The Definitive Guide to Data Governance
Content Marketing Manager, Sigma
Data governance has moved from being a subject discussed mainly at IT department meetings to a hot topic dominating conversations by leaders at all levels. As more companies aim to become data-driven, everyone is realizing that data governance has a big impact on corporate sustainability. Smart decisions can’t be made with inaccurate data. Technologies like artificial intelligence and machine learning must be built on reliable data or they fail to deliver value. And privacy concerns and regulations require that companies have a firm grip on data governance — or face fines and reputation fallout.
But that firm grip has traditionally come with the consequence of suffocating curiosity. In fact, while nearly every large company has its sights set on being data-driven, a 2019 survey by NewVantage Partners found that 72 percent of respondents from these companies say they haven’t been able to create a data-driven culture. When companies enforce stringent rules around who can access data, how, and when, domain experts can’t bring their expertise (and knowledge of what questions to ask for insights that will lead to growth) to the table.
That’s changing. In this guide, we explore the traditional limitations of data governance and how new tools are shattering these limitations, the benefits that data governance can offer, and exactly why data quality matters so much. We also look at best practices and how to build your own data governance framework.
What is data governance?
Data governance is the set of internal policies and practices, as well as the monitoring and enforcement of those policies and practices, that serve to guide data asset management within a company. Data governance, in essence, enables a company to manage its data in a way that helps keep the organization healthy and empowers it to take advantage of opportunities.
In the past, a strong data governance strategy kept data in the sole domain of data engineers, data analysts, and BI professionals. Data resided in an ivory tower, guarded by those with the knowledge and skills to extract meaning from it. There’s still a fear that if a company opens up access, data assets will turn into a Wild West where “anything goes” — security is lax, permissions are wide open, and data of questionable quality gets used to make important decisions.
But the reality of the modern world is that being data-driven is no longer a competitive advantage — it’s a necessity. Without domain experts making good use of your data, you’ll be left behind.
Fortunately, the modern world also has modern tools, like Sigma, which dramatically reduce the need for data professionals to be intimately involved in analytics. It’s becoming easier to centralize data, break down silos, and get data into the hands that need it — without compromising security or quality. It’s now possible to have the best of both worlds.
Data governance benefits make the challenges worth it
If you’re still not convinced that a strong-yet-flexible data governance strategy is worth prioritizing, consider the benefits that demonstrate why data governance is important.
Greater ability to comply with regulations
The most obvious benefit of data governance is that it allows you to more easily comply with regulations such as GDPR and CCPA. The rise of data lineage requirements and policies that promote the “right to be forgotten” have changed the game. Data governance strategies, by definition, create procedures for performing risk assessments, identifying the level of security needed for each dataset, implementing necessary tools and systems for compliant security, creating policies around data lakes and warehouses, etc. With a smart data governance strategy, there’s no need to wonder if you’re truly compliant or not.
Lower data management costs
Even apart from the savings that come from compliance and better strategic decision-making, data governance programs also deliver savings in the form of lower data management costs. First, there’s no need for the level of data duplication that most companies have in place. Data governance identifies which data needs to be duplicated and which doesn’t, freeing up resources. Second, making data management processes more efficient through governance also reduces costs.
Data governance delivers transparency in valuable ways. It improves collaboration between everyone in the organization, allowing you to take advantage of domain experts’ insights. When you’re able to bring a variety of perspectives to bear, you’re more likely to come up with a good solution quickly. Transparency also gives you greater visibility into your operations and processes so you can see opportunities more clearly.
Increases the value of data
Data governance makes your data more valuable because you can count on it to be accurate and up-to-date. Domain experts know where it’s located, what it means, and how to put it to use through high-quality data models. When you have confidence in your data, you can move forward boldly on new initiatives based on insights derived from that data.
Want to learn data modeling best practices with Sigma? Read our free guide.
Facilitates training and education around data assets
If you’re going to allow domain experts to access and use data, you need to teach them data skills so that they know how to make relevant queries and use data in context. You also must train them on where they can find vetted data. Data governance makes it easier to systematize education around data for those who don’t have experience in analytics. This promotes greater data literacy within organizations, and puts your company in a better position to extract the full value of your data investment.
Better monitoring and tracking for data quality
Data quality is a huge benefit of governance. Without quality, you’re essentially shooting in the dark — you have no way of knowing whether your decisions are good ones or whether your initiatives based on data will be successful. And any technology built on data of questionable quality won’t be truly valuable. Data governance puts monitoring and tracking processes in place that make it clear which data sources are vetted and which are not fully reliable. Good initiatives allow for data and business teams to work in concert so that the right data is available for business analysis at all times, not days or weeks after a question arises.
Ultimately, data governance results in cleaner data, better analytics, better compliance, smarter business decisions, and improved business results.
TALK WITH AN EXPERT
Have questions about making data accessible without sacrificing compliance and security? Talk to a Sigma expert.
Managing data quality for a stronger company
Trust is foundational to nearly everything important in life, and data is no exception. With a strong data governance program, you can quickly see which data is not to be trusted due to being inaccurate, out-of-date, or out-of-context. And you know which data is vetted and identified as relevant. Yet recent research by Trifacta showed that nearly a third of respondents have serious data quality issues. Quality is still aspirational for many companies, despite how important it is.
The impact of data quality
What’s the impact of data quality? Let’s take a look at a couple of different industries to get a high-level view of why quality is such a big deal.
Supply chain — The modern supply chain is complex. Many elements are involved, and a variety of factors influence demand, from influencer-generated fads to natural disasters. With customers expecting fast delivery, retailers must accurately track demand in every location for products to predict where inventory needs to go. Bad data could mean delayed deliveries, backorders, and lost sales.
Marketing — Marketing lives on data. The more a company knows about its target market segments, the closer it can match messages, offers, and media preferences — and avoid misdirected ad spend. Marketing professionals regularly analyze trends, keep up with KPIs, and track even the smallest actions taken by prospects and customers. Without accurate, up-to-date data, campaigns could easily miss the mark and prospects that should have gone into the sales funnel slip away.
Finance — Banks and other financial institutions rely on data for market analysis. Getting an analysis wrong can wreak havoc. Additionally, highly-regulated industries like finance (and healthcare) must ensure data quality or risk fines and reputation damage.
Essentials for data quality
The level of data governance required to ensure high-quality data takes work and planning — especially if you want to maintain democratization of data. And you’ll need the cooperation of people throughout the company. Here’s what you need to be thinking about as you build out your data governance framework for data quality.
Get buy-in from all departments — It’s essential that you get buy-in throughout the company because you’ll need everyone to do their part during rollout and in ongoing practice. Build a strong business case specific to your company to show that the initiative will have broad impact on the health of the company (which benefits everyone). Additionally, dive into the micro level and identify how data quality will impact the success of each department individually.
Define KPIs to describe what constitutes quality — Everyone has a different idea of exactly what “quality” means. How will you measure accuracy, completeness, relevance, and how clean and up-to-date data is? Define the KPIs that you are targeting in each of these areas.
Include data quality activities in your data governance framework — Describe the data quality activities that you’ll be implementing and their schedules in your data governance framework. It’s not enough to hope that these activities rise to the top of everyone’s to-do lists. It requires prioritization.
Keep a data quality issue log — As much planning and preparation as you do, data quality issues will sometimes arise. In these cases, it’s important to know the nature of the problem, where the problem occurred, and why the issue was able to make it through preventative measures. A data quality issue log will help you track and analyze these problems so you can iterate and reduce future occurrences.
Look to solve quality issues at the data onboarding point — Most issues happening downstream are caused at the onboarding point. Always track issues back to this point to better identify what’s causing the issue. You can reduce the chance of future issues by adjusting onboarding practices and procedures This is something that will need to be done regularly to maintain high data quality.
Better data quality makes a company stronger. With it, teams can better allocate resources and make smarter decisions that affect profitability and sustainability.
Data governance best practices
Data governance can seem unwieldy because each company needs to develop a customized framework with tailored systems and processes. But there are several best practices that will help you as you’re crafting your data governance strategy and as you’re implementing it.
What might be surprising to many people is that data governance best practices have existed for a couple of decades. The questions have already been answered. We now need to force ourselves into the discipline of following those best practices.
Chief Data Evangelist, Snowflake
Start small — Aim for quick wins. As stakeholders see the impact of even small steps, they’ll be more likely to get fully on board and participate moving forward.
Decide roles and responsibilities — Nothing gets done without assignments. You’ll need to identify who will be on the data governance team and what roles they’ll play, along with their responsibilities.
Develop standard definitions — Everyone needs to be clear and on the same page in order to work together. Create a solid data model to centralize data definitions so that anyone can use to find and understand the data in a data store.
Map infrastructure, architecture, and tools — In order for the program to run smoothly, you’ll need to be sure people understand your enterprise infrastructure and architecture and the tools for data governance.
Set goals and track progress — What does success look like in each area of your governance strategy? What milestones do you want to hit when? You’ll also need a way to track your progress (and don’t forget to celebrate when you hit your targets!).
Communicate — Clear, frequent communication is a must. Anytime you roll out a new initiative, it means changing habits, helping people learn something new, redirecting workflows. It can be challenging for people who have other primary responsibilities, and communication can help them overcome the hurdles that lie in the way.
Remember it’s an ongoing effort — Data governance isn’t “set it and forget it.” It’s an ongoing program that must stay a priority in order for the company to reap the benefits.
These guidelines will help you with implementation, but what about the framework itself? Let’s dive into the practicalities of putting together a data governance framework that serves your organization.
How to build a data governance framework
A data governance framework can be simple or complex, depending on what your company wants to do with data governance. The purpose of a framework is to enable data governance standards to serve your company. Only you know what you need in a framework. With this in mind, here’s an overview of how to make the decision of what your framework should address and, for reference, what most companies include.
Customize your data governance framework
Start by identifying what your company is struggling with. Do you have challenges around reporting, data quality, data access, duplication, siloed data, or other issues? Run a SWOT analysis on everything data-related in your company to find where weaknesses and opportunities are located. This analysis will guide you as you create your customized framework.
Components of a data governance framework
Most data governance frameworks will include the following:
Business drivers — It’s a good idea to include your business case for data governance and what’s driving the need for the initiative. This will keep the WHY front and center for everyone.
Objectives — Include your goals in your framework, specific to each area you’re targeting. For example, you’ll probably include goals related to data quality, data security and privacy, architecture and integration, and use of data warehouses for BI, among others.
Strategies and methods you’ll use — What are the systems, policies, and processes you will put in place to address each objective above?
Enforcement — How will you enforce your policies and procedures? Particularly think through how you’ll enforce without creating bottlenecks. Again, tools like Sigma can help with this.
Technologies — What tools and resources do you need to help you implement your data governance process? How will you use them?
Tracking and measurement — Detail exactly how you’ll track activities and measure your progress. This element also includes communication — how will you communicate to each stakeholder, and in what forms and frequency?
These components are a good starting point, but don’t feel forced to follow a set template as you craft your data governance framework. What’s important is that you think through all the angles that affect your organization and what you want to achieve to build a framework that will help you accomplish your goals on an ongoing basis.
A word about balance
The theme of balance keeps showing up as we discuss data governance, and for good reason: Your governance strategy must ensure compliance with regulations and dictate procedures that create data skill competency. But if your strategy is too strict, you’ll generate bottlenecks that hold people back, with too much reliance on data professionals. The balance you strike will be specific to your organization and your needs.
Ultimately, you’re aiming for a customized, flexible governance strategy that takes advantage of data professionals’ expertise while giving domain experts the power to access vetted systems and use data skills they’ve been educated on for day-to-day workflows. This balance will allow you to enjoy the freedom to utilize insights within the guardrails that ensure compliance and security.