Separate the Forest from the Trees with Enterprise Data Management

2008 Olympics Dashboard
August 13, 2008
Data Integration Trends: Future Shock
August 28, 2008
Show all

Separate the Forest from the Trees with Enterprise Data Management

published in Information ManagementThe “single version of the truth.” How many times have you seen that catchphrase in presentations by industry analysts, vendors, consultants or even your own IT group? How many times has it been used by industry experts writing their articles, columns or blogs? And it keeps being used for different solutions, technologies and products! Each new technology and silver bullet promises to deliver the single version of the truth, this time for sure.

The phrase started with data warehousing in the early 1990s. Then proponents of enterprise resource planning (ERP) systems used it during the last decade to encourage everyone to replace their legacy systems. Now master data management (MDM), customer data integration (CDI), operational BI and service-oriented architecture (SOA) are the latest applications to which this magic phrase is being applied.

Everyone is talking about it, but it still isn’t happening. Corporations that are just beginning to add MDM, CDI, SOA and operational BI to their ranks still have an entrenched and ever-expanding population of data silos. The experts tell them that the new solutions will deliver the single version of the truth. But the reality behind the hype is that those silos are still there and aren’t going away on their own.

Think Like an Engineer

I’m sure that most people who use the catchphrase “Think like an engineer,” are sincere and believe that this time it will be different. But sincerity is not enough.

We continue to build silos because we don’t think like engineers or architects when we examine data in an enterprise. For example; first, we look at all the pieces individually, rather than looking at the whole. We especially don’t look at how all the pieces fit together and interact.

Second, rather than looking at how something works, i.e., the underlying processes and interactions between processes, we characterize a solution by the product or technology used. All the limitations and constraints of that product are then associated with the solution. For example, rather than integrating their data, companies performed extract, transform and load (ETL) because that’s what the first round of data integration tools did. Enterprise application integration (EAI), enterprise information integration (EII) and SOA are considered totally separate because they work differently; at least they appear to if you do not examine the underlying processes.

Finally, we start off by stating how the newest solution is different. We in the high-tech industry are mentally conditioned to think that we are introducing a revolutionary product or idea rather than an evolutionary one. Vendors need industry buzz to sell their products, but expanding on past products or ideas doesn’t generate buzz. If you implement the “new” approach with the new products, they tell you, your data problem will be solved.

See the Forest

Working with MDM, CDI, operational BI and SOA is easy, relatively speaking. The hard part is addressing the intimidating, underlying issues with our data. Products alone cannot complete the data integration puzzle; people and process are crucial. The other missing puzzle piece is failing to envision a holistic architecture that covers everything – people, process, data and products.

That missing piece is enterprise data management (EDM). EDM encompasses managing data from its creation until its consumption as information. This management includes people and process, along with data and products. Data governance jointly administered by business and IT is key. EDM includes the information, data, technology and product architectural components. Enterprise data integration in all its forms (ETL, EII, EAI, SOA) and all its applications (data warehousing, operational BI, MDM, CDI, etc.) are all part of the whole rather than piece parts used to build more data silos.

Plan for Success

The biggest hurdle with EDM is that it is perceived as being too big and daunting. That’s why I don’t suggest taking on an EDM implementation all at once. If an initiative is too big, it’s apt to fail, so start with a plan and roadmap. Just as you use a blueprint for building a house, use one to build and manage your data – a fairly significant corporate asset.

When planning a house, you tell the architect what style you want, number of rooms, location and so on. The architect translates your wish list into a detailed blueprint that shows exactly how it will all come together, down to where all the pieces of the frame, wiring and pipes go through the house. With all the resources and money spent on data, it’s time for our industry to start building a holistic data blueprint just as an architect does for a house. We need to stop pitching new solutions destined to be the next data silos and start advocating EDM.

(For more on this topic, see the series of blog posts on enterprise data management on The Data Doghouse.)

Leave a Reply

Your email address will not be published. Required fields are marked *