rigour applied in projects within the oil and gas industry, such as pipeline construction and refinery development, contrasting it with the often insufficient rigour seen in the IT departments of large companies. They acknowledge that while some firms, like Roche and IKEA, may uphold high standards, the overall trend is troubled by the volatility of oil prices, leading to investment in best practices during booms but cutbacks and neglect during downturns.

Howard then highlights the importance of industry knowledge and data provenance, particularly in the oil and gas sector. He emphasises that a lack of experience can raise questions about comprehension of the business’s value chain. During a cyber incident recovery, prioritising applications was critical, with payroll and access systems identified as top priorities, showcasing the need to manage resources effectively.

The reality that basic business operations often rely on tools like Excel and PowerPoint is reflected in this. A business impact analysis of Data Stewardship reveals the significance of understanding the benefits of data management, as well as the importance of linking recovery time objectives to specific applications and data areas.

Figure 2 Learning Objectives

Understanding Data Management Artefacts

Data Stewards need to acknowledge the importance of performing data interrogation before utilising a dataset. Howard adds to this by recalling an experience with central bank data during the COVID-19 pandemic, where credit data was found to be unsuitable for decision-making three months later. This highlights the necessity of assessing data to ensure it is fit for purpose, thereby effectively addressing specific questions.

To facilitate this understanding, a framework titled “Questions About Data” has been developed, drawing inspiration from Bob Seiner’s categories. This framework outlines essential questions to consider and the data management artefacts needed to address them, emphasising the critical skill of interrogating data to enhance decision-making processes.

Figure 3 Learning Objectives Pt.2

Figure 4 DM Artefact Identification & Purpose

Figure 5 Question Categories & Scope

Figure 6 Types of Questions “About” Data

Business Data Definition Development

Business data definition questions focus on the logical characteristics of data, encompassing elements such as taxonomies, glossaries, dictionaries, and Metadata repositories, which help bridge the gap between logical and physical data planning. Answering these questions provides three key benefits: discovering the location of data, enhancing understanding through glossaries, and establishing trust through defined mappings.

Rationalisation questions address discrepancies in duplicate data across systems, identifying authoritative sources and documenting data similarities and differences, such as terminology variations like “business partner” in SAP versus “stakeholder” in CRM. Furthermore, business rule questions pertain to constraints and guidelines that exist within data modelling tools and documents. In contrast, data structure questions describe the physical data stored in databases, maintained by developers and business analysts. Effective Data Governance involves identifying subject matter experts, data owners, and stewards responsible for ensuring Data Quality and usage, as well as maintaining a data catalogue to facilitate reporting and analysis.

Figure 7 Business Data Definition Questions

Figure 8 Rationalisation Questions

Figure 9 Business Rules Questions

Figure 10 Data Structure Questions

Figure 11 Data Governance & Stewardship Questions

Training Data Stewardship

The focus of the training is to equip Data Stewards with the ability to understand and answer key questions regarding the scope and purpose of data within our data estate. Participants will learn to identify the necessary artefacts that provide guidance, such as data catalogues, and understand who is responsible for managing data effectively.

By addressing critical inquiries, such as the data we have and implications of poor knowledge management, Data Stewards will be trained to handle a comprehensive set of questions, potentially totalling 90, derived from 6 categories with 15 questions each. The training aims to provide these stewards with the knowledge and skills to build and contribute to these artefacts effectively.

Figure 12 Category Explanation

Implementing Data Stewardship and Self-Service Reporting

Howard focuses on the key questions related to business data definitions, rationalisation, business rules, data asset structure, and reporting and analysis. Important inquiries include how reports present and filter data, whether they sort or group information, and whether totals are provided. Emphasising the significance of these questions, participants aim to ensure Data Stewards can effectively categorise and answer them, enhancing their empowerment and efficiency in issue resolution. Additionally, Howard references the value of Metadata and engaging clients with a well-defined set of inquiries to inform business architecture.

The transition to self-service Business Intelligence (BI) has posed significant challenges for data citizens, particularly when it comes to joining datasets and ensuring data integrity. In a previous engagement, a company reduced its data team from 20 to just four members, primarily Data Stewards responsible for generating reports for account managers and sales teams. This reduction left a substantial gap in support and understanding for users trying to navigate self-service BI.

Many team members felt overwhelmed and lacked the confidence to ask relevant questions about data relationships, which hindered their ability to generate accurate reports. The change management process was stressful, as users struggled not only to understand the necessary questions but also to identify the correct answers, highlighting a critical gap in data literacy and support.

Figure 13 Business Data Definition

Figure 14 Reporting and Analysis

The Importance of Asking Questions in Data Modelling

Howard stresses the importance of asking critical questions when building data models, particularly in transitioning from an “as is” to a “to be” framework. A significant challenge arises when teams fall into the trap of maintaining the status quo without questioning the potential value of innovation or connecting disparate data sets.

It is crucial to consider the benefits of these efforts and to communicate this value to Data Stewards and stakeholders involved in the process. By addressing the question of “what’s in it for me,” team members can better understand the advantages of the exercise, fostering a mindset open to exploration and improvement.

Data Stewards need to understand the value of data and collaboratively share responsibilities rather than taking on everything alone. Howard thus encourages Data Stewards to actively engage by asking questions, even about unknowns, as this is essential for navigating Data Quality standards and expectations.

Howard emphasises the importance of understanding the quality of the underlying data used in reports and the necessity for stewards to possess the necessary knowledge and Metadata to address inquiries about the data they manage competently. The aim is to empower stewards to proactively challenge data sources and enhance their understanding, enabling them to respond effectively to questions and inquiries.

Figure 15 Question Questions

Data Management Artefacts

Howard then focuses on developing essential data management artefacts, emphasising the need to prioritise key items from the 193 available in the DM box. He raises a pertinent question regarding the alignment of these artefacts with their current initiative to compile a data asset register, specifically seeking guidance on relevant question categories.

Additionally, Howard shares an analysis of various questions and their corresponding data artefacts, highlighting the criticality of each artefact based on the number of questions he addresses, ultimately linking this to his data catalogue. This collaboration, Howard adds, aims to enhance understanding and efficiency in categorising data management efforts.

The importance of categorising artefacts by their audience—executives, Data Stewards, or technical teams—to better navigate their relevance is emphasised. A chatbot has been proposed to facilitate data inquiries, enabling users to access information independently, even if they are hesitant to approach Data Stewards directly.

Identifying responsible roles for each artefact is crucial to ensure that Data Stewards know who to consult for accurate information. The goal is to build a comprehensive system that allows efficient information retrieval while ensuring data artefacts are complete and well-documented.

Figure 16 Business Data Definition

Figure 17 DM Artefact Identification and Purpose

Figure 18 Types of Questions “About” Data

Data Lineage and Value Proposition in Business Architecture

Developing a comprehensive vertical lineage complements the existing horizontal lineage, allowing data to move seamlessly between positions. This vertical lineage traces data from its source up to its value proposition within the business architecture, addressing the critical question of why we possess this data and its contribution to business objectives.

Howard shares that his approach includes creating detailed definitions from business glossaries, dictionaries, and catalogues, taking into account the order in which these artefacts were created. He began my process with an examination of business strategy and data planning, incorporating Metadata collection and application planning. Notably, Howard recognises the need to include stakeholder perspectives in the value proposition, highlighting its importance in delivering relevant insights.

An attendee emphasises the importance of understanding the value chain and the critical metrics, such as KPIs, essential for organisational maturity. The challenges faced by a business process modelling team highlight issues with multiple documentation tools and the need for a cohesive approach to knowledge management.

The turnover of knowledgeable personnel has been a major concern, resulting in setbacks despite training efforts, underscoring the importance of making tools accessible to business Data Stewards for improved understanding and process continuity.

Howard emphasises the importance of training employees to prevent potential losses resulting from underperformance. He emphasises the significance of data management through a structured approach that includes vertical lineage elements, a data catalogue, a data dictionary, and a business glossary. These components enable comprehensive definitions of data products and attributes, allowing for the creation of an effective dashboard.

The process involves compiling an Excel spreadsheet to illustrate navigation through the data hierarchy and developing a meta-model of all worksheets and their associated Metadata. Additionally, the creation of a CRUD matrix helps identify which applications manage and interact with data, culminating in a detailed meta-model that encapsulates all elements and their classifications for a Development Bank.

Figure 19 Business Data Definition

Figure 20 Complete Vertical Lineage

Figure 21 Complete Vertical Lineage Pt.2

Figure 22 Development Bank Capabilities Model

Figure 23 Business Glossary

A Learning Process with Copilot

Howard reflects on their experience with development banks, noting a lack of understanding about their operations despite previous work with the African Development Bank and the Business Development Bank. After struggling to find clear information, they utilised Copilot to ask questions about the business architecture and operational capabilities of development banks.

This led to a rapid learning process, where they gathered insights on core capabilities, received references, and ultimately organised the information into an Excel spreadsheet. The speaker highlights a shift in how individuals are increasingly relying on AI tools like Copilot for data collection, rather than traditional search engines, emphasising the effectiveness of this approach in compiling comprehensive capability models.

The use of Microsoft’s CoPilot proves to be highly effective for users who have a clear understanding of their needs and the necessary Metadata, enabling the generation of relevant Excel outputs. Howard then highlights the significant role of development banks in funding infrastructure projects that commercial financial institutions often deem too risky.

Through collaborative efforts, users can create comprehensive lists addressing donors and funding sources while gaining insights into strategic planning, sector prioritisation, and policy advisory services. The learning curve has been steep, especially regarding loan origination processes and core business concepts within subject area models. Despite feeling somewhat outmatched by the capabilities of AI tools like CoPilot, users recognise the importance of starting points in their understanding of development banks, which facilitates meaningful contributions.

Figure 24 Complete Vertical Lineage Pt.3

Figure 25 Working with CoPilot to Fill in the Gaps

Figure 26 Overview of Conversations with CoPilot

Figure 27 Conversation with CoPilot

Figure 28 Building A Generic Business Architecture

Figure 29 Suggested Capabilities

Figure 30 Suggested Capabilities Pt.2

Figure 31 Business Process Example

Figure 32 Subject Area Model

Business Concepts and Data Models

Howard demonstrates the updated Subject Area Data Model, which he has integrated into Excel, resulting in detailed subject area models that feature business concepts, definitions, and examples. He shared that he was impressed by the comprehensive insights, including suggestions for ownership and value propositions, which supported strategic planning by identifying key stakeholders and measures.

This revelation allows for envisioning the creation of a dashboard to track the Business Value Chain and request wireframes for departments to visualise this structure. However, I encountered challenges when trying to generate a Metadata dictionary, as the descriptions in the business glossary were repetitive and lacked differentiation, which proved frustrating.

After struggling to compile information across various tables, worksheets, and columns, Howard sought assistance in generating VBA code, which successfully created a comprehensive meta-model of my work. This model allowed me to establish connections among capability definitions, core business concepts, and subject areas, ultimately generating relationships and business rules, such as links between repayments and loan agreements.

The standout revelation for Howard was the creation of a CRUD matrix, organised by subject area and capability application, which significantly enhanced my understanding of the data relationships I was working with.

The project involves generating a comprehensive capability list based on specific subject areas, which ultimately aids in building a CRUD (Create, Read, Update, Delete) matrix to establish ownership and manage data effectively. Utilising tools like wireframe generation, the process started with gathering questions and insights over a day, enhancing collaboration and clarity.

While the initial outputs require validation and refinement, they provide a valuable starting point, alleviating the challenges of beginning with a blank page. The ability to interact dynamically with tools like CoPilot enables ongoing inquiries and engagement, allowing for a deeper understanding of key metrics and models within the project framework.

Figure 33 Subject Are Model Pt.2

Figure 34 Value Chain

Figure 35 Dashboard Requirements and Wireframe

Figure 36 Wireframe

Figure 37 Using CoPilot Generated VBA to Finish

Figure 38 Core Business Relationships

Figure 39 CRUD Matrix

Data Governance and Data Lineage

An attendee expressed gratitude for insights on vertical data lineage, emphasising its importance in capturing business Metadata and effectively maintaining data catalogues. Additionally, Howard highlighted essential elements such as descriptions, classifications, access policies, retention policies, and governance workflows, showcasing how the system prompts users to enhance data management as they work.

This proactive approach impressed attendees, who acknowledged the significant value of templates in establishing streamlined Data Governance practices. Overall, the session fostered meaningful discussions about the roles of stewardship and collaboration in improving data management processes.

Figure 40 Application Solution Architecture

Figure 41 Core Business Crisis