The Value of Mastering Data

a man and woman looking at a computer screen

To deliver value, organizations must have data that’s accessible, trustworthy, and secure, which makes data governance a fundamental requirement. 

As a data architect, I’ve helped organizations of all shapes and sizes develop data governance strategies. A vital part of this is helping them articulate what the organization needs to look like, assessing their existing data governance program (if one exists), and guiding them on how to improve their data maturity.

In this article, I’ll explain the importance of uncoupling data from systems, building a data dictionary, and leveraging automation. With these key elements in place, organizations are far better positioned to build data maturity and drive business value. 

The Importance of Consolidating Data 

A key aspect of data governance is uncoupling data from systems. Today, every business has a huge amount of data. However, often it isn’t stored in one place and is very much system dependent. On one side, there’s data supporting infrastructure, such as marketing and finance systems. The custodians of that data don’t necessarily need to know how it works and where it goes, but it’s vital for them to understand the standards that go along with incorporating information about customers and vendors. Meanwhile, on the business-facing side, data stewards work with that data at the domain level. While they may understand their customers inside and out and be subject matter experts, they also have a responsibility for the data that’s being stored, who gets access to it, and the security systems around it. 

When it comes to measuring the value of data, there are two primary key performance indicators (KPIs) that are improved considerably by consolidating data:

  • Number of users—The more people who are able to access the data, the more value it can deliver. However, it’s vital that they’re doing so in a secure way and that they have access to the most up-to-date source of data. 
  • Quality of data—In legacy systems, this comes down to data ingestion and whether the application correlates with the data domain architecture.. When data is received from multiple sources, you need to be able to ascertain which one is the primary source and which one delivers the best quality. 

All data needs to be integrated and managed to ensure that it’s appropriately represented at every level within the business. 

How a Data Dictionary Supports Enterprise-Level Data

Being able to define who can access an authoritative reference for a given data domain is key. A data dictionary has terms and definitions, just like a standard dictionary, but goes on to define appropriate use and who’s responsible for each piece of data. However, getting businesses to adopt data dictionaries can be particularly challenging. Individuals see the data they use as their own, but at a data governance level, we need to cut straight across that. It can be an uphill battle to get that institutionalized. Ultimately, it’s the enterprise that needs to control and manage the data, not each data steward. 

In this way, a data dictionary changes data from societal or domains of business activity into enterprise-level sets of information. You can think of it as less is more. The more you use a data dictionary to consolidate data, the higher the value of the data. It’s also easier to maintain and more trustworthy; you have more confidence in where it came from and all the fingerprints that have touched it along the way, as well as who can access it. 

The Benefits of Automation

One of the biggest parts of understanding data depends on the descriptions behind it: who’s the owner, what are the valid values, and what are the associated security levels. This must be defined at a job role level; you need to be able to define who can see the data. At an aggregate level, someone might have access, but at an atomic level, they may not. Every business has a responsibility and has to be sensitive to whom data is exposed and how it’s used. 

When we started doing data warehousing and integrating large volumes of data, we quickly realized that automation was vital. We needed a process for integrating, standardizing, cleansing, and consolidating data. Doing all that manually not only required a lot of effort but became problematic. In the past, analytics was 80 percent data recovery and only 20 percent analysis. With automation, we’re able to flip that the other way around. Mature organizations don’t spend so much time on data preparation; they spend time on analysis. By automating processes, you can build trust. You know that data conforms to quality standards, as defined by your data dictionary, and is being used appropriately. In this way, products that support automation are extremely important; they enable organizations to spend less time preparing data and more time analyzing it and driving value.

Assessing Your Data Maturity

When I undertake a data maturity assessment, it’s pretty objective. I’ll try to find out what data is used for, where it comes from, which applications use it, and who has authorized access. It’s a standard step approach that takes businesses from having no idea about their data to full enlightenment. The five key stages are as follows:

  1. Chaotic—Manually merging data, no proper audits, poorly defined roles and responsibilities, and no metadata-related policies.
  2. Reactive—Limited data standards, incomplete plans, restricted tactical options, and tools used on an ad-hoc basis. 
  3. Defined—Defined set of KPIs, well-appointed data stewards, and compiled data dictionary, although often without precise rules. 
  4. Proactive—Automated reports for data usage, well-defined processes, and proactive measures and improvements.
  5. Predictive—Automated processes; top-down strategies; and people, processes, and technology working in harmony. 

To get from one to five, organizations need to take incremental steps, including forming a data governance charter, getting support from senior management, building strategic objectives, and defining an implementation strategy. 

How Data Governance Drives Value

It doesn’t matter who you are or what business you’re in—everyone does data governance whether they like it or not. In my experience, businesses that implement a data governance strategy show maturity and discipline, while those that take it less seriously struggle. There’s a significant price associated with the inability to leverage data; those organizations are data rich and information poor. 

In order to translate vast sums of data into pearls of wisdom, you must be able to access data securely and with confidence. By creating a data governance strategy with the support of things like a data dictionary and automation, you can achieve data democratization, create a structure that’s organic and adaptable, and truly build value. 

About the author: John Murphy is the senior data architect at Emids and the data quality solutions architect at Southern California Edison (SCE). He has over 20 years of experience as an independent consultant for Fortune 500 companies and specializes in data architecture, data modeling, and data warehouse design. 

From the Privacera team (Please note, John Murphy is not employed by Privacera, and his piece does not imply an endorsement).

For more on increasing the value of data governance, watch our webinar: The Keys to the Kingdom.

Interested in
Learning More?

Subscribe today to stay informed and get regular updates from Privacera.