The Persistence of Silos: Why Do We Keep Building Them?

Loraine Lawson

"Already I prophesied to my countrymen all their disasters. Ever since that fault I could persuade no on of aught."

- Cassandra in "Agamemnon" by Aeschylus, line 1203


Not to sound all Oprah-like, but I'm trying to make my peace with silos. As someone who focuses on integration, this is no easy task. I feel like Cassandra, issuing warnings no one will heed.


Just now, I was reading TDWI's just-published, free report, "Next Generation Master Data Management," and what should I see but this little depressing item:

44% We practice MDM, but in silos per department, data domain, or application.

This is out of 369 organizations surveyed, close to half developed MDM - an approach predicated on integration and dissolving data silos - in silos.


But that's not all. Even though many organizations want to push MDM hubs to the next level and make it a truly enterprise-wide resource, they're struggling to do so thanks in no small part to a lack of cross-functional cooperation. That ranked third in the top five challenges to enterprise-wide MDM, followed by two issues that are actually related to it: a lack of executive sponsorship and a lack of data governance/stewardship.


It's like we just can't learn. Silos created the very problem MDM is supposed to solve, and what do we do? We build MDM in silos.


And it's not just MDM. It's just about everything that happens, including things that would seem innately anti-silo, such as data governance.


I find it very frustrating. But I'm also getting older, and learning to accept the things I cannot change, as they say. Recently, I've started to wonder if silos are something we cannot change.


Oh, sure, there are integration projects that succeed in de-siloing data, applications and systems. It's not that we have to wave the white flag and just give up on integration.


But maybe silos are just a natural, practical, dare I say reasonable, way to incubate new programs, new projects, new approaches and new applications.


Even data people - who should know better - do this.


Dylan Jones, founder of Data Quality Pro and Data Migration Pro, recently pondered whether organizations are creating silos or islands of expertise, particularly when it comes to information management.


John Owens, who created the Integrated Modelling Method, originally voiced this concern in a LinkedIn data quality forum.


"John believes that in many ways we've gone backwards," writes Jones. "We often have limited unification between teams involved in information management. Business modeling and data modeling are rarely carried out correctly and many organizations have no functional models or logical models that reflect the business they're actually in."


Jones adds that recent visits to organizations within both the private and public sector show it's a real concern. He found out-of-date physical data schemas, no logical modeling and no target models to guide their business environment's evolution.


"But is this a blip?" he asks.


Given the persistence of silos, I'm inclined to think not. But am I just getting cynical from writing about integration? Have I just seen one too many stories warning about new silos and the lack of enterprise-wide strategies?


I'd love to hear your thoughts. You can post below, or send me a message on Twitter or LinkedIn. (And, yes, I'm aware of the irony of asking for responses in siloed applications.)

More from Our Network
Add Comment      Leave a comment on this blog post
Apr 5, 2012 9:41 AM William Sharp William Sharp  says:

I'd say there are a few major reasons for the continuation of data silos ...

1. Lack of enterprise data strategy vision

2. Independent ownership of budget for development

3. Cost

4. Speed

Basically summed like this ...

I didn't know there was a corporate policy or strategy around this type of data, I had the money to build exactly what I wanted and it was cheaper and faster for me  to move forward with my project

Apr 5, 2012 11:39 AM Deb Chase Deb Chase  says:

Great article! Having been a marketing strategy consultant for many years in the hi-tech industry, I'd have to agree that data silos will always be at play. IMO, this is mainly due to human nature and our desire to "protect" and control what gives us a competitive advantage (islands of expertise) within the organization (i.e. if I know more than the next guy and prevent him from knowing it I am perceived as more valuable)

For those who don't want to compete but simply make educated, data-driven decisions, it becomes a monumental task to obtain, clean, map, and analyze data because it is 'locked up' in legacy systems (with different data dictionaries) that don't talk to each other...and that means getting permission from each data store owner to use the data, then working with IT staff to actually get the data. There is a lot of time and level of effort required to do all of these.

That said, I am encouraged by the move to cloud services and MDM efforts that might help companies actually USE their mounds of widely dispersed data and more user-friendly BI tools to make good decisions that positively impact their businesses.

Apr 10, 2012 9:35 AM david k waltz david k waltz  says:


Like you, I have been a persitent critic of silos, and working in organizations that are full of them can get very frustrating.

I attended a conference last fall where someone mentioned they were good, which spurred me to think about this a little more and ultimately wound up being a post on my Treasury Cafe blog called "The Matrix". They do have some advantages, primarily in terms of professional knowledge and development.

I think you are correct that silos will persist. Below is the link to that post if you are interested.


Post a comment





(Maximum characters: 1200). You have 1200 characters left.



Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.