Why Your Data Catalog Might Be a Road to Nowhere
Julia Bardmesser
Accelerate the Business Value of Your Data & Make it an Organizational Priority | ex-CDO advising CDOs at Data4Real | Keynote Speaker & Bestselling Author | Drove Data at Citi, Deutsche Bank, Voya and FINRA
Imagine you've built a magnificent highway. It's wide, smooth, and designed for all types of vehicles – from tiny smart cars to massive semi-trucks. Sounds great, right?
But there's a catch.
The on-ramp to this highway is so narrow that only the smallest cars can squeeze through. And once you're on the highway? There's no exit ramp at all.
Not very useful, is it? Moreover, isn’t it a little absurd??
This is exactly what often happens with many data catalog implementations.?
Let‘s break it down:
Most catalogs make it easy to import technical metadata. If you have well-structured data models with clear naming conventions, you're golden. That's your small car on the highway.
But here's where it gets tricky. The real value comes from connecting business metadata to technical metadata. That’s a large semi loaded with valuable cargo you want on your highway.?
This process often requires manual input from busy business stakeholders. While everyone agrees it's important, it rarely gets the attention it deserves.
Many organizations moving to Agile are making the matters worse.?
In the waterfall days, we had detailed requirements documents that often included data definitions. But with Agile, many companies have swung too far in the other direction (and completely missing the point of Agile, but that’s a different newsletter), documenting only the bare minimum in user stories.
The result? We're actually losing valuable information that could be used to create robust business definitions in our catalogs especially with newly developed AI capabilities.
领英推荐
2. The Last Mile Problem: Making the Catalog Useful
Let's say you've solved the first problem. You've got all your data – technical and business – neatly organized in your catalog. Great job! But we're not done yet.
The next challenge is making sure people actually use it. This is our missing exit ramp.
Most catalogs are implemented as standalone tools. If you haven't thought about how to connect them to existing business processes, they'll sit unused. It's like having a beautiful highway that no one can get off of – what's the point of the journey if you can't reach your destination?
We all know what happens to unused data capabilities – they wither and die.
So, what's the solution?
Remember, a data catalog isn't a "set it and forget it" tool. It needs constant attention and care to truly provide value. So carefully weigh all your options before making it the linchpin of your data roadmap.
A well-implemented data catalog can be a powerful asset, driving efficiency and smarter decision-making across your organization. But to get there, you need to ensure you're not just building a road to nowhere.
Every Friday morning, I'll email you 1 actionable tip to accelerate the business potential of your data & make it an organizational priority.
This is such an insightful analogy! It really highlights the importance of accessibility in data systems. What do you think are the main barriers that organizations face in making their data catalogs more user-friendly?
Data Governance Lead in CSIRO Digital Office
2 个月Olga Lysenko Rebecca Ostergaard
Data Strategy, Data Governance, Data Quality, MDM, Metadata Management, and Data Architecture
2 个月Great post, Julia! Unfortunately, many data catalog implementations fail. Here are a few additional reasons: 1) Companies often spend $200-500K on software but only allocate 2-3 people to implement it—usually not even full-time. So, what can you expect from the data catalog implementation in such situation? Having data catalog software doesn't mean the data will automatically be cataloged. Someone has to create cataloging processes, collect metadata, and ensure that metadata remains current. A few part-time people cannot accomplish this effectively. 2) Those assigned to the implementation are often junior developers with little to no training in data governance, data quality, or metadata management. Management rarely invests even 1% of the software’s cost in proper training, leaving developers to rely on ...drum roll ... vendor-driven knowledge. This is the biggest issue. 3) Implementation is frequently outsourced to large offshore contractors. These firms prioritize billable hours over investing in the education or expertise needed for a successful deployment.
Strategy | Marketing | Sales | Brand | AI Fluent | Semi-professional Salsa Dancer ??
2 个月Perfect analogy Julia. ??
Enterprise Data Architect @ ABC Supply Co. Inc. | Leading Data Strategy, Data Governance, Data Accountability
2 个月Technology often likes to build tools then tries to figure out how to derive value. Just like any tool or solution, the purpose and value of a capability should be defined and agreed to up front in a business case. With this in mind, value can be prioritized and solution usability is available immediately, then grows with continuous refinement.