Abstract
In the pursuit of guiding practitioners in solving problems like: How do I understand better my information? The outlined framework becomes an instrument capable of supporting the thinking, organizing, and focusing on how to deal with the metadata. Therefore, it eliminates communication barriers between technical and business professionals. What does Metadata really mean and how can Metadata management help my business? This article clarifies those unanswered questions. The steps you need to take to define the effects metadata has on maintaining a data strategy is outlined, starting with a framework that team members can follow and properly express requirements for corresponding categorizations. This framework is the result of several years of work within many different industries and reflects best practices with metadata. Also, a step by step process ensures better adoption of metadata.
Metadata is Not a Technical Term Anymore
In recent conferences, I presented how I approach the dilemma of working with metadata in organizations. First, I noticed that most business professionals have no clue what it means, and that most of the time they are not interested in it at all. While surprising for some audiences, many Information technologists also hate to talk about it. Why is that so?
Complexity! The abstraction required to deal with metadata components is usually overwhelming to most professionals. It requires one to think of the right terminology for the business function that one is using and a different terminology for another. The technologists are usually only interested in the ones that satisfy the software platform they are using or are familiar with and assuming that there are so many out there, it’s impossible to know them all. Frustrating, no?
The only way to deal with what we have been calling the metadata conundrum is by defining it more clearly and organizing it in such a manner that you can focus on the component at stake at that particular moment.
The Business Metadata world uses many terms, names, data models, business stewards, data, information, knowledge, mapping, procedures, retention, privacy, policies, standards, definitions, categories, data inventory, processes, and so on, that can be conceptual, logical, and physical.
What does it all mean?
In reality, what matters is a much simpler list. There are business rules, relationships, calculations, derivations, sources, targets, and traceability – forward and backward. These are components that essentially will determine the context of how a certain datum is manipulated (captured, stored, moved, trashed). Essentially, the motive for using the datum has to be, otherwise we fall in the black hole trap, to solve a business problem. The problem is the key driver for gathering, processing, storing and retrieving the data.
It helps when you approach this dilemma with a definition that I have been preaching for the last ten years:
“Metadata provides a pedigree to the information: what the information is, where it came from and how it got there, what system it uses, and its relationship to the other pieces of information.”
Before establishing this definition, I noticed that my colleagues looked uncomfortable when I used the term metadata. Once I provided a definition that included the word “pedigree,” others then easily understood that data has a source and that it belongs to a certain “breed” that maintains its quality. I preach to those who need to talk about metadata with business professionals to use this definition, as the simple language applies often. I cannot say that it is a universal definition. However, it does apply to the majority of cultures.
I devised a 3-by-3 framework to explain the details of metadata further. It is a simple tool that allows organizations to be clear on the most known and relevant components of data, without requiring them to know formal classification systems.
On the horizontal axis, I use the most common sense grouping: descriptive, administrative, and lineage – I call this the structure grouping. Note that most of the earlier metadata authors emphasized metadata as descriptors of data, which in fact is the first cell in the matrix – cell (1,1).
On the vertical axis, I use business, platform, and operational. In the figure below there are examples of components in all nine cells.
The intent of the framework is to provide a tool for managing the large amount and types of metadata we face nowadays. While new technologies surface at every moment, I included the most promising one, block chain. Notice that the main three components of block chain fit nicely in the platform type considering that it contains the metadata embedded. And, of course, block chain is all about “pedigree.”
When a business dilemma arises, anyone can use the above matrix by selecting the cell where information needs to be improved, made more available and clarified further. This method ensures that efforts are well focused and that the organization achieves their desired goal.
On one occasion someone came to me with an issue where some data was altered in the middle of a large workflow and as a result caused surprising final calculations. The altered data was due to a currency rate exchange that was not recorded anywhere. The change in currency rate needed to be addressed, recorded, and available to flawlessly form the information pedigree. By cross-referencing operational and lineage in the above matrix (3,3), one would determine that a sky-hook analysis had taken place, and this is where the framework comes into play.
Also, my experience lets me suggest that a successful metadata implementation requires a process to follow through. The keys to establishing successful metadata include: defining, capturing, assessing quality, delivering, and sustaining. These steps provide a linear process for planning for metadata management.
Overview of the activities needed for planning and executing a metadata strategy
A step by step approach for firms and business owners help distinguish different levels of effort for managing metadata.
This step-by-step is guidance on what a typical businessperson will need to do to ensure metadata supports your quality efforts. These steps include defining requirements, analyzing what to inventory and developing a project plan, and how to select things for quality assessment. Communicate the delivery to stakeholders: it is paramount. Continuously validate the process to ensure the process is operationalized and sustained.
Define – make your words reverberate across the organization. Name things clearly and properly; use the cultural slang but innovate to improve understanding and improve communication.
Capture – ensure it is well documented and communicate, communicate, communicate. Use this opportunity to review on an item-by-item basis: (1) update/revise and check the terms, (2) delete duplications, or (3) rationalize similar ones.
Assess Quality – validate that it is in use appropriately with different audiences.
Deliver– Once updated, publish through all channels you can utilize.
Sustain – keep doing it, always.
