Top 14 Misconceptions About Master Data Management (MDM)

It is for enterprises and the requirements of enterprises cannot be met

Ideally MDM targets the whole enterprise, not just a single department, and must deliver applications that support multiple use cases while being capable of true enterprise‑wide deployment.

It is targeted at a single application, so it does not have to be a separate organizational procedure

MDM work should not be limited to a single application or domain; building a solid MDM foundation influences enterprise application delivery for the next decade, and tying MDM too tightly to an initial delivery hampers future development.

All of this is about ————

Fill the gaps with data quality, hierarchical management, merge/match processing, workflow/governance, real‑time integration, enterprise data models, etc. While any of these value propositions may spark a project, understanding the full range of MDM possibilities is essential.

Can I put master data in a data warehouse?

Yes, but batch‑processed data warehouses are too late in the data lifecycle for real‑time handling, and even real‑time warehouses often lack many MDM capabilities.

Most subject areas do not require workflow/governance

Many do not need complex workflow because they inherit standards from elsewhere, and MDM simply consolidates them into master data; however, even small workflows for authorization can add value.

No return on investment

Technically correct: unless MDM is part of a business application that delivers ROI, MDM itself is an investment. Projects that use MDM data become more efficient, reducing costs and improving functionality, and enterprise MDM lowers total ownership cost through reuse.

These projects all seem to fail

Failures attract attention, making MDM appear unlucky; if you get it wrong, projects fail. MDM needs business input to succeed and is not a pure IT project; successful MDM delivers high value across industries.

I start by selecting an MDM vendor

Begin with vendor‑neutral education and consulting; involve vendors later.

I can avoid data modeling with MDM

The data model is MDM’s most valuable component; everything you invest in the model yields multiple returns. Customize the packaged model and expect to invest effort in it.

To improve MDM data quality, the best way is to first find a data analysis tool and blindly run it on my data

If you enjoy meaningless work, do that. Data quality is custom; start by defining the rules the data should follow.

Data quality is vague and intangible

The most effective method is to score data quality and be very explicit about it.

All MDM tools are the same

All tools are not the same; some cannot perform all MDM tasks. Some provide extensive IP (models, reports, workflows) for specific domains, others do not. Decide what matters to you.

Organizational acceptance will take care of itself

Organizational acceptance is the hardest part; include change management in MDM implementation to achieve success.

Third‑party data is not suitable for MDM

Third‑party data usually extends important domain profiles that MDM governs. Introducing third‑party data can actually launch many MDM programs.

This article is part of IBM’s Mid‑Market Enterprise plan, which provides tools, expertise, and solutions for mid‑size companies to become smarter engines of the planet. My contribution is compensated, but the views expressed are my own and do not necessarily represent IBM’s position, strategy, or viewpoint.

Original source: https://www.mcknightcg.com/top-14-master-data-management-misconceptions/

Article URL: http://jiagoushi.pro/top-14-master-data-management-misconceptions

Discussion: Join the Knowledge Star “Chief Architect Circle” or the Feiliao group “Chief Architect Think Tank”.