Gaining qualitative insights into your data, and beginning to distinguish your data assets from liabilities, is the first step in the quest for data valuation. Fortunately, that’s a process that can be accomplished fairly easily, as I’ve noted in another post about qualitative data valuation. Why should gaining such insights be a priority? There are many possible answers — but one that comes immediately to mind is that it can point the way to new approaches to increasing revenue and controlling costs across an organization.
What I want to explore in this post is how to take your insights into your data to the next level — a quantitative understanding of your data’s value — which requires a different set of processes, toolsets and skills. I’ll discuss what may be the two most fundamental of these processes: data governance operations and business metadata acquisition. Both of these processes help you drill into metadata — that is, the data about your data.
If you’ve bought into the need to gain more insight into your data’s value, and want to get started immediately, that’s great. But first, a word of caution. You shouldn’t rush the process of business metadata acquisition, nor should you hope to simply buy a product, plug it in and have the metadata handed to you. Rather, it requires a specialized process that emphasizes engaging with the relevant data stakeholders over a period of time and collaborating around your data in a focused way.
The ultimate goal is to provide a context around your organization’s various data elements and your data supply chain, a perspective that can help you make smarter decisions about where and how to invest in your data in order to improve results for the organization as a whole — so it’s worth doing it right.
Data governance operations
The science of data governance operations (DGO) can be a powerful and central component of the data valuation process. In practice, however, it is often obfuscated or watered down with buzzwords and lingo.
At its core, DGO is a process for applying guidance, control and meaningful investment of resources to help capture ‘tribal’ and tacit knowledge about an organization’s data. Broadly, its objective is to create, mature and invest in the organization’s data community. More narrowly, for the purpose of data valuation, its goal is to derive the meaning and purpose of given data elements.
There are various approaches to deriving this meaning and insight, but what we have found works best is to start by gathering stakeholders from the broader data community in a room (either physically or virtually). Then, following a structured agenda, we collaborate to collect the participants’ unique viewpoints, perspectives and thoughts on a set of given data. By doing this activity in real time and using aggregated visualizations and a defined “social voting” protocol, one can quickly identify which data community members have the most valid insights into a given data element, versus those who simply care about it. (In our engagements, we refer to this process as “Controlled Chaos,” as we discuss in our white paper.)
Some of the specific areas of business metadata that can help drive valuation include:
Meaning and purpose: cataloging various names, definitions, contextual use cases, acronyms and aliases within the organization
Voice of the community: capturing which data individuals care the most about and why, the relationships of those data elements with the organization’s business model, and what problems need to be solved
Data community: gaining insights into each of the community members who care about a given business data element, including any named data stewards, custodians or business owners
Business processes: analysis of the macro- and micro-level business process steps in which data is either created, changed or edited, referenced or consumed
Business risks: the security tags and, more importantly, the combinations of data elements that would result in a high-risk breach
The next stage in the process is one that is already familiar to most data architects, ETL developers and the technicians who build databases and data warehouses. The traditional practices of data architecture and metadata management that arose in the heyday of relational database management are still highly relevant and useful in acquiring metadata — especially for gaining insights into all the metadata that one can’t capture through DGO, as described above. These include:
Technical & Core Metadata. The technical category is enabled by metadata repository tools that typically go out to a database or mainframe and ingest such metadata as its structure and platform — i.e., mainframe or database, Oracle or SQL, and so on. Core metadata, on the other hand, refers to the administrative aspects of the technical information, and is obtained by interviewing the database administrator who oversees each data element to understand the names, aliases and any related technical projects used to create or enrich the information. We generally recommend gathering technical and core metadata together, because either one without the other is basically useless.
People Metadata: For every data element in your organization (except those data elements that are truly worthless), there are people who actually care passionately about particular instances of the data. Depending on the level of data, these stakeholders might include anyone from the database administrator to marketing analysts or other people who consume the data in the course of their jobs. You need to capture the names, insights and knowledge of each of these individuals, and associate them with each data element you’re analyzing.
Positioning yourself for a deeper dive
When you’ve conducted the data governance operations and metadata acquisition processes, you have the foundation for being able to determine and rank the value of your data. You can answer many related questions much more definitively, such as exactly where various data elements live in your organization, who cares about them, what they mean (all the way down to their physical implementations), and how each data element interacts with, supports, changes or dictates your business processes.