2. DATA GOVERNANCE MISSION
ESSENTIALS
** No matter the driving initiative behind a data governance effort there
needs to be a three part process which includes:
1. Rule Design
2. Issue Resolution
3. Monitoring and Enforcement (of rules)
** FROM THE DATA GOVERNANCE INSTITUTE “BASICS”
HTTP://WWW.DATAGOVERNANCE.COM/ADG_DATA_GOVERNANCE_BASICS.HTML
3. RULE DESIGN
In my opinion there are two ways to develop governance rules.
1. Top Down
2. Bottom Up
The “top down” approach is one where rules are developed from the stand point of what data
should look like. These are idealistic rules in nature because they are heavily rooted in strict
adherence to business rules. In other words, they do not accommodate the legitimate
exceptions to rules. Also, these type of rules tend to turn less than perfect data into a violation
and produce voluminous exception reports.
The “bottom up” approach is one where rules are developed from extensive analysis of the
data, i.e. data profiling (a fundamental activity in data quality). This approach tends to be more
practical in that it deals directly with data related issues, rather than business rules, such as
common data entry and/or migration issues.
From my experience the “bottom up” approach tends to be more efficient and produce more
rules that improve the data over time. Another benefit of the “bottom up” approach is that data
profiling provides hard numbers on the occurrence of a specific violation. As a consequence,
profiling allows the governance committee to prioritize rule enforcement by exception volume
and create the most benefit for the effort.
4. ISSUE RESOLUTION
IR is one of those phases where it is not obvious what role data quality plays until
you use quality practices in the issue resolution process. Here are three ways to use
data quality in governance issue resolution:
1. Reviewing data profiles can provide hard facts around data value distributions
which can in turn provide direction on how to correct, or resolve, a data issue
2. Reviewing data quality scorecards can provide details around the volume of an
exception which can in turn provide direction on whether an issue resolution will
return a high yield on the remediation investment
3. Typically profiles and scorecards contain a source component or identifier. This
identifier can enable the governance committee to decide on who the
accountable parties are for developing the issue resolution
5. MONITORING AND
ENFORCEMENTSimply put … there is no rule monitoring or enforcement without data quality!
M&E is essentially a before and after process where the cycle in the
illustration below is continuously reviewed by the data governance
committee
Transform
rules into
scorecard
metrics
Examine
metric
exception
trends, volume
s and
instances
Build data
remediation
jobs to rectify
the exceptions
1. Transform data governance rules into scorecard metrics. This
enables monitoring and measuring data exceptions. At times this
can be straight forward because some rules are heavily data
intensive and can be traced to a data element. Other times this
process involves the decomposition of a business rule to it’s core
data element in order to transform.
2. Examining scorecard trends (is the data getting better or worse?),
volumes (how big is the issue?), and specific exception instances
(how is the rule being violated?). This is what data governance
committees should be talking about at almost every meeting!!
3. Building the data remediation processes is the post violation
enforcement of rules and a key component in the value the data
governance promises. Typically, this involves correcting the
exception to the defined and accepted value.
6. SUMMARY
o Data quality should play an active role in the core of any data governance
program
o Data profiling can provide direction on data issues, their volume
and, ultimately, the priority of data governance focus
o Data quality scorecards are the measuring and monitoring component of a
data governance program that provide clear evidence for governance
effectiveness
o Data quality remediation is the fundamental mechanism for turning data
governance issues into success stories