Proactively recognize when you need governance, build a solid foundation, and get buy-in
This article was co-authored with Otávio Leite Bastos (Global Data Governance Lead @ Contentsquare), Nandini Tyagi (Founders’ Office @ Atlan), and Prukalpa Sankar (Co-Founder @ Atlan).
As any data and analytics program evolves, it’s inevitable that data leaders will have to dive headfirst into data governance. Data governance has many facets such as data quality, access policies, data security, metadata management, data management, and more. But when is the right time? Who is needed? How do you even get started?
Historically, data governance was an afterthought to data and analytics architecture and use cases. This led to many organizations scrambling to reactively address problems related to data quality, data accessibility and data security, among others.
In this article, we explore how data leaders can more proactively recognize when you need data governance, build solid foundations for a data governance strategy, and get buy-in for your initiative.
Check out Atlan’s recent Masterclass from Contentsquare: How Data Governance Accelerates Contentsquare’s Analytics and BI.
Recognizing the Need: When to Buy a Data Catalog
A data governance team ensures the delivery of trust (verification of data sources and protection of PII) and clarity (well documented and accessible data products) around data to every decision-maker. To do this, there needs to be a centralized place that brings these data governance principles to life.
In the modern data stack, this is the data catalog. Unlike traditional data catalogs, these next-gen catalogs must be able to activate metadata to support all the facets of data governance.
So when is the right time to buy a data catalog?
There will be early signals that your team needs a catalog. Detecting these signals will be a mix of listening for qualitative feedback and quantitative analysis.
Some early qualitative signals to listen for:
- Analysts are unsure what data sets they can use, and not sure if they can trust them.
- Different teams are calculating the same metric in different ways.
- Analysts and business users are unsure what metric definitions even mean.
Quantitatively, signals that you need a data catalog revolve around the ever common time to value metric data teams often fixate on. Create a baseline time to value calculation and monitor how this metric changes over the course of a few weeks to a month. To calculate time to value:
[Project Delivery Date] — [Project Committed Date] = [Time to Value]
This can be done using data from your agile project management tool. If you aren’t currently using one, you can just as easily record these dates manually in a Google Sheet to calculate the difference.
As your team grows, any consistent increase in time to value (for example, quarter over quarter) is a sign you need to invest in a data catalog.
It’s common for these qualitative and quantitative signals to emerge or increase as your company and team grows. Naturally, demands for data will increase. In the early days, analysts will double as engineers (and vice versa).
However, when your team recognizes the signals mentioned above, you’ll need to divide and specialize into analyst tasks and engineering tasks. This is when you will start to need data governance and when you should start procuring a data catalog, as your next step after splitting data analysts and engineers will likely be creating the first data governance team.
Without a user-friendly data catalog, data governance teams would have to be significantly larger in an attempt to keep up with data demands. This is economically inefficient and not scalable.
Recognizing the inflection point of growth in data and analytics roles and the effects on time to value are tell-tale signs that it is time to formalize data governance efforts and procure a modern data catalog. Without this, organizations would have to invest excessive amounts of money on hiring to manually manage new data products — something that isn’t possible in the economic conditions faced in 2023.
Instead, investing in the right technologies early on in your data governance journey can ultimately save time and money down the road. Utilizing a next-gen catalog centralizes the management of governance rules but democratizes data discovery, leading to an efficient data governance program.
Ready to get a data catalog? Check out our ultimate guide to modern data catalogs and active metadata.
Defining the Vision
When starting your first data governance initiative, it is important to include key stakeholders in defining your purpose and vision. To do this, you need to have a deep understanding of what matters the most for your company, and also what challenges these stakeholders face without data governance.
Schedule multiple meetings with key leaders from executives, business leaders, operational teams, and data & analytics practitioners. Identify unique pain points and current processes using a value stream mapping framework. This exercise helps everyone align on what value could be achieved if effective data governance were in place.
As mentioned by Bill Schmarzo and Dr. Kirk Borne in The Economics of Data, Analytics and Digital Transformation, ‘The value of data isn’t in just having it (data-driven). The value of data is determined by how you use it to create new sources of value (value-driven)’. So, don’t be just data-driven, be value-driven!
Your vision and purpose should be clear and concise. It will serve as the North Star for what a great data governance program will look like for your organization.
Data governance covers many aspects in data and analytics such as data quality, data security, DataOps, master data management, and metadata management, among others. Your governance strategy will detail how each of these will be met, but your vision should communicate what success in each of these areas will deliver for the organization.
For example, a vision statement could look like:
Data governance at [your company name] enables democratized access to data assets that are trusted, well documented, and of the highest quality for a variety of stakeholders and use cases.
Communicate the vision consistently internally. Make it your purpose for everyone to know what data governance is and what is your purpose, vision and benefits once vision is accomplished. (More on this below.)
Crafting the Strategy
Once you define your purpose and vision, you will know how big of a mountain you are about to climb. There are no small mountains, so dividing your challenges in multiple smaller stages can help execute efficiently, keep stakeholders engaged, and maximize your chances of success.
The best way to do that is by creating a data governance maturity model. Consider that you are currently at Level 1, and call it something that resonates well with your current moment. Set a Level 5 name according to what you want to achieve if your data governance program is successful. Setting moonshot targets here is completely fine, as Level 5 will serve as motivation to make your internal governance engine move!
Define intermediate steps (i.e. Levels 2, 3 and 4) based on goals you need to achieve during your program. An example is below, but do customize it to your organization.
There are some basic pieces that your program should account for and anticipate at every level. Every level should contain (see below for an example):
- Milestones: Goals to accomplish during the rollout of every level.
- Action plan: Checklist of tasks to accomplish your goals in a specified timeframe.
- Scope: Which departments you will be covering on every level.
- Risks: As much as possible, anticipate the risks you will encounter during execution of every level.
- Expected results: Where your organization will be after the completion of every step. Expected results are a kind of transition criteria to the next level.
Securing Buy-in
At this point, you’ve taken the biggest step in your data governance program — getting started! Your vision and purpose are defined, your execution strategy is sound and clear, and your maturity goals are structured and modeled.
In order to bring the program to life, now it is time to get buy-in from decision makers, secure a budget, and identify the right teams for kicking off the program.
Take your time meeting those decision makers and carefully present your maturity model, emphasizing the expected results after completion of every level of your model. Your goal is to make these stakeholders believe data governance is the next step to skyrocket your company with data. To do this, be sure your expected results of the data governance program align with the business and/or operational goals of your stakeholders.
When your strategy is clear and people trust you to execute it, unlocking the budget to procure the necessary technology and hire the right people will become straightforward.
Launch!
There will be a moment when your first data governance piece of value is delivered. It could be implementing a data catalog or new KPI glossary, monitoring core data, defining ownership, etc.
Data governance deliverables, just like any other data product (data sets, tables, dashboards, etc.), are all about usage. Don’t forget to measure the adoption of everything you launch and, of course, measure the adoption of your data catalog using metrics like Weekly Active Users, Monthly Active Users, Product Stickiness, and Feature Usage.
Setting targets for these adoption metrics is key. It will help you identify where you are, where you want to be and what is still missing for you to get there. It will also help you better communicate with stakeholders on the progress you are making.
Overall, data governance is a key pillar to any data and analytics group’s success. To recognize the need for a dedicated data governance program, you must monitor the growth of the current data team, their challenges, and the impact on time to value for data consumers.
A clear vision for data governance in your specific organization is important, and it has to be accompanied by a strategy and roadmap that is easily understood by a variety of stakeholders. With maturity levels defined and a connection to business value established, it’s easier to get buy-in and budget for necessary technologies (such as data catalogs and/or data quality tools) and personnel from decision-makers.
Now it’s time to go out and launch your data governance program!
Coming soon: part 2 is all about sustaining engagement and growing your program!
To get that article in your inbox, subscribe to Metadata Weekly, Atlan’s newsletter with the latest links on active metadata, DataOps, and data culture.
Header photo: ThisisEngineering RAEng on Unsplash