Maintaining Data Quality

Have you been underestimating data quality issue?• Mismanaged data
Data quality is often addressed in the form of• Inadequate data acquisition process
employees or vendor-supplied consultants,• Multiple data silos maintaining the same data
accounting for 20-50% of the data warehouse• Redundant data across multiple channels
projects, labor over at least a few weeks and, in• Ineffective data update process
some cases, several months, depending onIt is important to identify the weak sources within
project size and complexity. Data, most often,an enterprise from where erroneous data flows.
loses quality when integrated. TraditionalOver the years, enterprises tend to maintain data
integration techniques most often fail as datasilos that keep expanding with faulty records. Due
flows from multiple sources (built on differentto disjointed network and insufficient integration
platforms) and in different formats. All datacapability, the data is further contaminated with
sources might not have effective informationerroneous information. Inaccurate, redundant and
sharing mechanisms, which primarily makesmissing data continues to grow till it becomes an
integrated data unreliable.unmanageable set of fragmented and voluminous
Poor data quality is known to damage milliondatabases. Different systems are programmed to
dollars for any enterprise. Spending onaccess data from the different faulty databases.
implementing large CRM, BI or integration projectsCRM, ERP, BI and integration technologies that fail
are a waste till the quality of data flowing toin enterprises are most often the result of
these systems remains low. In fact in the longsubstandard data. The data quality metrics are
run, bad data can lead to ‘low customeronly the roadmap to improve the quality of data.
satisfaction’ and decreased customerEstablishing a process to monitor results and set
retention. The three aspects critical to data areimprovement goals is the next vital phase of data
accuracy, consistency and timeliness. High qualityquality lifecycle. Cost, effort and time spent on
data is dependent on these three standards.the data quality process should be able to provide
Inaccurate data is junk for any enterprise, thus,business benefits, accordingly. The improvement
accuracy is important. Though all departments ofprocess is a set of business cases and specific
an enterprise need data for different purposes, itareas that need to be addressed. With the
is crucial for the entire enterprise to haveimprovement areas, the goals and the priority
consistent data. In 86% cases, low customerattached with each should be clearly identified.
satisfaction is the result of obsolete customerPersonnel responsible for executing each
data that may exist in several departments asimprovement task should clearly understand the
redundant customer records. Duplicate customerentire execution process. The improvement
records also increase the volume of databases.process can be taken as the yardstick to
Data at the right time, for the right peoplemeasure future data quality improvement.
determines the operational efficiency of anyAccording to Gartner, through 2005, IT-driven
enterprise. It is the right data that becomes thedata quality efforts will be largely ineffective in
basis for taking operational, tactical and strategicachieving improvement goals. For improvement
decisions.efforts to succeed, senior management must
The main factors that render low quality to datarecognize and adopt data quality as a key
are:business priority.