Being a perfectionist by nature, I have always found it difficult to aim for less than perfect. While that might sound like an ideal answer to the question “What is your worst quality” during a job interview, I must admit that in the data & analytics space, perfectionism is not a superpower. 

Data governance is a step-by-step journey

When I started out as an analyst, I wanted each report I built to be perfect. That meant clear titles and correct data, but also perfectly aligned visualizations, carefully chosen color schemes and reviewing report labels until their font size was optimal. This approach leads to great results and very happy users when you have to build a report. Not very much so if you need to build more than a hundred before the project goes live. While of course through experience I became more productive, I also learned that in order to move forward in any data initiative, you need to focus on the core added value. n my case: was there a need for more perfectly designed reports in a nearly overdue project or was it about offering business users the full spectrum of data insights they needed to do their jobs in a timely manner?

Advising organizations in implementing our metadata management solution, I often see data governance and IT leaders sharing my character trait. They set the bar high at the start, but in doing so miss out on quick wins and increase the risk of undue delays and loss of enthusiasm from eager users.

When thinking about the available documentation templates, they tend to add too many information fields, trying to cover all potential needs up-front. When thinking about their governance model, they try to foresee several detailed roles rather than start with achieving the most vital one first: data ownership. When planning content creation in the glossary, they wish to have 100% of their data definitions validated before considering exposing them to a broader audience. While setting the bar high is admirable, it’s important to remember perfectionism comes at a cost. This opportunity cost might be time, money, missing out on learning opportunities or all of these.

What then, should ambitious data leaders aim for? That same high bar they are used to, but using a different approach. In my case, rather than obsessing about perfectly designed visualizations, I should have prioritized useful reports, foreseeing time to turn them from good-looking to great-looking after users were able to start using them. In the case of implementing metadata management, that means getting started with straightforward templates focused on vital info. It means starting with a simple governance model, extending it as needed as you grow. It means unlocking your 20% most important content first, before worrying about the other 80%.

Our latest customer went live with a fully operational report catalog within a month

The beauty of this approach is that in the end, not only will you notice you are able to achieve that same high bar you were aiming for, you’ll be able to do it faster. Our latest customer went live with a fully operational report catalog within a month, including over 200 reports clearly documented and organized by validation status, sensitivity, domain and organization. This customer did not set out to leverage the full capabilities of our metadata solution all at once. For example, they are not yet focusing on the business glossary to harmonize data definitions. Neither did they aim to fully document 100% of the cataloged reports, rather focusing on the 200 most vital ones. But they did achieve what they set out to do, offering a fully operational report catalog to all users across the organization. In less than a month, they succeeded in making their reports much easier to access and understand. They now plan to continue this journey by starting to catalog and document their key data warehouse models.

This blog post is not a call for lowering your expectations nor the goals you wish to achieve (that’s not in my nature either), but it does make a case for achieving progress step by step. If you are planning to implement a metadata solution, in your first iteration don’t try to perfectly design each template or cover 100% of all your users’ needs at once. Think carefully about the most valuable use cases and prioritize that portion of the content that will offer the most value to the most important group of users. Once you’ve done that: learn, reflect and take the next step.

Do that, and that high bar you are aiming for will transform from an elusive and ever-delayed finish line into a snowflake starting a downhill journey, allowing you to build momentum and realize growing value for your users over time.  

This insight was shared by dScribe co-founder Pieter Delaere on February 22, 2022. If you are a data governance manager, make sure you visit this page too.