Metagenix Reverse Engineers Data Into Information

Event Summary

Metagenix, Inc. has designed its flagship product, MetaRecon to, as they put it, "Decipher Your Data Genome". The product reverse engineers all of the metadata ("data about data") from data sources and generates information that is very helpful to developers in designing specifications for a new data store, and assists greatly in preparing for cleansing and transformation of existing data. The product will also generate ETL (extract/transform/load) scripts based on the information in the MetaRecon repository that can then be compiled and executed in Ascential Software's DataStage product (DataStage was formerly a product of Ardent Software, which was acquired by Informix in 1999.

Ascential was formerly known as Informix Business Solutions after Informix decided to split their product lines in 2000, basically between database products and business solutions products. IBS did not officially become Ascential until January 15, 2001. It was a circuitous process, but DataStage remains a market leader.) MetaRecon 2.5 also includes support for ETL products from Informatica.

The latest development on the MetaRecon front is support for UNIX as a server platform (currently on Sun Solaris, HP/UX, and Linux, with the possibility of support for other UNIX variants based on customer demand). Since the back-end product was written in PERL, which is a portable computer language, implicit multi-platform support has always been available. According to Mr. Leman, future developments include support for middleware, such as product offerings from New Era of Networks, Vitria, and Metagon.

Among other unique approaches taken by Metagenix:

• All employees are stockholders, and the company's financial books are open to them at all times.
  
  
• Sales Engineers report to Rob Klink, Vice President of Operations, instead of Sales. This eliminates conflicts of interest, between the group that is tasked with demonstrating the product to prospects, and the group tasked with actually selling the product. This practice is unusual among software vendors.
  
  
• Along the same lines, Quality Assurance also reports to Mr. Klink, instead of the vendor standard which is QA reporting to Development. (It's hard for QA to argue a product "feature" with a developer when he/she is your boss.)
  
  
• The company has established a team-based system pitting development against QA. Development competes based on how fast they can fix a bug, and how low they can keep the bug count. QA competes on the basis of how many bugs they can find. This is a good way to encourage higher quality software when the product is finally released.

Market Impact

There are currently few competitors in this market space. MetaRecon is similar in functionality to the Axio Suite from Evoke Software (for more information about Evoke's offering see Evoke Software Releases Axio Data Integration Product), and offers the addition of ETL script generation and a lower price point than Evoke. MetaRecon 2.5 includes a desktop version which provides profiling and analysis, transformation mapping, repository maintenance, reports, DDL and XML generation in a one client/one server license arrangement on Windows NT/2000 for $25,000 per year which competes directly with Evoke Axio on a feature-for-feature basis.

Prior to the release of these products, the only solution was custom coding by in-house IT staff, an expensive and laborious task. Many IT shops still argue that they "can do it better" than an off-the-shelf solution, but IT employee turnover and code maintenance issues make a strong argument for commercially available software.

User Recommendations

Before attempting to acquire software in the "data migration" category, companies should study the scope of their data warehouse/data mart/data consolidation efforts to determine what type of solution is appropriate. Many market surveys have revealed that one of the largest hidden costs in these type of projects is incorrect or "impure" data, so any effort to automate the process of uncovering problems sooner rather than later (typically at the worst possible time, which is when the actual target data load is attempted) will yield significant cost savings to the company.