Data governance is the collection of processes, policies, roles, metrics, and standards that ensures an effective and efficient use of information. This also helps establish data management processes that keep your data secured, private, accurate, and usable throughout the data life cycle1.
When a C-level executive thinks of data governance, eventually thoughts run to, “How will this affect my budget?” or, more specifically, “What is this going to cost me?”
In fairness, setting up data governance, especially in a shop which is newer, or where data governance is a new concept, can cost quite a bit. On the brighter side, once standards are in place, you can often save enough money to pay for the costs of being organized and compliant.
Let’s set aside the costs involved and talk about some of the financial benefits.
Let’s start with indirect benefits, then we’ll move to direct benefits.
Treating data servers in a consistent manner focused on industry-exceeding best-practices improves up-time in a variety of ways. According to think tanks (and experiences CIOs), most estimates place long-term application development costs at about 7X the overall development costs (i.e. $1M in development costs, $7M in maintenance costs over the life of the application). In other words, minimizing maintenance costs over time saves a lot of money, and data governance (i.e. defining your practices) is a critical component of this treatment.
Next, downtime is minimized when standards are set and adhered to, which includes processes around production changes, especially data changes.
In addition, you get client and internal benefits when you’ve taken the time to identify and implement applications and processes that require high availability.
Finally, data governance reduced the chances of having sudden, embarrassing, or catastrophic expenses. This includes:
- Recovery time objective (RTO) and Recovery point objective (RTO) are sometimes assumed and not quantified, the effect of which means that you can never meet (much less exceed) expectations because they are not set.
- Defining these means you can set and meet real expectations, rather than be victim of assumed ones
- Losing the one person who knows passwords
- Territoriality amongst team members does not lend itself to long-term success with access during emergencies, or after that person wins the lottery and suddenly disappears, or is unavailable on vacation or at a movie
- Dealing with a ransomware attack and finding you don’t have resources or a plan to deal with it
- The need for things like air-gapped backups and a tested plan to deal with it, is paramount
- It helps if you’ve run this plan through your board so that they know what will happen and on what timeline. Nobody ever got fired for creating a plan, getting it approved by a board, and following the plan
Actual cost-savings get more interesting. Let’s look at an example.
We recently implemented data governance for a customer, a $70M health-care services company with a lot of data needs, a few DBAs who were territorial, and insufficient talent to manage the environment.
They were losing money in several places, due to excessive server downtime and poor database performance, costing some customers and losing fines due to Service Level Agreements (SLAs).
The first step was stabilizing the environment. This stopped the bleeding (customer loss, continuous “war-room” calls) and gave us time to take subsequent steps.
Next step was database tuning, which improves performance to the end user but also reduces server load.
This gave us time to gather requirements and to identify specific needs pertaining to uptime and high availability.
The reduction in server load became critical, for a variety of reasons. First, the plans to increase CPU and memory became unnecessary. This is more of an indirect benefit (eliminating the need for hardware purchases), but think about license costs, which are typically dependent upon core count. Improved performance = reduced core needs, which means we can reduce the number of cores, which means we can reduce the licensing costs of individual database servers.
The next step, though, is to analyze the overall architecture and reduce the number of servers in play. At a $70M company, we’ve reduced overall server costs by $250k/year, which reduction is reflected every year.
Reducing server and CPU counts has other cost benefits. This includes:
- Reduced footprint
- Reduced hardware costs
- Reduced upgrade costs
- Reduced security costs
While the costs of data governance is high, the benefits are also high, and over time, often lead to reduced costs and smarter spending.
Jeff Garbus is the Co-Founder and CEO of Soaring Eagle Database Consulting. Jeff is responsible for the technical direction, vision, image, and long-term growth of the company.
Thirty-plus years ago when Jeff consulted around the country on complex database problems, he observed a need for software to track his client’s database performance from anywhere in the world. Thus, Soaring Eagle was born.
Since then, Jeff has grown Soaring Eagle Database Consulting into a multi-million dollar firm. An expert in MS SQL Servers and SAP ASE (formerly Sybase) SQL Servers, Jeff has written 20 books to date on these subjects in order to help businesses overcome their database issues.
Jeff holds a Bachelor of Science in Computer Science from Rensselaer Polytechnic Institute. When he’s not consulting on IT issues, Jeff enjoys playing poker and tennis, traveling with his wife, and spending time with his kids and grandkids.