Building Trust
Building
trust into data means continuously monitoring and improving data quality in
terms of accuracy, completeness, validity and consistency. Trust in data is
demonstrated through competence in one’s approach to data stewardship.
Stewardship is more than simple oversight by a business owner -- it is a
complete lifecycle plan joining activities from business, development and
testing.
The Path to Data Trust:
Understand – Your Data
Integrity Needs
Understanding what is quality data and the consequences of losing data
integrity is the first step to data trust. From this understanding, develop your
data quality goals along with measurable objectives. SMART* objectives is a good
method to obtain your goals. (*Specific,
measurable, achievable, realistic and timed)
Innovate – Data
Stewardship and Quality
Create an innovative approach that tightly joins data owners, development and
testing. Most QA processes lack the tight collaboration across the different
disciplines required for data testing. However, innovation does not mean
wholesale QA process replacement. Rather, develop a creative and effective
trust approach specific to data integrity that enhances data stewardship and
testing processes.
Stabilize – Data
Testing Approach
It is inevitable that things don’t go as planned, so allow time for your trust
approach to work and make small changes as needed. Use incremental steps that
concentrate on the important integrity objectives first. Whether it is new
automation technology or a higher degree of collaboration, let your team
practice and get used to testing data.
Optimize
– Data Trust Processes
Once elements
of the process are stabilized, take steps to improve your process. Optimize the
existing activities while increasing the capabilities and coverage for verifying
data integrity. Also, consider how ongoing validation is achieved or how to
migrate processes to new areas of integration.
Check – Measure Data
Quality and Act
The keystone
to any data trust process is the ability to continuously validate. For data
trust, it is an absolute that integrity must be continuously validated. It is
difficult to have 100% test coverage and far too easy to implement untested
changes into data integration. It is inevitable that data deficiencies will be
detected, whether through ongoing regression testing or field reported defects.
Use these deficiencies as an opportunity to strengthen your process, not tear it
down.
Key Steps to Improve Data
Quality:
-
Identify the potential sources for bad data within your organization
-
Establish goals with measurable data quality objectives
- Create
data integrity controls as a matter of policy
- Ensure
that quality tests live beyond the project of the day
- Build
proactive defensives to faulty data
- Define
more specific collaboration points with all parties
-
Automate data testing for ongoing integrity verification
-
Build-in continuous quality improvement process
- Invest
in the quality of your staff
-
Cultivate quality attitudes and behaviors
It is
important to build a workable process improvement behavior. Process, polices and
procedures alone do not change behaviors especially when it comes to improving
data quality. There must also be a change in attitude and human behavior. Change
is best accomplished through practice; not policy.
|