Executive Summary
This webinar outlines key themes and trends in contemporary data management practices, highlighting the importance of effective governance and integration of artificial intelligence within the aviation sector. Merrill Albert discusses the evolution of chatbots and their user experience, emphasising the need for comprehensive guides and workshops to enhance data leadership and business continuity. Additionally, the challenges of data lineage, metadata, and the significance of referable data metrics are examined, alongside a critical analysis of data operations and architecture. The webinar highlights the importance of having well-defined policies, processes, and standards to navigate the complexities of Data Governance and ensure organisational effectiveness.
Webinar Details
| Title | Data in a Dangerous Time – The Modern Approach to Managing Data with Merrill Albert |
| Date | 2025-08-26 |
| Presenter | Merrill Albert |
| Meetup Group | DAMA SA User Group Meeting |
| Write-up Author | Howard Diesel |
Contents
Modern Approach to Data Management
The Journey of Writing ‘Data Crimes Against Data’
A Legal Case Study of AI and Chat Bots in Aviation
The Evolution and User Experience of Chat Bots
The Mad Method: A Comprehensive Guide to Data Management
The Importance of Workshops in Data Leadership and Business Continuity
Data Governance and Its Challenges in Organisations
Data Management and Its Importance
Integration of AI and Data Governance
Data Management Metrics and the Role of Referable Data
Role of Lineage and Metadata in Data Modelling
The Role and Challenges of Comments Fields in Data Modelling
Data Operations, Storage, and Metadata
A discussion on Data Architecture, Data Governance, and Standardisation
Data Management: Policies, Processes, and Standards
Information Architecture and Data Governance
Modern Approach to Data Management
Howard Diesel opened the webinar and introduced Merrill Albert. He then recalled the high engagement of her previous webinar, where she discussed data quality issues highlighted in her book. Drawing from her extensive experience since studying at the University of Waterloo, Merrill noted the significant evolution of Data Governance over the years. She pointed out that many organisations still rely on outdated methods from 15-20 years ago, contrasting these with more contemporary strategies. While advocating for a modern approach, Albert acknowledged that adapting to these changes is not mandatory, leaving the decision to organisations. Ultimately, her insights serve as a call to action for professionals to reconsider their data management methods in light of current practices.
Figure 1 Data in a Dangerous Time
The Journey of Writing ‘Data Crimes Against Data’
With a devoted career in data management and analysis, Merrill shared that her writing inspiration began with the process of clarifying her thoughts and sharing valuable insights. This journey led to the publication of two impactful books, starting with “Crimes Against Data,” which arose from discussions surrounding the misuse and abuse of data, revealing numerous alarming cases. Following a brief hiatus after the first book, Merrill returned to her initial goal of documenting important lessons learned in her field, diligently setting a deadline and preparing her work for publication. Ultimately, her writing journey not only reflects her expertise but also serves to educate others about the critical importance of data ethics.
Merrill’s journey as an author has led her to produce three impactful books, each addressing different aspects of data management and governance. Initially, she met her writing deadline with “Lessons Learned,” which focused on insights derived from her experiences. However, during the editing process, she realised that her work had evolved, leading to the publication of “Stop the Data Madness,” which encapsulated her original concept. After a brief hiatus, she delved deeper into solutions for Data Governance issues, culminating in the release of her latest book, “Data in a Dangerous Time,” published by Technics Publications at the beginning of the month. Through her work, Merrill underscores the vital role of effective data management in the context of the growing conversations around artificial intelligence, highlighting its significance in today’s rapidly changing landscape.
Figure 2 How it started
Figure 3 Why Data’s now more Important than ever
A Legal Case Study of AI and Chat Bots in Aviation
Merrill recounted a poignant story that underscores the challenges individuals face when seeking fare assistance during emotional crises. A man urgently needed to book a flight due to a family bereavement and sought a bereavement fare on the airline’s website, but he found the information difficult to navigate. To expedite the process, he turned to the airline’s chatbot feature, which misleadingly advised him to book the flight first and promised a partial reimbursement afterwards. After completing his journey, he was met with disappointment when the airline denied his claim for reimbursement. This experience illustrates the importance of clear communication and support from airlines during such critical times.
The man attempted to claim a refund from an airline after following the rules presented by a chatbot, which indicated he could qualify for a bereavement fare at a reduced rate. However, the airline denied his request, arguing that the chatbot operated as a separate legal entity responsible for its own actions. The man took his case to court, and the court ultimately sided with him, rejecting the airline’s defence. This case highlights a growing trend among companies to integrate AI while overlooking established business rules and existing solutions, suggesting that an effective transition to AI requires leveraging pre-existing strategies and data rather than creating entirely new departments for decision-making.
Integrating existing data processes within businesses is crucial for effective operations and risk management. A compelling example illustrates this point: a lawyer successfully leveraged a conversation with a chatbot as evidence in a case against an airline, demonstrating the relevance of records management in capturing and maintaining these interactions. Such moments underscore the need for treating chatbot conversations like telephone call records, ensuring they are easily accessible for future verification. As AI continues to evolve across various sectors, we can expect more cases that highlight the importance of comprehensive data integration and management practices.
The Evolution and User Experience of Chat Bots
The recent expansion of the data models within the party master data domain to include a third level—categorising entities as individuals, organisations, and bots—marks a significant evolution in how we classify entities. This enhancement introduces bots into the framework for the first time, reflecting the growing complexity of digital interactions. However, this shift raises concerns about the definition of a “party,” as it now encompasses a diverse array of entities, both human and non-human, complicating the legal implications associated with each classification. Thus, while this development demonstrates innovation, it also necessitates careful consideration of how we define and manage these varied entities.
An attendee shared their belief that many companies harbour a misconception about chatbots, failing to recognise them as separate legal entities, which impacts their effective use. This misunderstanding is compounded by the limitations of earlier chatbot iterations, which often did not save conversations, prompting users to seek their own methods for preserving essential interactions. Additionally, the attendee expressed dissatisfaction with current chatbot performance, likening them to verbal FAQs and urging companies to showcase their capabilities; they argue that the distinction between human interaction and chatbot responses is clear. Consequently, they opt to avoid automated voice prompts and unnecessary automated interactions altogether, advocating for more authentic forms of communication.
Many customers express frustration with automated phone systems that often require lengthy navigation through prompts before reaching a human representative. This process can take as long as 15 minutes, with issues arising from circuitous prompts that fail to accurately interpret voice responses, resulting in repeated cycles. Numerous complaints highlight that these systems are not user-friendly, leading to misunderstandings where customers are made to feel at fault for selecting options based on the available choices. Overall, there is a strong sentiment that companies are not adequately addressing these concerns, leaving customers feeling unheard and dissatisfied.
The Mad Method: A Comprehensive Guide to Data Management
The data management wheel presented in the book outlines a structured framework comprised of six distinct capabilities, each clearly defined to avoid overlap. This model streamlines data management practices and draws on a variety of approaches used by previous companies with which Merrill has collaborated. If organisations do not already have their own data management frameworks, this wheel serves as a practical starting point. The design is complemented by a logo, representing Merrill’s brand, Merrill Albert Data, which coincidentally aligns with her middle initial “D,” adding a personal touch to the framework.
The book “Data in a Dangerous Time” presents a compelling framework for effective Data Governance by emphasising clarity and intentionality within data management practices. The author introduces the “mad method,” which advocates for focused activities and the elimination of unnecessary elements, emphasising the importance of clear definitions to address common misunderstandings, particularly those related to data quality. Lastly, Merrill underscores that establishing clear boundaries and maintaining focused efforts are essential for fostering effective communication and collaboration within organisations, paving the way for successful data management.
Balancing theoretical knowledge and practical application in data management presents a significant challenge, especially among experts in fields such as metadata. While these professionals have a deep theoretical understanding, it is essential to translate this knowledge into clear, actionable insights that engage team members who may not be interested in the complexities of the theories.
Effective communication about the importance of data management should focus on articulating specific benefits, demonstrating how these practices contribute to the overall success of the company, rather than merely highlighting their presence on audit reports. Ultimately, fostering this understanding is key to ensuring that all team members recognise the value of data management in driving business results.
The key to effective data management is establishing a solid foundation in Data Governance before tackling additional capabilities. Prioritising Data Governance involves assembling the right stakeholders and typically can be accomplished within eight weeks. Once governance is in place, it lays the groundwork for managing metadata and data quality without the necessity of extensive, costly projects. Instead of initiating large-scale data quality or metadata projects that may not attract enthusiastic participation, focus on critical business needs—such as creating a report with previously unexamined data elements—by concentrating on specific areas where data quality management can be effectively implemented.
The approach to building and managing metadata emphasises the importance of understanding definitions and concepts before assessing quality. Instead of a linear alphabetical progression, the strategy focuses on gradually developing resources based on immediate needs. This method allows for a more flexible and organised process, enabling the team to prioritise tasks and take the time to ensure thoroughness, while ultimately covering all necessary aspects.
Figure 4 Data Management Wheel
Figure 5 Philosophy
Figure 6 Approach
The Importance of Workshops in Data Leadership and Business Continuity
A workshop-based approach is essential for accurately defining key business concepts such as “customer” and “product.” By engaging business leaders in collaborative sessions, skilled data leaders can gather diverse feedback and insights from multiple stakeholders. This inclusive process not only promotes a thorough understanding of terminology across the organisation but also empowers clients to sustain these practices independently long after consulting engagements have concluded. Ultimately, this method fosters collaboration and ensures that definitions align with the broader business context.
Merrill underscores the significance of sustainability in their role, emphasising a commitment to knowledge-sharing rather than secrecy. By conducting workshops and providing instruction, they ensure that others are well-equipped to understand and continue the processes after their departure. On their final day, Merrill expresses confidence in the team’s ability to take ownership of the work, wishing them success as they step into their new responsibilities. This approach fosters an open dialogue, as evidenced by the questions raised by participants, highlighting the importance of collaboration during the transition.
Data Governance and Its Challenges in Organisations
Data Governance is a crucial element for the success of organisations and should never be treated as an afterthought. The discussion emphasises the need to implement governance initiatives in key business areas, taking inspiration from Bob Seiner’s approach to non-invasive implementation. A significant hurdle to effective governance is the presence of siloed technology managers who operate independently, often without alignment to the overall company strategy. This challenge is further complicated by the lack of engagement from Chief Experience Officers (CXOs), who frequently delegate governance responsibilities without fully grasping their implications. To combat these issues, the speaker introduces a comprehensive framework for establishing a Data Governance organisation, highlighting the importance of coordination and proactive involvement at all levels. Ultimately, prioritising Data Governance will strengthen organisational alignment and operational effectiveness.
The effective implementation of Data Governance requires a business-driven approach rather than an IT-centric one, as evidenced by experiences from its early development 15-20 years ago. Initially, organisations mistakenly approached Data Governance with a focus on IT, which proved ineffective. To address this, a shift towards leadership by business professionals within governance structures is essential. While some resisted this change, recognising the need for collaboration led to the inclusion of business roles on the Data Governance organisation chart, facilitating better alignment and cooperation between business operations and data management.
In addressing the collaboration between business and IT, it is important to recognise counterparts on the organisational chart, as this has fostered better integration despite evolving roles within IT. Initially, including IT in the org chart seemed unnecessary, but it became clear that their presence was vital to team dynamics. Tapping into areas with significant energy, such as the risk schedule, has proven beneficial. Interactions with the Chief Risk Officer highlighted the critical reliance on data in their operations. It’s essential to focus efforts where enthusiasm exists, while avoiding areas lacking momentum to ensure productive outcomes.
The approach involved integrating specific columns into the risk schedule to emphasise data dependencies across both enterprise and project risks, ensuring that each risk had a data-related aspect. This strategy leveraged existing energy around data management, allowing for a gradual implementation of governance while simultaneously educating the organisation and project teams. A significant focus was placed on the financial and risk management elements necessary for effective governance. When issues arose, it provided a clear reference point to highlight failures in addressing data quality before initiating work, demonstrating the importance of thorough data practices. The speaker reflects on their unique background in logical data modelling, noting that few share a similar path from IT to a focus on data.
Effective data management begins with comprehensive data modelling to prevent long-term quality issues. Many individuals encounter data-related challenges only after dealing with the consequences of poor quality, underscoring the need for thorough definitions and lineage documentation before data enters production. Initially, organisations employed sound practices that prioritised data quality, avoiding extensive metadata issues. However, as time progressed, a shift occurred, leading to the release of inadequately managed data into production, which has contributed to current data management challenges. This situation underscores the crucial importance of adhering to effective data management practices from the outset to prevent unnecessary complications in the future.
Data Management and Its Importance
Merrill presents a comprehensive framework for effective data management that addresses crucial capabilities and challenges. She begins by identifying potential roadblocks and introducing a conceptual wheel that outlines six key capabilities, emphasising the necessity of a strategic approach. The largest chapter focuses on Data Governance, underscoring its importance and the author’s dedicated attention to this area. Each capability is thoroughly examined, supplemented by a discussion on relevant metrics. In conclusion, Merrill offers practical recommendations for implementing a successful data management strategy, including a sample roadmap and task list to guide project execution.
The concept of a data management wheel, which includes six core capabilities, underscores the necessity of a comprehensive approach to data management. While data strategy is crucial, it exists independently from Data Governance, suggesting that governance represents only one component of a broader data management framework. This emphasises the integration of various elements, such as governance and metadata, to maximise the value derived from data management. Ultimately, recognising the interconnectedness of these components enables organisations to adopt a holistic data management strategy that enhances overall effectiveness.
Figure 7 Table of Contents
Integration of AI and Data Governance
The integration of AI into existing Data Governance frameworks is essential for organisations aiming to adapt to emerging technologies. In a recent discussion, Merrill highlighted that her evolving Data Governance program incorporates AI, emphasising the need to view AI and Data Governance as interconnected rather than separate entities. She noted that although her latest book does not specifically address AI Governance, the principles of governance are universally applicable across various domains, including data management and analytics. Ultimately, by understanding these governance principles, organisations can strategically navigate the challenges presented by AI and other burgeoning technologies.
Integrating AI Governance into existing Data Governance programs is essential for organisations seeking effective solutions. Successful collaboration between business and IT is crucial, as attempting to manage AI Governance separately can complicate processes and hinder progress. By leveraging established principles from data and information governance, organisations can avoid unnecessary complexities, foster collaboration, and create a more streamlined approach to governance. Ultimately, a cohesive strategy that embraces existing frameworks will enhance overall effectiveness in managing both data and AI initiatives.
Data Management Metrics and the Role of Referable Data
Merrill emphasises the necessity of a comprehensive approach to metrics in data management, advocating for a breakdown of metrics by six key capabilities rather than a singular emphasis on data quality. While recognising that many organisations predominantly focus on data quality metrics, she insists on the importance of examining a wider array of metrics that encompass Data Governance, data architecture, and metadata. By encouraging this broader perspective, Merrill aims to address the limitations of the common narrow focus on data quality, ultimately promoting more effective data management practices.
Merrill emphasises the complexities involved in developing a data strategy and metrics, pointing out that participants often lack specific questions or capabilities in this domain. She asserts the importance of gaining diverse perspectives from consultants, as predicting the future of data over a decade remains uncertain. A notable suggestion involves seeking insights from AI entities, such as Sophia the robot, to broaden our understanding beyond human viewpoints. Additionally, there is intrigue regarding the omission of reference master data and data warehouse capabilities from the proposed framework; a participant elaborates that the term “reference data” is often problematic due to its inconsistent definitions among stakeholders, which has led to its exclusion from their discussions. Ultimately, these insights highlight the multifaceted challenges of navigating data strategy in an evolving landscape.
The usage of the terms “reference data” and “master data management” has evolved over time, often leading to confusion, as they can refer to different concepts. Initially, reference data was discussed separately, but it was later merged into the concept of master data management. Some users mistakenly equate reference data with code values, such as product type codes. Due to varying interpretations, explaining these terms repeatedly can be cumbersome, prompting some to abandon using them altogether. However, it is important to recognise that reference data, when viewed as code values, is part of metadata, while master data is a subset of broader data management practices, ultimately contributing to the establishment of a single source of truth through effective data architecture.
The complexity and evolving nature of terminology in data management, particularly regarding reference data and Master Data Management (MDM), necessitate a clearer understanding among professionals. Merrill points out the challenges posed by overlapping meanings and expresses a desire to simplify the language used in this field. Reference data, often categorised under MDM, provides standardised and stable information across various domains such as products, customers, and vendors, facilitating efficient data referencing without the need to duplicate entire records. Yet, differing interpretations of these terms persist, underscoring the need for well-defined terminology in data management practices.
Role of Lineage and Metadata in Data Modelling
The relationship between data architecture and metadata is a crucial aspect of data management, particularly in terms of concepts such as lineage and glossaries. Lineage is categorised within the framework of data architecture, while glossaries fall under the domain of metadata. The speaker expresses a preference for maintaining a distinction between these concepts, although they recognise the technical complexities associated with metadata. They underscore the vital role of metadata in logical data modelling, asserting that models lack completeness without clear definitions and understanding of lineage. Ultimately, despite earlier debates on the necessity of differentiating “metadata,” the decision to include the term in the speaker’s second book reflects its significant relevance and common usage in the field.
The Role and Challenges of Comments Fields in Data Modelling
In data modelling, the use of comments and definition fields varies by the tool employed. Some modern data modelling software features distinct fields for definitions and comments, leading many data modellers to utilise the definition field exclusively, as comments typically do not appear in reports. However, traditional tools like Oracle’s SQL Developer, which is an older and currently free option, include a built-in definition field from their inception, emphasising the importance of field definitions in effective data modelling.
The effective use of comment fields in databases is crucial for clarity and consistency in data management. Despite over 30 years of experience, the speaker notes that these fields are often underutilised, making it challenging to generate accurate reports from data modelling tools that rely on well-defined entries. This issue is further complicated by the need to manage different data models across various departments, highlighting the potential benefits of a centralised glossary to standardise definitions. However, maintaining such a glossary requires careful coordination to ensure all definitions are accurately reflected in each data model. Ultimately, while perfect solutions may be elusive, establishing organisation and discipline in the use of comment fields can significantly enhance the overall effectiveness of database management.
Data Operations, Storage, and Metadata
Concerns were raised by an attendee regarding the presentation’s emphasis on data management and the clarity of the information presented. They highlighted that while the book addresses data control, it lacks a clear discussion on metadata, which is vital for providing context to other data types. Additionally, the attendee questioned whether the presentation adequately covered the actual data itself, suggesting that it may have only been implied in the discussion about data operations and storage. This feedback underscores the importance of clearly distinguishing between data and metadata to improve overall understanding.
The significance of metadata lies in its role as descriptive information that enhances our understanding of data within a model or lineage framework. While operational data is systematically managed and governed, it often lacks explicit descriptions of the objects it encompasses. To effectively address data quality issues, it is crucial to establish a clear resolution process. Furthermore, the discussion raises an important point about the absence of considerations related to data storage options, such as SQL or NoSQL databases, in the operational framework being analysed. In conclusion, recognising the interplay between metadata, Data Governance, and storage solutions is vital for a comprehensive approach to data management.
A discussion on Data Architecture, Data Governance, and Standardisation
Merrill highlights the distinction between data architecture and the broader themes explored in her latest book, which does not serve as a comprehensive guide to data modelling. While the book addresses critical issues such as data retention, destruction, and privacy, it prompts readers to consider fundamental questions about the duration of data storage and the rights to access it. This volume is part of a trilogy, with a particular emphasis on Data Governance and management, rather than the technical aspects of data. It suggests that a thorough understanding of the concepts requires engaging with all three books.
Effective Data Governance is essential for managing diverse data types while maintaining a cohesive framework. While specific data categories, such as customer and product information, may require tailored strategies, the fundamental principles governing data management, security, and architecture should remain consistent across all data types. By establishing a Data Governance organisation that emphasises a unified approach, organisations can ensure comprehensive oversight without the need for extensive modifications based on the data being managed. Ultimately, this strategy fosters a balanced governance model that respects the unique characteristics of different data types while promoting overall efficiency and security.
Standardisation in data management plays a vital role in achieving effective Data Governance across organisations. Consistent processes across different departments are essential; adherence to uniform standards prevents the complications that arise from varying approaches due to differing data types. Personal experiences from various projects, particularly those involving the implementation of data quality and metadata tools, underscore the necessity of a cohesive governance framework. Ultimately, embracing standardisation leads to smoother data management practices and significant improvements throughout the organisation.
Establishing a robust Data Governance organisation at the outset of projects is crucial for long-term success. Effective Data Governance goes beyond simply drafting policies; it encompasses essential elements such as data quality and metadata management, which should be integrated into a comprehensive framework before initiating individual projects.
By ensuring that governance aligns with organisational goals, businesses can navigate the complexities of data management effectively. Ultimately, a strategic approach to Data Governance not only strengthens operational efficiency but also fosters a culture of accountability and informed decision-making within the organisation.
Data Management: Policies, Processes, and Standards
In a Data Governance project, implementing policies, processes, and standards is crucial, and the willingness to adopt these can vary significantly across different industries and companies. While some organisations are very policy-driven and eager to create data policies to ensure consistency, others may find the process lengthy and cumbersome; for instance, one company estimated a two-year timeline for implementing its policy. In such cases, it may be more effective to focus on developing processes that can be implemented by the company itself, as having a written policy in hand can serve as a foundation for creating actionable procedures. Overall, the approach to Data Governance is highly dependent on the specific company culture and industry context.
Figure 8 Barcodes
Information Architecture and Data Governance
Merrill Albert highlighted the multifaceted nature of architecture, emphasising that it extends beyond mere design to include the management of operational aspects and the implementation of changes. Her experience with information architecture for a nuclear power plant serves as a prime example of the complexities involved in architectural projects, illustrating the critical need for effective change control. Ultimately, this discussion not only fostered valuable connections among participants but also showcased the diverse expertise present within the group.
The integration of architecture and governance in building processes is crucial, as many assume that once a project is operational, the involvement of an architect is no longer necessary. However, challenges arise in managing data, such as master Data Governance, which still requires oversight post-implementation.
Effective collaboration between IT and business professionals is vital, as business stakeholders possess essential operational knowledge and must communicate their requirements clearly to technical teams. Addressing the transition from architectural design to ongoing management ensures that systems are appropriately maintained and aligned with business needs.
Effective collaboration between business and IT is crucial throughout the data lifecycle, particularly during problem investigation and testing, where both parties should remain engaged. Often, businesses tend to relinquish data responsibilities to IT, neglecting their role in defining business terms that align with metadata and data elements. This disconnect can hinder proper Data Governance, which must be managed during the operational phase. It’s vital to differentiate between “data as stuff,” focusing on logistical aspects like storage and movement, and “data as meaning,” which emphasises business relevance and context. A unified approach ensures that both technical and business insights are integrated to enhance the overall data management strategy.
Involving business leaders in Data Governance is essential for effective data management within organisations. Since these leaders are responsible for the data, fostering collaboration among business personnel, rather than isolating a data team, can lead to more aligned and informed decision-making that mirrors the overall business strategy.
A data leader plays a vital role in guiding stakeholders through this process, ensuring their active participation in data management. Ultimately, this collaborative approach not only enhances Data Governance but also contributes to a more unified vision in achieving business objectives, with additional resources, such as a recommended book, available for further insights.
If you would like to join the discussion, please visit our community platform, the Data Professional Expedition.
Additionally, if you would like to watch the edited video on our YouTube please click here.
If you would like to be a guest speaker on a future webinar, kindly contact Debbie (social@modelwaresystems.com)
Don’t forget to join our exciting LinkedIn and Meetup data communities not to miss out!
