How to succeed in governance projects  

Data governance is a necessary step in creating a healthy and robust environment that drives business decisions and eases barriers to information across any organization. That being said, it’s gotten a bad rap. Finding someone who enjoys doing documentation is a rarity, and at its core, that is the foundation of governance. The goal is to improve transparency, break down barriers to data, and increase time to value, but that can become difficult when there are other more visible priorities.

Demystifying Governance Misperceptions 

One of the biggest concerns I hear about data governance is around scope. The topic is massive. Especially when approaching a long-established data infrastructure, this can quickly become intimidating. And on the surface, it is. Companies often have sporadic documentation, data dictionaries or model mapping from years past that are no longer relevant, and conflicting use of terms depending on the division or business unit. Getting your arms around that is quickly overwhelming when considering taking on a governance project. Attempting a governance project has become so stigmatized as difficult and unachievable that the word governance itself can cause team members to put up their defenses.

That doesn’t have to be the case though. That shouldn’t be the case. Governance can be approached in bite-sized chunks with quick implementations and incredible gains in efficiency. Learning a language takes time, and knowing the basics is important before moving on to more complex topics. The same applies to governance. The goal is to have your organization speak the same language about your data. Start small with foundational needs.

For example, an insurance organization centers around receiving and handling claims. The word claim is often used differently depending on your team, what context it’s being used in, and the conversation being had. An account manager may see a claim in a completely different way than an analyst or a developer. Miscommunication in this can, at the very least, cost a lot of time and stress attempting to request a report or create a new development project based on client feedback. At worst, this can mean incorrect or misinterpreted data going out to a client, an auditor, or the public. Focusing on this single term and coming to an enterprise consensus on the definition can immediately mean less time in meetings arguing over what a claim is and more time doing the work that needs to get done. Beyond that, identifying technical resources across the organization, fitting the claim definition, and assigning them that definition in a centralized place reduces technical teams’ time to map resources across systems. Lineage ensures that transformations and sourcing are consistent and transparent. And productization gives the business rep access to that claims report at the push of a button instead of requesting another redundant report.

Tactics for Success 

As I mentioned, the word governance can send people scattering. Internal marketing, change management, strictly defined scope and proper resourcing are all key factors in ensuring a governance project doesn’t become a failed experiment used to deter future attempts..  

Internal Marketing 

Perception is important, especially when implementing a large-scale and long-term initiative. While your approach to governance should be small, it will immediately be thought of as massive, a “boiling of the ocean” unless you quell those concerns early on. Work with the core team to define the scope of the first project strictly (more on that below) and identify ways to communicate it with the wider organization as part of a kickoff. Governance has an impact enterprise-wide, so ultimately, everyone will be involved in the initiative long term. Socializing early in a more formal way will give people time to get used to the idea and not feel as overwhelmed when you begin to reach out with questions and requests.

This early communication is also where you want to incorporate a show of buy-in from executives and individual contributors alike. There will be some early adopters of the program who understand the need for governance and are enthusiastic about being involved from the outset. Try to identify those individuals before announcing to the enterprise and get them up to speed on the initiative. These early adopters will be able to communicate about governance needs from a business or development perspective and indicate this is not just a data team initiative but something for the whole company. This is an important concept to communicate so all the work doesn’t fall to the individual team and instead is spread across all divisions.

Beyond individual contributors, front and center executive sponsorship will show support from the outset. When meeting with the broader organization, they should express the need for governance and how it will impact other company priorities. This not only incentivizes more people to attend the kickoff but will also give additional authority to later requests.

Change Management 

Governance is a big change for any organization. Often, there is new technology to learn, time allocation to contend with, and established processes teams are unwilling to part with. Effective change management for governance always has to start with collaboration and meeting teams where they are. Understand what processes and documentation already exist and what will work going forward. Hear their pain points and identify where governance can solve these problems. Provide training and workshops along the way that focus on functional ways to use the tools and strategies that have been created. Ultimately, the goal of governance is for greater data usage. Easing into the process through early collaboration will ensure this habit becomes part of their day-to-day as opposed to an on-time project. 

Scope Definition 

Governance projects can get big fast. There is a lot of data hiding in nooks and crannies, and it can become easy to continue adding data assets and domains to an individual project. But governance is an initiative, not just a single project, and top-loading with a large scope will squash the enthusiasm and momentum gained. This is a sprint, not a marathon, and strict scope definition and adherence will drive the first and all subsequent projects to success. A few ways to build a well-defined scope for a pilot governance solution include: 

1.  Identify a priority of the organization, or a new initiative that requires cross-departmental involvement. 
2.  Define a specific use case supporting that priority or initiative.  
3.  Working with data owners, developers, and business users, understand the biggest pain point they’ve had with the use case data. 
4.  Consolidate that pain point into a few data assets; tables or schemas. We recommend using no more than five in your pilot program. 
5.  Define critical data elements for those assets. These are the elements that are impacting or causing the previously identified pain point.  
6.  Understand which features to implement, such as catalog, lineage, or quality. 
7.  Build a scope document. This can be a quick write up, a deck, or even an email but ensures that those involved are aware of the scope. 
8.  Assign a project or program manager to the initiative who will be responsible for strictly enforcing that scope. 

The curation of assets as part of data governance can be a complex and time-consuming process. Initially, there may be pushback on the rigidity of the pilot scope or attempts at scope creep, but ultimately, remaining steadfast on this will ensure a project delivered on time with all requirements met. This sets the governance initiative up for success with a visible win that will build the momentum for a long-term iterative solution.

Resourcing 

Documentation is a fundamental aspect of any profession, especially as business units have become more siloed and tasks have become more complex. Unfortunately, this is often overlooked and only comes to light when it’s most needed. Everyone is busy, and documentation ends up falling by the wayside. Ultimately, data governance time spent can be transferred from one form of documentation, maintenance, and communication strategies to another more centralized strategic method. But if documentation isn’t currently happening, no time can be allocated to governance. On top of basic interactions, additional roles must be filled to steward, curate, manage, and drive a governance program forward.

Current employees can do some of this and will not require a full-time resource. However, especially at the outset of this type of initiative, a significant amount of time needs to be spent on strategy, long-term planning, training, and overall coordination. A data governance center of excellence is needed to focus the initiative and be a central point of contact for departmental data stewards, owners, and users. Even large enterprises can maintain this as a small (usually less than five people) using the hub and spoke governance method to scale. It is essential though, to support current resources as they begin their governance journey and continue to iterate

The Cost of No Action 

Data governance can be big, and it can be scary. It can also enable organizations to be more proactive in their data habits and usage, improve data quality, allow for transparency, break down silos, initiate collaboration, create data consistency, reduce miscommunication, and so much more. Not taking the leap into data governance is costly in a number of ways, including time and reputation.

Time: Data and data use is a language in and of itself. So when using data there is often need for translation. A business analyst requesting a report may need to understand payments, but do they need completed payments, open payments, late payments, payments for a certain geography, payments with tax and fees, payments without tax and fees? These conversations to scope out a business need require highly skilled resources and a significant amount of back and forth to come to a resolution.  

Additionally, without governance developers spend large portions of their day searching for the right columns in the right systems, attempting to understand downstream impact, and fixing bugs caused by upstream changes. Understanding the full lineage of an organization’s data infrastructure is complex even in the best state. Offloading this work into a governance platform reduces time spent and errors caused, allowing for more time to do the work of development and analysis. 

Reputation: An organization cannot function if it does not have the trust of its clients, employees, and the public. Inconsistent or inaccurate data can cause both internal and external clients to develop skepticism towards reporting functionality and may move away from using it altogether. Beyond data accuracy, the real threat of data leakage looms over all organizations maintaining any type of data. PII or PHI are especially concerning, and without insight into what data is in those categories, where it is in the system, or who is using it removes the opportunity for preventative measures in risk management and data security.  

Product Opportunity Cost: Most organizations have piles of data that is underutilized. Information that can be productized for clients or utilized by internal DS/ML and analytics capabilities often go unnoticed or are maintained with so much other data that these opportunities are obfuscated. Without governance there needs to be a trigger, a request, to drive how data is manipulated or used. By organizing, cataloging, and presenting data assets in a consumable way there are untold possibilities to what can be built. And by making this browsing capability available to the entire enterprise you open up the opportunity for more people to understand and have creative ways of using the data for revenue generation and margin improvement. 

Final Thoughts 

Governance is a continuous, iterative process that should become an engrained habit in any organization looking to optimize its data usage. Focusing on small wins early on and driving enthusiasm through collaboration will ensure a successful and long-standing governance program. Focusing on enterprise priorities while understanding and alleviating individual concerns is important, a fundamental principle here at CTI. The best time to start governing was when your organization started using data; the second-best time is today, so let’s get governing!

Amanda Darcangelo is a Senior Consultant at CTI Data.

Contact us for more information on how we can help support your Data Governance initiative.