WHAT IS A DATA QUALITY SOLUTION? 

A true Data Quality solution takes into consideration all aspects of data management in an organization. This involves the entire "data flow", from collection to final utilization.  Every step in the process is an area that can either enhance data quality, or detract from it. 

Even though a Data Quality solution may start as a single "project"; it is in reality a never ending process.  As data is accumulated and utilized  - the tools, systems and knowledge need to be in place to accommodate changing business processes, client expectations and source systems.  The need for consistent monitoring of data quality becomes apparent once one considers the increasing rate at which data is collected, and the increasing value placed upon that data by its end-users.

KEY COMPONENTS OF A SOLUTION

LAUNCH A DQ INITIATIVE:

Gain board level support. Many times the IT department has authority to manage data, but not change business process or behavior.

Set initial objectives and goals.

Determine the gaps between desired process, procedure and quality standards and actual operations at present.

DEVELOP A PROJECT PLAN:

Determine which key data elements will be measured for validity and then cleaned and monitored.

Set specific goals (i.e. less than 1% duplication rate).

Determine ROI for project.

Create action plan.

Build systems to continually monitor DQ levels.

DEFINE ROLES AND RESPONSIBILITIES:

Key functions to support successful implementation of DQ  need to be identified and staffed appropriately.

REVIEW PROCESSES:

Identify and review all business processes for collecting, recording and using data in the organization.

Identify who owns the data and who uses the data.

Determine which processes require modification and suggest metrics for measuring.

Review all technical infrastructure to determine where systems can be modified or replaced to improve DQ.

ASSESS CURRENT DQ:

Identify common data defects (i.e. missing, incorrect, duplicate).

Domain Profiling to identify business rule violations.

Create metrics to detect defects as they enter the system.

Work with subject matter experts who understand both the business and data.

Define cleansing rules and application points.

Develop metrics for measuring DQ on continual basis.

CLEANSE THE DATA:

Fix as close to the source as possible; prevention is least costly method.

Repair takes place when clients have already been affected.

Do not assume all elements should be cleaned as business value may not be there.

Determine cost / benefit ratio as it relates to cleansing target metrics.

MONITOR DATA:

Develop program to audit data on a regular basis.

Develop service level agreements that specify tolerances for data elements and rewards / penalties for exceeding / missing.

IMPROVE BUSINESS PROCESSES:

Attitudes may need to be changed.

Corporate stewardship and firm-wide commitment to DQ is critical.

Educate, train and reward end users.

Agreement by senior managers on codes, rules and definitions.

 back to top