Extreme Information Management

Extreme Information Management

A “live blog” of Mark Beyer’s “Extreme Information Management” session at the Gartner MDM Summit 2012

I’m attending the Gartner MDM Summit this week, and one of the first sessions I dropped in on was Mark Beyer’s “Extreme Information Management”.

Mark started out by talking about the extreme information challenges for the 21st century CIOs, saying it is no longer possible for one person or group to control how an idea like “big data” is used. But Gartner thinks the term Extreme Information Management provides more clarity.

The human brain is a purpose built processor, tuned to recognize threats more than anything else. One of the biggest threats today is stress due to information overload. He quoted Thomas Edison saying “I have not failed, I’ve just learned 10,000 ways that do not work”.

So today’s session is about mitigating the stress that comes from extreme information.

Through 2015, organizations integrating new information types will outperform their industry peers financially by more than 20%.

Unstructured data comprises 60-80% of all enterprise data. Anecdotally, Gartner is seeing 5-15% of analytic efforts starting to use some form of sentiment or social analysis to augment structured data analytics.

Information is exploding, growing at 59% per year. The new mantra is “include all the information assets, deal with the volume, and derive the business value”.

And gut feel probably differs from an answer derived from analytics 79% of the time.

Leaders are able to create new data points to prove that they’re more effective than the competition.

Extreme information management is the concept that your current information infrastructure must be intentionally managed along 12 complementary dimensions to meet the challenges of the 21st century Information Age.

Mark pointed the audience to a research note on Gartner.com: “Big Data is Only The Beginning of 21st Century Extreme Information Management”.

If volume is increasing at the same time as variety, that is a “big data” problem. Does the create case match the use case? If the business reports that volume is an issue, then ask if the variety and the velocity are changing as well.

The 12 dimensions are:

Big Data















    Pervasive Use

Qualifying the Information








You will always be stretching out on these 12 dimensional aspects – big data initiatives are never finished, and Fidelity is the culmination of the other 11 dimensions.

Traditional application data growth is related to 87% of performance issues, costs are 3-10 times more to store data as to capture it, and this represents 15% of your problem.

The variety of types of information assets is crushing.

Operational technology – small apps, millions of devices, leads to billions of instances.

Governance models in too many places can challenge the contracts around the use of data. Profiling the data in an ERP system allows the IT organization to work proactively with the business, to look for changes in the data that might mean changes in the business process. It’s very common to see four columns crammed into one (overloaded attributes).

There are a lot of opportunities in social media.

A network of related data is far more reliable than any one datapoint, especially if you know when, why and how it was related.

The “cloud” aspect of big data allows you to get more power, more data, and use an alternative deployment.

Big Data MapReduce Hadoop


    Introduce a metadata management program


    Plan on extending existing repository management environment to use cloud storage


    Build at least one pilot analysis that combines text, social and operational data

I found Mark’s session very thought provoking and intellectually stimulating, and it helped me to think about where today’s clients who are doing MDM and data governance might need to go next.

Tags: , , , , , , ,

Comments are closed.

%d bloggers like this: