Tag Archive for 'GIGO'

Avoid the pitfall of bad data.

In organizations where there are gigabytes or terabytes of data collected through the course of business, there exists a risk that, over time, some of the data can be bad, thus losing their business values.  This happens when the data cannot be consistently understood or interpreted to correctly represent the real world life.

Bad or unclean data creates confusion to the data users, making businesses lose insight on their performances. It also drives bad decision-making, causing damages or loss of profit to the business, and in the really bad scenarios, to customers.

(Read about GIGO: http://en.wikipedia.org/wiki/Garbage_In,_Garbage_Out)

The two keys reasons for bad data in an enterprise are loopholes within application/system design and incorrect usage by the end users.

In respect to Banner ERP, SunGard designed this system with SOME integrity in it. This ensures, to some extent, the business processes are adhered to, based on the normal work flow in most higher learning institutions. E.g. a student cannot register for a course before the registrar’s office creates it, or a classroom can’t be assigned twice for the same time slot in the same building. In the meantime, SunGard also left some flexibilities open for institutions to exercise their own policies and procedures. This leaves room for the second cause of bad data – incorrect usage.  Since there is not much Davenport staff (including IT) can do about the fundamental design of Banner, or any other purchased software the University is currently using, correct usage of these vendor products becomes the key factor for data integrity.

The following are some data quality suggestions that functional departments can implement

  1. Determine consistent and standardized data values within departments. This is the most important but also at-times ignored factor
  2. Validate data against the values developed from step 1 before entering into the system
  3. Enter/change data based on the result of step 2
  4. Developing training and knowledge transfer documentation to assist end-users with data validation/entry process
  5. When impact of data crosses to other departments, work with other departments and/or ITS to bridge departmental gap to ensure institution wide data quality
  6. Don’t try work-around to circumvent the established process/standards. If needed, re-evaluate the process and make process changes with consistency and integrity. When in doubt, seek advices from SunGard and/or ITS.

Kane Zhang