Showing posts with label data. Show all posts
Showing posts with label data. Show all posts

Wednesday, June 29, 2011

Reliability Of Shares

Shares of reliability, often initiated by the trades and in consultation with the project owners, are intended to identify areas of non-quality, identify the causes and identify pragmatic ways to mitigate or delete.

The first actions are almost always in the form of manual corrections. These projects mobilize substantial charges to align the repository with the reality-duplication, enhancement or correction of signs, etc ... In addition, these actions if they can have a satisfactory short-term, must be renewed frequently to maintain quality and fight against the progressive drift.

In the long run, it is best to think of more sustainable solutions and, in this regard, the levers are very diverse appointment of quality managers in the contributing entities, workflow validation of customer data, standardization of concepts , comparisons with external sources, management of several criteria of uniqueness ... The possible solutions are many but their cost, period of implementation and impact on data quality is variable. However, the prioritization of these actions is often subjective: some effects more "visible" are given priority while others are ignored because their quality impacts are unknown.

The success of the business plan requires a continuous improvement in performance. It should be laid down in processes that involve the third party repository, a real cornerstone of the IF bank. Based on the quality approaches used in industry (six-sigma, Total Quality Management ...), the virtuous cycle is divided into five phases: definition of indicators and quality objectives of the standard, indicators measuring, analyzing results, identifying actions to improve the quality control of the effect of implemented solutions. We propose in this article, to limit ourselves to the first two phases that we consider most important.