4 Reasons Why Metadata is Broken and Changing the way you think about It

Metadata is more important now than ever. While new technologies enable business users to work directly with data, the consumerization of IT means people expect systems to be intuitive and require little training. With so many people using data to support so many kinds of decisions, it’s critical that your data is described, defined and understood. Still, too many systems require a slow, rigid approach to metadata. By changing the way you think about metadata, you can make it faster and easier for the business to make sense of their data.

4 Reasons Why Metadata is Broken and Changing the way you think about It. Source: Tableau
4 Reasons Why Metadata is Broken and Changing the way you think about It. Source: Tableau

Read this article to find answers to four common reasons your metadata is broken.

Businesses are transitioning from traditional business intelligence solutions to more nimble solutions that enable end-users with analytics for decision support. Many IT managers wonder what the role of metadata is in such an environment.

The truth is that metadata is more important than ever. New technologies enable businesspeople who have traditionally not been analysts to work with data. The consumerization of IT means people expect systems to be intuitive and require little training. With so many people using data to support so many kinds of decisions, it’s critical that your data is described, defined and understood. That’s the role of metadata.

But too many systems still require a slow, rigid approach to metadata. This approach decreases the flexibility of a business intelligence solution and ultimately reduces the benefit you can get from deploying self-service analytics.

IT managers today face these issues when managing metadata. By changing the way you think about metadata, you can make it faster and easier for the business to make sense of their data.

Content Summary

Pre-defining metadata takes too much time and slows down a deployment.
Metadata isn’t as flexible as you need it to be.
Metadata isn’t discoverable in the flow of analysis.
Metadata slows users down rather than helping them.
Conclusion

Pre-defining metadata takes too much time and slows down a deployment.

In traditional business intelligence systems, organizations must model their entire enterprise as a first step. This is a time-consuming and complex process that pushes out an enterprise deployment by weeks or months. The start-up costs are high and the benefits of analytics are delayed.

A better approach is to look for a solution that can support analysis immediately. This not only has the benefit of delivering useful analytics more quickly, but also means that your metadata model can be built iteratively as you learn more about how people use the data. This more agile approach typically leads to a more robust and realistic metadata model.

A good way to get started is to leverage metadata from existing systems wherever available. For example, why take the time to define all your date fields as dates in your analytics solution when the database already has them defined that way?

Metadata isn’t as flexible as you need it to be.

Traditional metadata models are difficult and expensive to change. As a result, they don’t change often. This means they slowly fall behind in their ability to accurately represent the business data. IT may be responsible for keeping metadata up to date, but they may not have the information they need to respond to change.

In a world of fast-changing trends and opportunities, this is a severe competitive disadvantage for a business.

New definitions and calculations are necessary on a fairly regular basis. Business users who are asking and answering questions with data are often the best source of new metadata. They may create a hierarchy of category → product, or group territories into a region. A flexible analytics solution will allow IT a way to evaluate and then promote new metadata objects to production so they can be shared by all users.

If you lack flexibility in your metadata, you lack the ability to evolve your understanding of your business.

Metadata isn’t discoverable in the flow of analysis.

Another issue facing IT managers who want to enable end users is that metadata may not be discoverable. If your field names look like this:

Metadata may not be discoverable. Source: Tableau
Metadata may not be discoverable. Source: Tableau

and your business users have to hunt for the meaning of a field by searching in an intranet or referencing some document, you might not be getting the benefits of self service analytics. Users are blocked by their inability to understand what these fields mean—they need metadata. And when metadata is hard to find or difficult to access in the flow of analysis, users may give up. Then you’re not using data to improve your business and your end users are frustrated.

Look for a system where you can expose metadata to users easily, through readable field names and field descriptions. Make sure users can get that information when they need it.

Metadata slows users down rather than helping them.

Business users shouldn’t need to understand metadata to be successful. Metadata should help business users do their work, not stand in the way of an analysis that would have supported a better decision.

Consider the following scenario: a business user requests access to data and an analysis tool for a high-priority strategic project. Before that person can start to mash up data and analyze trends, he or she is required to take a training class on the organizations’ data and metadata. The class is scheduled infrequently and requires a good bit of time. This person can’t fit it into the schedule in the timeline he or she has for the project, and instead relies on an old analysis that doesn’t address the core questions.

What if, instead, the user could immediately get access to data, analyze without any special training, and discover the meaning of fields and calculations as needed? In addition, what if they could share this with their peers, organization or company in a secure way? This should be a requirement of any self-service analytics solution.

Conclusion

The heart of an enterprise-scale, self-service analytics solution is programmatic support for end users. Metadata is one of these programmatic elements and is a critical enabler of business users. IT managers who are striving to provide data discovery for their organization are uniquely in the position to manage shared data and metadata. Look to technologies that can support your need for flexible, fast and useful metadata. Your users will thank you.

Source: Tableau