SAS Data Quality recognised as best data quality package on market

SAS, the world leader in business intelligence, recently announced that Yphise, an independent analyst, has awarded it the Yphise Award for the SAS Data Quality Solution, which is designed to provide quality data prior to its use in analysis, modelling, simulations and other applications.

Yphise recently carried out a study of the data quality packages on the market, published under the title of "Software assessment: data quality packages - May 2004". This study, using Yphise's ISO 9001 certified methodology, looked at more than 300 functional items or techniques classified on the basis of five evaluation criteria:

* Having quality data to make decisions reliable.
* Ensuring the operational handling of data quality.
* Controlling the impact of changes on data quality.
* Reducing the data-related risk in projects.
* Reducing the cost of data quality.

On the basis of these five criteria, SAS Data Quality came out on top.

Data quality, a significant challenge, determines the reliability of information. By improving data quality, organisations enhance potential sources of savings and competitive advantage.

"The relevance of business processes depends on the quality of the data they handle. Low quality can lead to incorrect results, to processes not being available, to regulatory constraints not being complied with.

Operational information systems manage data in numerous databases, which are, moreover, very heterogeneous. This makes guaranteeing data quality much more complex. Data quality packages have powerful functionality to make all the data correctly usable. These packages are now mature. They represent a key investment," notes Xavier Benmoussa, analyst at Yphise.

KRC Research conducted a separate survey that showed 66% of marketing managers think that bad data quality has an impact on the performance of their company. SAS solves these problems.

Data held by a company may be incomplete, erroneous, duplicated or obsolete. This bad quality comes from a number of sources in the company's management processes and systems:

* Lack of upstream control on the data entered into management systems.
* A glitch when transferring data in the information collection process.
* Delay in updating information.
* Heterogeneity of the coding systems.
* Multiplicity of calculation rules in the management systems.
* Multiple inputting in several systems in the same entity.

The consequences of bad data quality are often harmful and costly for the company. They can call into question the production of performance indicators, reduce the credibility of the information system and even lead to financial losses in the case of marketing campaigns.

Data quality is thus a critical factor in the success of a business intelligence project. The process of improving data quality with the SAS Data Quality solution is based on a methodology that is broken down into four phases operating in the form of an iterative cycle.

Trek to the summit

00Days

:

00Hrs

:

00Min

:

00Sec

SAS analyses the data quality and makes it possible to easily obtain, in the form of reports, the level of data missing for every variable of, for example, a client, the erroneous data for a measure or the duplicate entries in client files.

Data quality is determined by the input process in management systems upstream of the company's operational processing chain. Corrective procedures may be put in place, directly at the source, after a data audit to improve the quality.

The SAS Data Quality solution:

* Restructures complex text fields, such as addresses, into several fields according to postal standards.
* Standardizes the content of fields, such as name, position or company.
* Identifies and links clients or suppliers in a single reference table.
* Eliminates duplicate entries in a client file through the use of an algorithm.
* Recognises different descriptions for a single product in different ERP systems.

SAS's ETL (extract, transform and load) capability puts in place alerts in the data collection process for all types of checking: numeric interval, coding, nonexistent record, etc, then automates the data correction even for large volumes.