Whether the topic was Master Data Management, regulatory compliance, Metadata, or overall Data Strategy, a number of sessions at Enterprise Data World 2015 Conference both implicitly and explicitly addressed challenges in Data Governance.
The need to do so stems from the fact that the overall reliance on data for consumer, industrial, and governmental processes will keep increasing; numerous applications of Big Data, the Internet of Things (IoT), social media, mobile, and Cloud technologies will help continue the on-going data deluge.
What is less apparent, however, is the effect these technologies are having on Data Governance. In an interview on the vendor exhibit floor at the conference, Ian Rowlands, the Vice President of Product Management at ASG Software Solutions, commented on the ramifications of these various technological improvements on governance:
“We see that enterprises used to do governance to save money and then for regulatory compliance. Now people are starting to say we want to do governance to make sure that the way we use data matches our corporate values and treats our customers properly.”
The progression of each of those drivers for Data Governance only solidifies its importance across vertical industries. Similarly, the proper governance of Reference Data (the different categories of data that organizations maintain that are common across business domains, yet rarely governed as such) and its relationship to Metadata and business glossaries helps to reinforce that importance, and may be key for accounting for the copious quantities of data organizations will soon encounter. Rowlands reflected on the fact that:
“You start with what sounds like in all honesty a pretty utilitarian type of technology: a way of managing codes and code tables. But, as you start to think through the implications, it’s a really important thing for people to have under control.”
Reference Data First
There are numerous ways in which the proper governance of Reference Data is instrumental for not only managing Big Data applications such as the IoT, but also monetizing them. Governing Reference Data is innately related to maintaining proper MDM, Metadata, and definitions that are linked with business glossaries. Whereas Metadata “describes all of the other classes of data” according to Rowlands and Reference Data consists of different categories of data frequently found as codes, Reference Data actually depends on Metadata and definitions in the business glossary to provide meaning to Reference Data. Thus, platforms expressly designed for Reference Data, such as ASG’s ASG-metaRDM, stratify these three important facets of Data Governance through the use of Semantic standards. According to ASG Software Solutions’ Ken Pericak:
“Some Reference Data code tables are relatively static, but some of them that are industry specific change quite frequently. You need to make sure you’ve got an infrastructure in place to manage them, so when you’re working with all these different systems, we’re all talking about the same things. It could be a code of 01. That has absolutely no context unless you can tie that back into the semantic meaning associated with it. That’s one of the strengths of what we do; we tie back into the business glossary.”
Across the Enterprise
Reference Data platforms not only provide value for the management of Reference Data, Metadata, and business glossaries to generate meaning, but also do so holistically to avoid the silo-based tendency of managing Reference Data. One of the primary benefits of such a holistic approach is it reduces the penchant for producing costly mistakes that are bound to happen when there are uniform governance protocols for Metadata backed by an enterprise-wide business glossary, yet disparate means of managing Reference Data across business units. Eliminating Reference Data silos, therefore, aids with regulatory compliance and data quality. According to Pericak:
“The other key aspect of this is the governance aspect and regulatory requirements, especially if you look at the financial industry. You need to be able to define values and attributes of your data and be able to trace it back and that’s required [with] compliance. So certainly there’s a risk management aspect associated with it, but the solution also brings a governance infrastructure in place as well so you can have your stewards make sure that they’re managing the data.”
Big Data Governance
Well defined governance of Metadata and Reference Data is an integral part of deriving meaning from Big Data, which not only complicates conventional Data Governance due to the high speeds of ingestion, but also because much of it involves external, non-proprietary data. The confluence of business glossary definitions, Reference Data, Metadata, and semantic standards—ideally managed through a single platform with a unified user interface such as ASG’s ASG-metaRDM—can help provide valuable context to data that is otherwise meaningless, or which simply require too much time to sort through to leverage with fast-paced business opportunities. “Big Data can come at you from any which place in any format.” Rowlands observed. “You need to put some kind of discipline around that so you can make use of it.”
The practical example of a singular approach linking Reference Data, Metadata, glossary definitions, and standards is perhaps best illustrated with sentiment analysis from any popular social media site such as Twitter. The non-conventional ways that consumers discuss a product or service over these platforms can utilize various aspects of slang, abbreviations, and even different languages that can confound typical definitions of products and services. “Now how are you going to rationalize that?” Rowlands queried. “What you do is you use a Big Data technology to say if they say this or this or this or this, they’re really talking about this and we can rationalize that. That goes back to Metadata and Reference Data.”
The Internet of Things
With predictions for the IoT including estimates of nearly 50,000 connected devices, all continuously transmitting data simultaneously, this application of Big Data is perhaps its most formidable. The correlation between Semantics, Metadata, Reference Data, and business glossaries can help to provide numerous instances of targeted advertising for this phenomenon which is still in its earliest stages—giving organizations precious time to decipher ways to monetize it. The key to generating business value from the IoT may ultimately pertain to usages across what Rowlands denoted were its three principle stratifications: consumer, industrial, and governmental applications. He commented on one potential use case:
“[There’s] 18 million [smart] electricity meters. Now we start to aggregate data across them. We can start to tell people ‘you know what? This geography here is the areas where most power is being consumed. Maybe that’s somewhere where you can go and sell energy reduction solutions’. You can detect trends in usage and use that to generate marketing capabilities. That’s monetization: [you] take that data, aggregate it, and sell it.”
There are also substantial governance implications for the IoT that must be accounted for to take advantage of its marketing capabilities. Some of these are specifically related to Reference Data and the fact that although most of these data are relatively static, there are certain data types for which real-time management capabilities are needed on the consumer side. Certain transactional data, for instance, will need to be classified according to Reference Data and require near real-time capabilities to look up and identify these data, which reinforces the need for effective Big Data Governance.
As technologies evolve and data-driven processes become more refined, it’s natural for factors of Data Management to do the same. In this respect Data Governance is merely expanding to reflect the growing importance of data to the enterprise and the surrounding world. The hallmarks of governance—Reference Data, Metadata, standards-based approaches, business glossaries, and the processes and people that enforce them—are now firmly entrenched in a need to demonstrate organizational competence of Data Management to customer bases that today interact with businesses much differently than they did just a few years ago.