Subscribe to Dr. Granville's Weekly Digest

Data Silos Obstruct Quest for Competitive Advantage

Data and information silos are a significant problem for organizations getting full value from data. Data silos are separate databases or data files that are not part of an organization's enterprise-wide data administration. An information silo is where parts of the information management system is unable to freely communicate with other information management systems. A siloed application is an application that does not interact with other applications or information systems.

The data science revolution depends on collecting, storing, analyzing and distributing massive volumes and varieties of data to turn into knowledge and valuable, actionable insights. The real value is from mixing different internal and external data sources and sharing information within a culture of collaboration. 

Strong evidence suggests that organizations utilizing a variety of both internal and external data sources - in conjunction with data science and business analytics - outperform firms that only rely on internal data and have data silos that prevent data science practice, information sharing and collaboration.

Unfortunately, managers have strong incentives to silo information to maintain power. As a result, organization leadership must have an information management strategy and policy of sharing information to break the deadly "data silo" status quo.

Data and information silos often exist because managers control the flow of information and access to the silo, and they perceive (1) their power and careers depend on information control; (2) there is not enough benefit from sharing information; (3) information might not be useful to folks in other systems; (4) costs to integrating the information systems is not justified.

In addition, data silos are a danger to data integrity - increasing the risk that current (or more recent) data will accidentally get overwritten with outdated (or less recent) data. When two or more silos exist for the same data, their contents might differ, creating confusion as to which repository represents the most legitimate or up-to-date version.
.
According to a recent CompTIA survey:

  • Eight (8) in ten (10) organizations report high or moderate degrees of data silos - collections of data that have grown across the organization or within specific departments which are not connected in a cohesive plan.
  • Only thirty-one percent (31%) report being able to provide a comprehensive single customer view (SCV), meaning potentially valuable customer data points are not fully integrated.
  • Two-thirds of organizations report definitely or probably experiencing some degree of shadow (or rogue) data repositories, whereby staff may be maintaining their own contact or prospect list.
.
To fully benefit from the data science revolution, organizations must change the incentive structure for managers to hoard and control information and prevent data and information silos from impeding the quest for competitive advantage.
.

Views: 493

Tags: Collaboration, Data, Information, Practice, Science, Sharing, Silos

Comment

You need to be a member of Data Science Central to add comments!

Join Data Science Central

Comment by Tony Maull on February 18, 2014 at 2:33pm

Doug - I don't agree entirely.  Let's look at customer information.  

Silos appear within product groups, between partners and between functions even under such things as loyalty programs - call center and email.  Sometimes it can be traced back to not creating a single view of the customer since you can't treat the customer across products.  

But, I believe, most information sharing doesn't occur because of a lack of an incentive.  "What in it for me to educate you about my customer interactions"?  In a bank, sharing home-equity-line-of-credit customer information could steal customers from the same bank's credit card product.  

Sharing customer data between channel partners is even worse.  It represents pricing power and access to buyers which means access to trade support and rebates.  Sharing information means a change in sharing money.  

If you want to break down silos, you have to dig into the business reason why the silo is there and create a more powerful reason for the two of you to tear it down.  

Comment by Douglas Dame on February 14, 2014 at 10:09am

Silos are indeed a problem, but IMO you have a totally wrong read on the actual nature (cause) of the problem. People generally do not silo data with the intent to maintain power.

USING data is all good. Whereas OWNING and MAINTAINING data is just a pain and an expense, and additionally makes you accountable for any and all errors and holes in the data. Therefore, as a manager, if you and your people can do the first without the burden of the second, that is by far the best place to be. 

People silo data ... and create their own guerilla reporting/analytics ... because the main IT shop lacks the interest and/or resources to satisfy all the needs. Anytime the word "ENTERPRISE" is used heavily, that is usually code-speak for "these are the things deemed to be really important to many people [and that IT/BI is working on], and anything else by definition is less important [and IT/BI is not working on.]" 

Operating units have ongoing information needs. If the central IT authorities don't satisfy those needs, then managers with money will start building their own stuff to fill the void. Improvise, adapt, overcome. It's human nature/organizational behavior 101, and entirely predictable.

Guerilla data silos do not "imped[e] the quest for competitive advantage." Quite to the contrary, they are a symptom of parts of the organization trying to achieve competitive advantage despite the impediment of a lack of support from central IT for their activities and needs.

* * *

An under-appreciated aspects of silos is that many organizations have silos of subject matter expertise about their data, even when the data itself is centrally organized and maintained. That also keeps the organization from maximizing the use of data assets. There is often a lack of redundancy for knowledge and skills on the people side of the mgmt system. The hardware side of the system will usually be designed with a "no single point of failure" mentality; the wetware side, not so much.

d.d.

Follow Us

Videos

  • Add Videos
  • View All

© 2014   Data Science Central

Badges  |  Report an Issue  |  Terms of Service