Could "End User Buy-In And Support For Accurate Data" Resolve The "Garbage In, Garbage Out" Issues of Databases?

In a previous blog, I referenced a database privacy quotable "Garbage In, Garbage Out" to address not only the concerns for data privacy, but the accuracy of that data. Database accuracy is such a big concern because information about every aspect of our lives continues to be recorded and stored in infinite datawarehouses to be accessible by individuals who have the power to grant or deny us privileges. It is not enough that our ability to make decisions is no longer an autonomous process, but that which should be our sole right is now dictated by powerhouses of information that could be dangerously inaccurate. We see the effects of bad data maintained in businesses when individuals experience identity theft; a keeper of the data stored about us makes a clerical error; and or when a company never continuously validates the integrity or quality of data it stores.

I recall reading an excerpt of Database Nation (Simson Garfinkel) in which he discusses how outdated data is at a lot of the companies who continuously collect and store data about individuals. This old or bad data may be rarely updated or never validated in a lot of companies whose sole function is to share information for the purpose of conducting business. We know this to be true because of the credit reporting, billing, and mailing errors that many of us have to live with daily. How many times have you received someone else's mail? And how many times have you received calls at your residence for the same wrong person-wrong phone number even after the passage of multiple years? Someone or some company simply is not, properly, maintaining the data being stored. This failure to validate and update data is not only annoying, but can be detrimental if this bad data is used to make critical business decisions.

We recognize this problem obviously, but what can we do about it? How do we begin to tackle this problem when it seems like an infinite one that will take infinite lifetimes to resolve? When companies maintain bad data while simultaneously collecting and storing new data, how can they reasonably expect to achieve data accuracy and consistency? It is possible we may never solve this problem, but if we choose to tackle the problem then we might minimize the number of errors and their impacts. There are companies who have taken steps toward approaching a solution.

In the article that is referenced below, "The Secret to Successful Business Intelligence A Top Notch Data Warehouse" Rensselaer Polytechnic Institute seeks to gain end-user endorsement and support for achieving accurate data. An essential first step for the institution was to review how data was being defined, stored, and used by various entities within the institution. It appeared that data was being managed with no real guidelines in place. There was chaos in a sense because no "data definitions" were established (Daniel 1). Different departments and business functions used their own definitions and methods for looking at data (Daniel 1). This presented a lot of problems for Renassler as the following excerpt reveals.

"...Finally, the admissions staff needed more timely demographic information about its applicants to inform student selection decisions.

Getting a handle on the data has been critical because higher education today is a tough arena. Government funding is down, requests for financial aid are up and admitting a diverse student body—in terms of gender, geography, ethnicity and academic achievement—has become more challenging. All these factors make balancing the supply of enrollment acceptances and financial aid with the demand from student applicants more challenging than in the past. The better Rensselaer could optimize its administrative resources and time, the more revenue it would have for courses and scholarships to attract the best and the brightest.

The answer was a business intelligence and enterprise data warehouse implementation (Daniel 1)."

Ultimately, Renassler decided to "create enterprisewide processes for collecting and using data (Daniel 1)" which included communication, training, and support for end users (Daniel 1). They implemented this new process via the following focused steps as listed in the article:

1. Create cross-functional support.
2. Think big, start small, deliver quickly.
3. Create one version of data truth.
4. Provide support for new behaviors.


I like Renassler's approach to solving an ongoing business intelligence problem. It showed some maturity (e.g. as it applies to Capability Maturity Model® Integration) on the part of the institution to go from chaos to at least recognizing the problem and trying to find a valid solution. Also, I believe that their new approach of putting more value on the keepers of data by providing "broad user support (Daniel 2)" could serve as a best practices methodology for other companies seeking to improve the quality of their information they store. Renassler's redirection serves as a good best practices because "enterprise data warehouse and business intelligence projects' success depends on broad user support and because consequential business decisions are made on the faith that information is accurate (Daniel 2)."

If every datawarehouse infrastructure applies this ideal of continuous proces improvement (CPI) or total quality management (TQM) of data, then we may get closer to weeding out the garbage that could potentially go into databases thus minimizing the garbage that goes out of them as well. It looks like it did well for Renassler in terms of ROI and "optimized expenses (Daniel 4)." See the section titled "An A+ for Rensselaer's Business Intelligence" on page 4 of the article.


Referencing Article:
http://www.cio.com/article/151601/The_Secret_to_Successful_Business_Intelligence_A_Top_Notch_Data_Warehouse
"Outdated information and disagreement over data definitions was impeding Rensselaer Polytechnic Institute's progress. To the rescue: a business intelligence plan that emphasized end user buy-in and support for accurate data"