Anthem Consulting, LLC

One of the deliverables of a successful data quality effort is a current-state assessment of the organization's data. Without this initial step, there is no baseline against which to measure on-going efforts. New systems or applications do not fix the underlying quality issues.

Accountants cannot determine the financial state of a business without first conducting an objective audit. The information accounting is an objective view of the state of the data assets of a business. Objective measurements of data characteristics, such as missing values and frequencies, distributions, integrity constraints, etc. are taken. These measurements are then weighted and prioritized by business analysts and experts to arrive at a current data state.

Once the information accounting is completed, the organization can develop a plan of action based on the priorities and needs identified. Depending on the type of project, the information accounting can be used to

• Value the data resources that are within the scope of the project
• Develop project data models that are the basis for database design
• Construct data cleansing or standardization routines for a transformation process

Anthem Consulting offers three levels of information accounting.

Basic

Consists of evaluating data sources for standard concerns and issues such as, but not limited to, presence of NULLs, blanks, zeros, unusual values, and frequencies, primary key validations, foreign key validations, and simple orphan analysis. Business and technical definitions of structures are recorded. Initial quality scores are developed, based on priorities/weights of problems and/or structures. Brief and infrequent business analyst or business user knowledge needed.

Standard

Consists of further validating organization-specific characteristics or rules concerning the data. Things such as detailed business rule validation, complex orphan/duplication analysis across systems, prioritization of quality issues, and detailed scoring of data sources are accomplished. Detailed metadata, such as source file growth rates, update percentages, enumerated domains and ranges, is captured. Significant Subject Matter Expert time and effort is required, since business decisions are being made.

Advanced

Consists of project or organization-specific cleansing routines, tranformation specifications, and mappings to fixed or designed target data structures. On-going, repeated evaluation of data sources, on a monthly basis, based on previous information accounting efforts. Trending reports on data quality, and data standard compliance are delivered.