Earlier this month, I had the privilege of delivering the keynote at the Data Governance Winter Conference. I’ve been a data governance trainer, practitioner and enthusiast for well over a decade. Yet I can’t recall a period of time that is as exciting, as momentous, and as dynamic as the here and now. Today. Living in a data-never-sleeps reality where we produce 2.5 quintillion bytes of data daily. Where data-driven is not just a vague corporate sentiment but truly a cultural shift — and anchor — for many organizations. Where nearly 90 percent of businesses have adopted or have plans to adopt a digital-first strategy.
Data has become the lifeblood of our digital economy. And in order to extract its full value, data must be managed and governed. Hence the title of this article and the theme of my keynote: the future of data governance is now.
And … data governance is different now.
Evolution of Data Governance
Data governance is by no means a new practice — but a constantly evolving one. When I first started in the industry, it wasn’t really called “governance,” but people were doing it anyway. They were coming together to make decisions around data, such as how to define customer, how to determine the best representation of a customer and how to ensure the record is of high quality, among countless others. Many times these activities were done in response to another project, such as an ERP implementation, the creation of a data warehouse, or the implementation of an MDM solution. Even though the governance was project-specific, it was still governance. And as demand grew, companies realized the value of being more organized around governance.
Over time, governance was given extra fuel with regulatory requirements such as Basel I, II and III and Dodd-Frank, which sharpened organizational focus on regulatory reporting. As new trends, technologies and business drivers like self-service analytics and data privacy emerge, data governance will continue to shift to meet the demands of data-driven businesses.
It was energizing to be speaking to an engaged crowd at my DG Winter keynote, “The Future of Data Governance is Now: Key Focus Areas for 2019.”
Digital Transformation Journeys Are Underway
Whether the goal is to enable an improved customer experience, increased speed of innovation or improved time-to-market, it’s clear most businesses have embarked on a journey toward digital transformation. This is happening on a daunting scale across all industries, with different levels of maturity.
And this is time-critical. It’s the feeling that if everyone is doing it, then the first one across the finish line wins. Companies are scrambling to figure out what digital transformation means to them. The purpose and pace will be different to different companies, but there is an aspect that will mean the same thing to all of them — the amount of data created with digitization will skyrocket.
Because data governance ensures the right people are involved in determining standards, usage and integration of data across projects, subject areas and lines of business, sound data governance will underpin all digital transformation initiatives.
Where to Focus Governance for 2019
Where should you align your data governance efforts for maximum results? Since my keynote was only 50 minutes long, I targeted these key trends/areas where an organizational focus can unlock significant value for any business:
- Data Understanding
- Data Privacy and Ethics
- Expansion of Data Governance
- Data Governance in the Cloud
- Diversity in Data Teams
I’ll share some key concerns and considerations for each focus area, and I encourage you to consider and implement the ones that will drive the most impact for your business.
Data Understanding
Data understanding is the ability to know where the data comes from and what it means. In a nutshell — I’m talking about provenance, source and lineage. Data is everywhere in the enterprise and is being acquired externally at a rapid pace. The provenance, or lineage of that data, is critical to ensuring appropriate use. If you don’t know where it comes from, how can you trust it? How do you know its level of quality?
Metadata is foundational to all data work and should be the top priority of a data governance program. Without an understanding of what data means, where it comes from or how it’s classified, it’s virtually impossible to progress anything else. You cannot define quality levels, access rights or usage guidelines. So if you’re going to understand both the data and the value you can derive from that data, then you need to have metadata.
And it’s important that your metadata strategy supports key business objectives and direction. Metadata for metadata’s sake is not a sustainable program. Determine what metadata you need to support the digital transformation, to comply with the General Data Protection Regulation (GDPR) or other regulations, to control data breaches, etc. If your digital transformation effort is focusing on maximizing customer experience, then you focus the process to gather, manage and improve metadata on those data elements that are needed to action the Customer Experience program — customer master data, contact data, sentiment data or purchasing patterns.
The same approach goes for data lineage. It will be a very expensive and lengthy exercise to identify and record all the lineage associated with all of your data. Focus on the amount of lineage that is important and create some criteria to say when “enough is enough.” This goes for the granularity of the data lineage as well. You can start with the big picture data flows, then dive down into which attributes of the data elements come from which systems, then how they get combined and transformed. And for some data that is used for regulatory reporting, it will be necessary to have significant granularity and accuracy. For others, perhaps not so much.
If time and money is invested in metadata, it’s critical that people can access it and use it to ensure they understand what the data means, where it comes from, its accuracy and timeliness. This means being able to easily see the metadata for data at the application layer, by a hover-over capability, for example. Maybe it’s an intuitive and easy-to-access data catalog to help report builders and analysts know what data is available, how others are using it or how frequently it is used. Preserve instances of collaboration and the associated artifacts to drive productivity and efficiency. Focus on tool capabilities that not only produce, but also propagate work.
Lastly, building out a corpus of metadata can be a long and arduous process. If it is up to a select few individuals, it will take a while. Determine a way to leverage crowd-sourcing and then when people do contribute, provide them the recognition and appreciation for their effort to encourage them to continue to do so — and influence others to contribute, as well.
Data Privacy and Ethics
Privacy is the concept that certain data is so sensitive to a person that it shouldn’t be shared, whether that is because of risk associated with that data, or because it’s personal. Privacy has a legal basis and legislation to protect it in most countries in the world. By contrast, ethics is a voluntary code that outlines personal responsibility. Ethics comes into play in both the management of the data, as well as the usage or how that data is involved in decision-making.
Privacy regulations are becoming more voluminous and more stringent: GDPR, ePrivacy and the California Consumer Privacy Act (CCPA), to name a few. With a focus on the individual and the consumer, the most granular level of customer information, these regulatory requirements can be quite challenging.
In our digital world, we are creating, using and sharing data constantly. Simple guidelines and standards cannot cover all circumstances. As data governance and management professionals, we need to recognize that we are at the center of how data is created, collected, shared and used, and that the decisions resulting from this can be ethical or unethical.
Keep in mind that what people consider to be private or ethically acceptable depends on culture and age and can change with time.
Although the average cost of a data breach last year was $3.62M, the actual cost of non-compliance to GDPR or the reputational damage created by either a privacy or an ethical breach could be fatal. Governance has to be involved in the projects that use the data to ensure they comply with the privacy standards. Initiate privacy and impact assessments for all projects, and incorporate privacy by design into the development phase of projects and software enhancements. And create a culture of data privacy by educating people on how to integrate privacy practices into all data management activities, data sharing and data usage.
You may also want to consider creating a data ethics framework as a way to enable an ethical culture. The framework should call the principles and guidance around what is considered ethical (and what isn’t) and could include a data ethics policy. This sort of “compelled” approach can help people recognize the purpose and value of data ethics and can help to start a more ethics-aware community.
Ultimately, the goal is to not just have a policy, but have a data ethics culture that uses that framework to make ethical decisions with data. To make that a reality, it’s important to leverage all your organizational change management capabilities to drive awareness and adoption of a data ethics mindset.
Expansion of Data Governance
Governance needs to demonstrate value beyond regulatory requirements to show direct business value from data. And the potential for governance extends beyond merely data. For example, understanding, transparency, auditability, quality and trust are also required for reporting, analytics and models, too.
Build upon the foundation you have already created and don’t reinvent the wheel. Be focused and have an incremental growth strategy. You will never be able to capture 100 percent of metadata, nor 100 percent of lineage. So focus on what provides business value and how you can start with a kernel of value and then incrementally expand.
Remember when we all rallied around the data governance ethos of “think program, not project?” Now it’s time to “think operationalization and embedment, not program.” For governance to unlock full value enterprise-wide, it needs to be for data what Human Resources is for people.
Demand for technology to support data governance is also becoming greater as the adoption of governance extends across the enterprise. More work needs to be done more rapidly by fewer people with more transparency and auditability than ever before.
This changes the perspective of data governance to providing support as close to the point of data usage as possible, with the goal of empowering data citizens to govern themselves. Of course, with freedom comes great responsibility. This is where creating a culture of data privacy, a culture of data ethics — and, in fact, a culture of data governance — is critical. At a certain point, a governance team can only do so much. If they can focus on empowerment, not control, they will be able to support more people with higher volumes of demand and data.
And it’s not just the volume. The range of capabilities needed to be involved creates demand for technology support. The pace has changed, so take advantage of machine learning and AI to complement the people involvement and get the work done in the time needed to do so.
Lastly, don’t forget to measure value delivered. A 2018 McKinsey study showed in digital transformation efforts, the most successful ones focused on assessing and measuring the impact of the change created. Among respondents who reported their organizations monitored KPIs as part of implementation, 51 percent reported success, compared with only 13 percent among those who did not monitor KPIs.
Data Governance in the Cloud
In a cloud-first environment, there is a feeling of less control over the data and, in fact, the movement of data to the cloud creates demand for even more governance. Do you know what data you have in the cloud, who put it there, whether it should be there and how it’s being protected?
Data integration between multiple cloud-based applications can be complex. The new cloud-based toolsets require integration of data governance practices. While there may be a clear value proposition for putting your actual data in the cloud because of the volumes, when it comes to metadata, it may be more difficult because you have to understand your metadata of on-premise solutions, as well as that of cloud solutions.
Although this goes without saying, make sure you’ve verified the vendor’s governance approach. In some cases, the security and privacy standards are higher than you can accomplish in-house. Taking a data-centric approach to the development strategy means understanding your data and information requirements, as well as your functional and technical requirements. When implementing cloud solutions, it’s important to understand the data needs of and for those solutions, so you can incorporate those requirements into the cloud implementation and determine the gap between what is provided in the solution, and what you need to do to complement it.
Cloud is becoming the new normal because it provides a low-cost alternative to what you have today, so many companies have cloud-first strategies. If you are moving to a cloud-first strategy, take advantage of the opportunity this creates to potentially re-engineer from a data-centric perspective. Determine how shared data will be governed, and create data-sharing standards and leverage metadata to track and manage that sharing.
Diversity in Data Teams
There’s a great quote by Jesse Jackson that I included in my presentation. “Inclusion is not a matter of political correctness. It is the key to growth.”
For data management, inclusion means crowdsourcing and diversity, which facilitates innovation and, therefore, drives competitiveness. One of the byproducts of inclusion is cognitive diversity or differences in perspective and information-processing styles. Essentially, it’s the difference in the way people think and solve problems. Diversity increases creativity, the opportunity to understand a wide variety of perspectives and the ability to consider a variety of viewpoints when trying to make a change or initiate a new program. This broad lens is critical to successful data initiatives, because data is everywhere and has a varied impact on people across an organization.
Again, because I had to focus on the time available during my presentation, I highlighted gender diversity in particular. In my home area, the tech industry, specifically VC, is king. Did you know that of the top 100 VC firms, only eight percent of investing partners are women?
Building a data future with women is important for two reasons. In this fast paced, data-first world of change, we need to do everything we can to be competitive and ensure success. Secondly, there is a huge shortage of skilled workers. According to the World Economic Forum, 54 percent of the workforce will need re-skilling to transition into the Fourth Industrial Revolution, spurred by the fusion of new technologies.
The good news is that in data, we are making great strides in gender diversity. In fact, Gartner predicts that by 2021, the Chief Data Officer role will be the most gender-diverse of all technology-affiliated, C-level positions. I encourage you to keep the momentum going to drive success of your programs.
As the pace of change accelerates, we need to increase the opportunities for innovation, and create teams that are more likely to produce creative solutions. Teams with gender diversity are more likely to experiment, be creative and share knowledge. Recent studies have also shown that gender diversity is correlated with both profitability and value-creation.
It’s not about preferences without expected results. It’s a recognition that there is a shortage of skills in the marketplace, and we need to take advantage of all the resources out there. Digital transformation alone will require new skillsets and resources.
And for my women in data peers, don’t be afraid to reach out and ask to own the data governance program if no one is stepping up — or to apply for a data position that seems beyond your reach.
I’ll leave you with one of my favorite quotes by the author and poet Erin Hanson.
And you ask “What if I fall?” Oh but my darling, What if you fly?”
